Dissertation Natascha Pavlovic

Herd Behaviour in Consumer Buying
Decisions in the Age of Social Media

Studies on different influencing aspects
in the mobile communications industry
Der Rechts- und Wirtschaftswissenschaftlichen Fakultät /

dem Fachbereich Wirtschafts- und Sozialwissenschaften
der Friedrich-Alexander-Universität Erlangen-Nürnberg

zur
Erlangung des Doktorgrades Dr. rer. pol.
vorgelegt von
Dipl.-Kffr. Nataša Pavlović-Höck,
aus Nürnberg
Als Dissertation genehmigt
von der Rechts- und Wirtschaftswissenschaftlichen Fakultät /

vom Fachbereich Wirtschafts- und Sozialwissenschaften
der Friedrich-Alexander-Universität Erlangen-Nürnberg
Promotionstermin: 7. Juli 2020
Tag der mündlichen Prüfung: 22. Juni 2020
Vorsitzende/r des Promotionsorgans Prof. Dr. Markus Beckmann
Gutachter/in: Prof. Dr. Michael Amberg

Prof. Dr. Sven Laumer
Acknowledgements
__________________________________________________________________________
Acknowledgements
The research work presented here was created in the last years, when I was an ex-
ternal PhD student at chair of IT Management at the Friedrich-Alexander University of
Erlangen-Nuremberg. I would like to express my special appreciation and thanks to
my doctoral supervisor Prof. Dr. Michael Amberg, who gave me the opportunity to do
my external doctorate at his chair with the flexibility I needed, and at the same time
supported me with his generous advice and trust in my work. Further, I would also
like to thank Prof. Dr. Laumer for taking over the second opinion and Prof. Dr. Beck-
mann for accompanying the defence.
My special thanks go to all people who supported me over the past years and without
whom the dissertation project would not have been possible. First of all, I would like
to thank my husband Ben for his active and untiring support, always motivating me
and believing that I will succeed. Further, I would like to thank my sons, Emilijan,
Damjan, and my newborn son Nikola who helped me to focus my efforts, particularly
in the final stages. I would like to thank Nadja for proofreading parts of my disserta-
tion, her support during my whole dissertation time, and her valuable advice. My sin-
cere thanks belong to Markus, who has proofread the whole dissertation. Further-
more, I would like to thank the colleagues at the chair, especially Doris, Tuba, Sebas-
tian and Sabrina, who each in their own way, have given me the best possible sup-
port.
I would especially like to thank two researchers Prof. Dr. David Sundaram and Dr.
Gabrielle Peko, whom I met during my doctoral thesis and who supported me ex-
traordinarily in presenting my publication.
Last but not least, I would like to express my thanks to my family and circle of friends
as regards to their understanding and moral support during the recent years. I would
especially like to thank my sister Maja who has supported me in many parts of my re-
search work, my cousin Mirjana who has made it possible for me to devote time to
my work. My parents Milanka and Radoslav for their unshakeable belief in my work
and their engagement.
III
Abstract
__________________________________________________________________________
Abstract
Social media has developed to become an integral part of people's lives. It is used on
a daily basis to communicate, share or rate, and thus the opportunities to observe
other consumers' decisions has significantly increased. Likewise, considering herd
behaviour — according to which people discount their own information to imitate oth-
ers — the transparency of other people's opinions by social media might influence
consumers' buying decisions. As social media has become one of the most important
sources of information in consumers' purchasing decision process, firms need to face
these challenges accordingly. This can result in both opportunities (e.g. through the
targeted use of social media to promote products) and risks (e.g. uncontrolled nega-
tive exchange of information about products). It is therefore essential for firms to build
up a better understanding of customer behaviour along the entire buying decision
process in the context of social media. This is particularly critical for industries, whose
business model is based on contractual relationships with customers and which op-
erate in a highly competitive market. The mobile communications industry is a good
example of such an industry; therefore, it is in their interest to deal with the entire
buying decision process to ensure that customers once acquired remain loyal to them
for as long as possible.
Nonetheless, despite a large body of literature exploring the single topics of herd be-
haviour, buying decisions and social media, little is known about the complex inter-
play of the three topics, and the focus on the individual pre-buying, buying and post-
buying decision phases. Thus, based on a comprehensive literature review, catego-
ries were developed through a qualitative summary of the material, which give a ho-
listic overview of the combined research area. Then, the topic is specifically ad-
dressed in the mobile communications industry from different perspectives in three
further empirical studies.
Through the analysis of user-generated content of three different social media plat-
forms relating one mobile provider, and using a qualitative content analysis, the so-
cial interactions on mobile communications topics are examined more closely. The
findings provide novel social interaction categories, which indicate that some interac-
tions seem to be specific to certain platforms.
IV
Abstract
__________________________________________________________________________
Based on conducted expert interviews, determining factors of herd behaviour in buy-

ing decisions influenced by social media in the mobile communications industry are
identified. The findings show that there is no single factor, but rather multiple factors
like drivers of social media usage, other influencing groups, segment or product spe-
cifics or social media activity.
To develop a deeper understanding on herd behaviour along the buying decision

phases in the context of social media, an online survey was conducted with regards
to mobile products. The findings show that subjects use the information from others
as cues in making buying decisions, whereby the influence decreases along the buy-
ing process and in the post-buying phase, where people no longer seem to follow the
herd. Additionally, recommendations of other consumers exert a stronger influence
on subject choices than those of experts. Finally, friends seem to influence consum-
ers' choices more than social media. Product Involvement shows a certain influence
in the buying phase, and with regards to social media activity, people who consider
themselves to be participants are probably most likely to be influenced by rating plat-
forms.
In summary, this dissertation provides a comprehensive insight into the field of re-
search on herd behaviour in the consumer buying decision process, considering dif-
ferent influencing aspects. Moreover, it shows interesting opportunities for further re-
search in this field and presents valuable suggestions for practitioners in the mobile
communications industry.
V
Abstract (German)
__________________________________________________________________________
Abstract (German)
Soziale Medien haben sich zu einem integralen Bestandteil im Leben der Menschen
entwickelt. Sie werden täglich genutzt, um zu kommunizieren, zu teilen oder zu be-
werten. Damit haben die Möglichkeiten, die Entscheidungen anderer Konsumenten
zu beobachten, erheblich zugenommen.
So kann die Transparenz über die Meinungen anderer durch soziale Medien die
Kaufentscheidungen der Verbraucher beeinflussen, wenn man das Herdenverhalten
betrachtet - wonach Menschen eigene Informationen verwerfen, um andere nachzu-
ahmen. Da die sozialen Medien zu einer der wichtigsten Informationsquellen im
Kaufentscheidungsprozess der Verbraucher geworden sind, müssen sich die Unter-
nehmen diesen Herausforderungen entsprechend stellen. Daraus können sich so-
wohl Chancen (z.B. durch den gezielten Einsatz von sozialen Medien zur Vermark-
tung von Produkten) als auch Risiken (z.B. unkontrollierter negativer Informations-
austausch über Produkte) ergeben. Für Unternehmen ist es daher unerlässlich, ein
besseres Verständnis des Kundenverhaltens entlang des gesamten Kaufentschei-
dungsprozesses im Kontext von sozialen Medien aufzubauen. Dies ist besonders kri-
tisch für Branchen, deren Geschäftsmodell auf Vertragsbeziehungen mit Kunden ba-
siert und die in einem stark umkämpften Markt agieren. Die Mobilfunkindustrie ist ein
gutes Beispiel für eine solche Branche, daher liegt es in ihrem Interesse, sich mit
dem gesamten Kaufentscheidungsprozess zu befassen, um sicherzustellen, dass
einmal gewonnene Kunden so lange wie möglich gebunden bleiben.
Dennoch ist trotz einer großen Menge an Literatur, die sich mit den einzelnen The-
men Herdenverhalten, Kaufentscheidungen und soziale Medien befasst, nur wenig
bekannt über das komplexe Zusammenspiel der drei Themen und die Betrachtung
der einzelnen Kaufentscheidungsphasen darin. Daher wurden auf der Grundlage ei-
ner umfassenden Literaturanalyse durch eine qualitative Zusammenfassung des Ma-
terials, Kategorien entwickelt die einen holistischen Überblick über das kombinierte
Forschungsgebiet geben. Anschließend wird das Thema, speziell in der Mobilfunk-
branche aus verschiedenen Perspektiven in drei weiteren Studien, beleuchtet.
Durch die Analyse von nutzergenerierten Inhalten dreier verschiedener 'Social-

Media-Plattformen' bezogen auf einen Mobilfunkanbieter, werden mit Hilfe einer qua-
litativen Inhaltsanalyse die sozialen Interaktionen zum Thema Mobilfunk näher unter-
VI
Abstract (German)
__________________________________________________________________________
sucht. Die Ergebnisse liefern neue Kategorien zu sozialen Interaktionen, die darauf
hindeuten, dass einige Interaktionen für bestimmte Plattformen spezifisch zu sein
scheinen.
Auf der Grundlage von durchgeführten Experteninterviews werden entscheidende

Faktoren des Herdenverhaltens bei Kaufentscheidungen, die durch soziale Medien in
der Mobilfunkbranche beeinflusst werden, identifiziert. Die Ergebnisse zeigen, dass
es nicht einen einzelnen Faktor gibt, sondern mehrere Faktoren wie Treiber der Nut-
zung von sozialen Medien, andere Einflussgruppen, Segment- oder Produktspezifika
oder Social-Media-Aktivität an sich.
Um ein tieferes Verständnis für das Herdenverhalten entlang der Kaufentschei-

dungsphasen im Kontext von sozialen Medien zu entwickeln, wurde eine Online-
Umfrage zu Mobilfunkprodukten durchgeführt. Die Ergebnisse zeigen, dass die Test-
personen die Informationen von anderen als Anhaltspunkte für ihre Kaufentschei-
dungen nutzen, wobei der Einfluss entlang des Kaufprozesses und in der Phase
nach dem Kauf, in der die Menschen anderen in ihren Entscheidungen nicht mehr zu
folgen scheinen, abnimmt. Zudem haben Empfehlungen anderer Konsumenten einen
stärkeren Einfluss auf die Kaufentscheidung als die von Experten. Schließlich schei-
nen Freunde die Entscheidungen der Verbraucher stärker zu beeinflussen als soziale
Medien. Das Produkt-Invovement zeigt einen gewissen Einfluss in der Kaufphase
und in Bezug auf die Aktivität in sozialen Medien werden Personen, die eine mittlere
Aktivitätsrate aufweisen, wahrscheinlich am ehesten von Bewertungsplattformen be-
einflusst.
Zusammenfassend bietet diese Dissertation einen umfassenden Einblick in das Ge-

biet der Erforschung des Herdenverhaltens im Kaufentscheidungsprozess, unter Be-
rücksichtigung verschiedener Einflussaspekte. Zudem werden interessante Möglich-
keiten für weitere Forschungen auf diesem Gebiet aufgezeigt und wertvolle Anre-
gungen für das Management in der der Mobilfunkindustrie abgeleitet.1
1
The German translation of the title of this dissertation is: Herdenverhalten im Kaufentscheidungspro-
zess im Zeitalter sozialer Medien; Studien zu verschiedenen Aspekten in der Mobilfunkbranche
VII
Overview of Contents
__________________________________________________________________________
Overview of Contents
Acknowledgements .................................................................................................. III
Abstract .................................................................................................................... IV
Abstract (German) ................................................................................................... VI
Overview of Contents ........................................................................................... VIII
Table of Contents .................................................................................................... IX
List of Figures........................................................................................................ XIV
List of Tables ......................................................................................................... XVI
List of Abbreviations ........................................................................................... XVIII
1 Introduction ............................................................................................................ 1
2 Research Background ......................................................................................... 14
3 Research Design .................................................................................................. 45
4 Study I: State-of-the-Art Analysis ....................................................................... 64
5 Study II: Social Interactions on Different Social Media Platforms ................. 103
6 Study III: Factors affecting Herd Behaviour in the Context of Social Media. 119
7 Study IV: Herding along the Consumer Buying Decision Process ............... 134
8 Overall Discussion ............................................................................................. 200
9 Conclusion ......................................................................................................... 207
References ............................................................................................................. 218
Appendix A: Supplement to Chapter 3 ................................................................ 249
Appendix B: Supplement to Chapter 4 ................................................................ 251
Appendix C: Supplement to Chapter 5 ................................................................ 276
Appendix D: Supplement to Chapter 6 ................................................................ 284
Appendix E: Supplement to Chapter 7 ................................................................ 298
VIII
Table of Contents
__________________________________________________________________________
Table of Contents
Acknowledgements .................................................................................................. III
Abstract .................................................................................................................... IV
Abstract (German) ................................................................................................... VI
Overview of Contents ........................................................................................... VIII
Table of Contents .................................................................................................... IX
List of Figures........................................................................................................ XIV
List of Tables ......................................................................................................... XVI
List of Abbreviations ........................................................................................... XVIII
1 Introduction ............................................................................................................ 1
1.1 Motivation ......................................................................................................... 2
1.2 Research Focus ................................................................................................ 5
1.3 Objective of the Dissertation ............................................................................. 6
1.4 Structure of the Dissertation ............................................................................. 9
1.5 Bibliographic Notes ......................................................................................... 12
1.6 Conventions .................................................................................................... 13
2 Research Background ......................................................................................... 14
2.1 Herd behaviour ............................................................................................... 15

2.1.1 Introduction of Terms ............................................................................. 15
2.1.2 Applied Theories .................................................................................... 20
2.1.3 Established Frameworks........................................................................ 22
2.2 Consumer Buying Decision Process ............................................................... 24

2.2.3 Established Models ................................................................................ 30
IX
Table of Contents
__________________________________________________________________________
2.3 Social media ................................................................................................... 35

2.3.3 Established Frameworks........................................................................ 41
3 Research Design .................................................................................................. 45
3.1 IS Research Paradigms and Methods ............................................................ 46

3.1.1 IS Research Paradigm ........................................................................... 46
3.1.2 IS Research Methods ............................................................................ 48
3.2 Applied Research Methodologies ................................................................... 50

3.2.1 Qualitative Methods ............................................................................... 50
3.2.1.1 Nonreactive Data Collection on the Internet .............................. 51
3.2.1.2 Expert Interviews ....................................................................... 52
3.2.1.3 Qualitative Content Analysis ...................................................... 53
3.2.2 Quantitative Methods ............................................................................. 55
3.2.2.1 Survey ...................................................................................... 56
3.2.2.2 Statistical Analysis ..................................................................... 57
3.3 Research Strategy .......................................................................................... 58

3.3.1 Research Objectives and Questions ...................................................... 58
3.3.2 Research Paradigm ............................................................................... 61
3.3.3 Research Approach ............................................................................... 61
4 Study I: State-of-the-Art Analysis ....................................................................... 64
4.1 Literature Review Approach............................................................................ 65
4.2 Results of the Quantitative Analysis ............................................................... 71

4.2.1 Contributing Research Streams ............................................................. 72
4.2.2 Applied Methods .................................................................................... 73
4.2.3 Applied Theoretical Lenses.................................................................... 74
4.2.4 Overview of identified Key Topics .......................................................... 76
4.2.5 Overview of aspects affecting the social media impact .......................... 77
4.3 Results of the Qualitative Analysis of the Literature Set ................................. 78

4.3.1 Key Topics regarding Herd Behaviour in purchasing decisions ............. 78
4.3.1.1 Information Type ........................................................................ 79
X
Table of Contents
__________________________________________________________________________
4.3.1.2 Information Sources ................................................................... 82

4.3.1.3 Technology Adoption ................................................................. 85
4.3.2 Aspects affecting the impact of social media on herding in purchasing . 87
4.3.2.1 Popularity ................................................................................... 87
4.3.2.1 Information Quality ..................................................................... 89
4.3.2.2 Valence ...................................................................................... 90
4.3.2.1 Product Type.............................................................................. 92
4.3.2.2 Homophily .................................................................................. 94
4.3.2.3 Others ...................................................................................... 95
4.4 Summary and Interim Discussion ................................................................... 98
5 Study II: Social Interactions on Different Social Media Platforms ................. 103
5.1 Objectives of the Study ................................................................................. 104
5.2 Theoretical Foundation ................................................................................. 106

5.2.1 Social Media Classification .................................................................. 106
5.2.2 Social Interactions on Social Media ..................................................... 107
5.3 Methodical Approach .................................................................................... 109

5.3.1 Empirical Field ..................................................................................... 109
5.3.2 Data Collection .................................................................................... 110
5.3.3 Data Analysis ....................................................................................... 111
5.4 Results .......................................................................................................... 112
5.5 Summary and Interim Discussion ................................................................. 117
6 Study III: Factors affecting Herd Behaviour in the Context of Social Media . 119

6.2.1 Herd Behaviour in the Context of Social Media ................................... 121
6.2.2 Buying Decisions in the Context of Social Media ................................. 122

6.3.1 Empirical Field ..................................................................................... 123
6.3.2 Data collection ..................................................................................... 124
6.3.3 Data analysis ....................................................................................... 125
XI
Table of Contents
__________________________________________________________________________
6.4 Results .......................................................................................................... 126
7 Study IV: Herding along the Consumer Buying Decision Process ............... 134

7.2.1 Herd Behaviour in Buying Decisions .................................................... 137
7.2.2 Different Information Sources .............................................................. 138
7.2.3 Product Differentiation ......................................................................... 139
7.2.4 Social Media Activity ............................................................................ 141
7.2.5 Personality Types ................................................................................ 143
7.3 Hypotheses ................................................................................................... 146

7.3.1 Herd Behaviour in the Buying Decision Phases ................................... 147
7.3.2 Herd Behaviour by Information Sources .............................................. 149
7.3.3 Herd Behaviour by Product Type ......................................................... 151
7.3.4 Herd Behaviour by Social Media Activity ............................................. 152
7.3.5 Herd Behaviour by Personality Type ................................................... 154

7.4.1 Procedure and Sample Profile ............................................................. 155
7.4.2 Research Design ................................................................................. 156
7.4.2.1 Structure of the Survey ............................................................ 156
7.4.2.2 Comparison Groups ................................................................. 159
7.4.2.3 Items and Scales ..................................................................... 162
7.4.3 Data Preparation and Analytical Methods ............................................ 166
7.5 Descriptive Results ....................................................................................... 167

7.5.1 Demographics...................................................................................... 167
7.5.2 Information Sources ............................................................................. 168
7.5.3 Social Media Activity ............................................................................ 171
7.5.4 Product Involvement per Product Type ................................................ 172
7.5.5 Personality Types ................................................................................ 175
7.6 Statistical Analysis ........................................................................................ 176

7.6.1 Herd Behaviour along the Buying Decision Process ............................ 176
7.6.2 Herd Behaviour by Information Sources .............................................. 179
XII
Table of Contents
__________________________________________________________________________
7.6.3 Herd Behaviour by Product-Involvement ............................................. 182

7.6.4 Herd Behaviour by Social Media Activity ............................................. 185
7.6.5 Herd Behaviour by Personality Type ................................................... 190
7.6.6 Strongest Influence on Herd Behaviour ............................................... 192
8 Overall Discussion ............................................................................................. 200
8.1 Contribution to IS Research .......................................................................... 201
8.2 Contribution from a Management Perspective .............................................. 204
9 Conclusion ......................................................................................................... 207
9.1 Research Questions addressed .................................................................... 208
9.2 Limitations..................................................................................................... 211
9.3 Further Research Opportunities .................................................................... 215
References ............................................................................................................. 218
Appendix A: Supplement to Chapter 3 ................................................................ 249
Appendix B: Supplement to Chapter 4 ................................................................ 251
Appendix C: Supplement to Chapter 5 ................................................................ 276
Appendix D: Supplement to Chapter 6 ................................................................ 284
Appendix E: Supplement to Chapter 7 ................................................................ 298
XIII
List of Figures
__________________________________________________________________________
List of Figures
Figure 1: Dissertation Overview - Chapter 1: Introduction ........................................... 1
Figure 2: Number of Active Social Media Users Worldwide by 2020........................... 3
Figure 3: Research Focus ........................................................................................... 6
Figure 4: Structure of the Dissertation ......................................................................... 9
Figure 5: Dissertation Overview - Chapter 2: Research Background ........................ 14
Figure 6: Herding Framework .................................................................................... 23
Figure 7: Information Processing Model .................................................................... 28
Figure 8: Model of Buyer Behaviour .......................................................................... 31
Figure 9: Model of Consumer Behaviour ................................................................... 32
Figure 10: SR-Model - Schematic Diagram ............................................................... 33
Figure 11: SOR Model ............................................................................................... 34
Figure 12: The three Phases of the Buying Decision Process................................... 35
Figure 13: Classification Scheme of Social Media..................................................... 43
Figure 14: S-O-R Framework for Social Media.......................................................... 44
Figure 15: Dissertation Overview - Chapter 3: Research Design .............................. 45
Figure 16: Method Profile of WI in the years 2007-2012 ........................................... 49
Figure 17: Dissertation Overview - Chapter 4: State-of-the-Art Analysis ................... 64
Figure 18: Steps of Literature Review ....................................................................... 65
Figure 19: Temporal Classification of Identified Publications .................................... 71
Figure 20: Contributing Research Streams of the Literature Set Investigated ........... 72
Figure 21: Applied Methods of Literature Set Investigated ........................................ 74
Figure 22: Applied Theoretical Lenses of Literature Set Investigated ....................... 76
Figure 23: Key Topics Identified in Investigated Literature Set ................................. 77
Figure 24: Identified Aspects affecting the Social Media Impact ............................... 78
XIV
List of Figures
__________________________________________________________________________
Figure 25: Summary of the Qualitative Literature Review Results ............................ 99
Figure 26: Dissertation Overview - Chapter 5: Study II ........................................... 103
Figure 27: Classification of Social Media ................................................................. 107
Figure 28: Dissertation Overview - Chapter 6: Study III .......................................... 119
Figure 29: Overview of the Seven Identified Influencing Factors in the Mobile
Communications Industry (Main Categories)........................................ 126
Figure 30: Dissertation Overview - Chapter 7: Study IV .......................................... 134
Figure 31: Overview of the Distribution Process to the Comparison Groups ........... 158
Figure 32: Information Sources within the Buying Decision Process....................... 168
Figure 33: Overview of Social Media Activity........................................................... 172
Figure 34: Product Involvement per Product Type .................................................. 175
Figure 35: Distribution of Personality Types in the Sample ..................................... 176
Figure 36: Product Involvement per Product Type .................................................. 183
Figure 37: Dissertation Overview - Chapter 8: Overall Discussion .......................... 200
Figure 38: Dissertation Overview - Chapter 9: Conclusion ...................................... 207
XV
List of Tables
__________________________________________________________________________
List of Tables
Table 1: A Sampling of Herding Definitions and History ............................................ 16
Table 2: Theories of Herding ..................................................................................... 20
Table 3: A Sampling of Social Media Definitions from IS Perspective ....................... 36
Table 4: Overview of Personal Behaviour Theories .................................................. 38
Table 5: Overview of Social Behaviour Theories ....................................................... 40
Table 6: Overview of Mass Behaviour Theories ........................................................ 41
Table 7: Research Methods applied in Each Research Sequence ............................ 62
Table 8: Overview of Research Sources ................................................................... 66
Table 9: Reduction Process of the Literature Review ................................................ 68
Table 10: Frequency Distribution of Social Interactions across Platforms (Main

Categories) ............................................................................................. 113
Table 11: Frequency Distribution of Social Interactions related to Mobile

Communications Topics across Platforms (Subcategories) ................... 114
Table 12: Cross-tabulation: Top 5 Main Categories and Top 5 Subcategories ....... 116
Table 13: Overview of Frequency Distribution of the Subcategories ....................... 129
Table 14: Overview of developed Hypotheses ........................................................ 146
Table 15: Structure of the Survey ............................................................................ 157
Table 16: Provided information in the Pre-Buying Phase ........................................ 160
Table 17: Provided information in the Buying Phase ............................................... 161
Table 18: Provided information in the Post-Buying Phase ...................................... 162
Table 19: Data Cleansing ........................................................................................ 166
Table 20: Other Information Sources within the Buying Decision Process .............. 169
Table 21: Reasons for choosing the Information Sources ....................................... 170
Table 22: Product Involvement of a Smartphone .................................................... 172
Table 23: Product Involvement of a Mobile Contract ............................................... 173
XVI
List of Tables
__________________________________________________________________________
Table 24: Product Involvement of an Entertainment Service ................................... 174
Table 25: Herd Behaviour in the Pre-Buying Phase ................................................ 177
Table 26: Herd Behaviour in the Buying Phase ....................................................... 177
Table 27: Herd Behaviour in the Post-Buying Phase .............................................. 178
Table 28: Herd Behaviour in the Pre-Buying Phase with Different information
Sources (1st information round) ............................................................ 179
Table 29: Herd Behaviour in the Pre-Buying Phase with Different information
Sources (2nd information round) ........................................................... 180
Table 30: Herd Behaviour in the Buying Phase with Different information Sources 181
Table 31: Herd Behaviour in the Post-Buying Phase with Different information
Sources .................................................................................................. 182
Table 32: Herd Behaviour by Product Involvement in the Pre-Buying Phase .......... 184
Table 33: Herd Behaviour by Product Involvement in the Buying Phase ................ 184
Table 34: Herd Behaviour by Product Involvement in the Post-Buying Phase ........ 185
Table 35: Herd Behaviour by Social Media Activity in the Pre-Buying Phase ......... 186
Table 36: Herd Behaviour by Social Media Activity in the Buying Phase ................ 187
Table 37: Herd Behaviour by Social Media Activity in the Post-Buying Phase ........ 187
Table 38: Review-/Rating Platform Influence depending on Social Media Activity .. 188
Table 39: Expert Rating influence depending on Social Media Activity ................... 189
Table 40: Facebook influence depending on Social Media Activity ......................... 189
Table 41: Herd Behaviour by Personality Type in the Pre-Buying Phase ................ 190
Table 42: Herd Behaviour by Personality Type in the Buying Phase ...................... 191
Table 43: Herd Behaviour by Personality Type in the Post-Buying Phase .............. 191
Table 44: Overview on Results of the tested Hypotheses ....................................... 196
XVII
List of Abbreviations
__________________________________________________________________________
AIS Association of Information Systems
ARPU Average Revenue per User
B2B Business to Business
B2C Business to Consumer
B2P Business to Partner
eWOM electronic Word of Mouth
FB Facebook
FFM Five-Factor Model of personality
I&KM Information & Knowledge Management
ICT Information and Communication Technology
IS Information Systems
IT Information Technology
MBTI Myers-Briggs Type Indicator
NEO PI-R Neuroticism-Extroversion-Openness Personality Inventory-
Revised
PSI Para-social interaction
RQ Research Question
SEO Search Engine Optimisation
TIPI Ten-Item Personality Inventory
TAM Technology acceptance model
TRA Theory of reasoned action
TPB Theory of planned behaviour
UGC User-generated-content
XVIII
__________________________________________________________________________
UGT Uses and gratifications theory
WI Wirtschaftsinformatik
WOM Word of Mouth
XIX
Introduction
__________________________________________________________________________
1 Introduction
Chapter Overview: This chapter introduces the dissertation at hand, hereby outlin-
ing the motivation behind why such research has been carried out (section 1.1), plac-
ing the research in context as well as presenting target of the doctoral research pro-
ject (section 1.3) and the research focus (section 1.2). Then, the structure of the dis-
sertation with the contents of the chapters therein are outlined (section 1.4), and bib-
liographic notes in terms of published articles (section 1.5) are listed. In the last sec-
tion of this chapter (section 1.6), central conventions are defined.
PART A: Foundations
1 Introduction
2 Research Background
3 Research Design
PART B: Conducted Studies
4 Study I: State-of-the-Art Analysis
5 Study II: Social Interactions on Different Social Media Platforms
Study III: Factors affecting Herd Behaviour in the Context of Social

6 Media
7 Study IV: Herding along the Consumer Buying Decision Process
PART C: Summary
8 Overall Discussion
9 Conclusion
Figure 1: Dissertation Overview - Chapter 1: Introduction
1
Introduction
__________________________________________________________________________
1.1 Motivation
The phenomenon that “everyone does what everyone else is doing, even when their
private information suggests doing something quite different” is not new and has al-
ready been explored by Banerjee in 1992. Herd behaviour is an integral part of hu-
man behaviour, and is an influential and well documented feature of human behav-
iour which has been studied in a number of research disciplines, particularly in social
psychology and economics (Raafat et al. 2009b). "Herding is a form of convergent
social behaviour that can be broadly defined as the alignment of the thoughts or be-
haviours of individuals in a group (herd) through local interaction and without central-
ised coordination" Raafat et al. (2009b, p. 420). Herding in consumers buying deci-
sions is common where people infer information from other people's choices and in-
corporate that information into their buying decision making process (Duan et al.
2009).
With the dissemination of Internet technologies, the transparency of other consumers'

choices has significantly increased, therefore making herding more feasible. Con-
sumers can take a shortcut to follow a herd, according to which people discount their
own information to imitate others. In a time of 'always on' with everyone using the
Internet and being connected through social networks, information and opinions are
shared very quickly. Nevertheless, due to the low level of informativeness, herd be-
haviour remains fragile, e.g. a new information at a later stage might change some-
one's decision (Banerjee 1992; Bikhchandani et al. 1992).
In particular due to the increasing spread of mobile end devices, the use of the Inter-
net and social media has risen. Social media usage has increased by over 80% from
2015 to 2020, and in January 2020 there were about 3.8 billion active social media
users2 worldwide (Figure 2). In Germany, over 50% of the population actively uses
social media (Outlook 2019a). Thus, people increasingly interact with each other to
discuss topics, exchange ideas and seek support. "Social media has become the
pacemaker in modern life - many start and end their day with it" (Bitkom 2018).
2
Active usage refers to usage at least once a month (DataReportal 2020a)
2
Introduction
__________________________________________________________________________
Number of active users in billion

5
3,8
4
3,5
3,2
2,8
3
2,3
2,1
2
0
2015 2016 2017 2018 2019 2020
Figure 2: Number of Active Social Media Users Worldwide by 2020

(DataReportal 2020a)
Young people especially are using social media to a high extent, with more than 50%
of them doing so daily or several times a week (mpfs 2018). Facebook (FB) enjoys
the exalted position of being the market leader of the social media world, with 2.45
billion monthly users worldwide; in Germany, FB is the most popular social network,
with 84% of social media users regularly using this service (DataReportal 2020b).
Given that people rely on the opinions of others, individual behaviour may be influ-
enced by other members of the community (Zhou 2011). This development impacts
consumers as well as firms because this easily accessible information may greatly af-
fect consumers' buying decisions. Therefore, social media are not only popular for
private use, but are also highly relevant for firms. Thus, spending on social media
advertising is continuously increasing and is expected to rise to 3.75 billion euro by
2023 in Germany, which would mean an increase of over 40% in five years (Outlook
2019b).
Today, consumers depend on various information inputs before taking a buying deci-
sion. Thereby, traditional sources like family and friends, test reports, or expert re-
views are crucial in most product purchases (Nakayama et al. 2010). But the emer-
gence of the Internet has introduced a new information input channel: electronic word
of mouth (eWOM), which gives customers the opportunity to easily benefit from the
experience of others for both online and offline buying decisions (Shen et al. 2014).
3
Introduction
__________________________________________________________________________
Thus, it is easier to record and convey information about purchases and opinions of
peers, such as previous sales, reviews/ratings, or online reviews (Liu et al. 2017).
Especially, "online consumer reviews have shown as an important source of informa-

tion that affects individuals’ purchase decision-making" (Shen et al. 2014, p. 1). The
results of an international study in the service sector of the University of St.Gallen
(Maas et al. 2014) show above all the importance of social networks, review/rating
platforms and forums in the buying decision process as a source of information be-
fore a buying decision is made. It is not surprising that rating and review sites, whose
main purpose the evaluation and review of products and services are the most popu-
lar. These are followed by forums and social networks, both social media types in
which friends, acquaintances or even strangers can help one another(Maas et al.
2014). The importance of social media as a source of information for purchasing de-
cisions is particularly evident among younger consumers, since 42.8% of them use
this source of information to buy products (MindTake 2019).
Social media is changing communication and consumers' behaviour patterns and

thus confronting firms with completely new requirements. There are new opportuni-
ties for firms to use social media to foster knowledge sharing within the organisation
and thus to benefit from people's metaknowledge (knowledge of 'who knows what'
and 'who knows whom') (Leonardi 2015; Majchrzak et al. 2013). Reviews and prod-
uct recommendations enable consumers to easily evaluate alternative products be-
fore buying them. Companies generate recommendations that refer customers to a
greater variety of products. By showing customer-generated online reviews, consum-
ers are supported in evaluating the recommended products (Jabr and Zheng 2014).
Moreover, information asymmetries between suppliers and consumers disappear or

are reversed, giving networked consumers power (Akehurst 2008). Firms and organi-
sations have realised that the fact the growth and popularity of social media and their
influence on consumers' decisions have become a powerful tool for marketers. They
can promote their products, and consumers share their opinions.
Social media not only enables consumers to communicate with each other, but also
enables firms to communicate with their customers both during the purchase itself as
well as before and after the purchase. This allows firms to use the post-buying phase
to strengthen customer loyalty, build brand loyalty and influence future buying deci-
sions (Kaplan and Haenlein 2010).
4
Introduction
__________________________________________________________________________
Specifically, the buying decision process involves the consideration of the individual
buying phases (pre-buying, buying, post-buying) to enable a holistic picture in the re-
search approach. This holistic picture is crucial for industries like the mobile commu-
nications industry since its business model is geared towards longer customer rela-
tionships.
Against this background, the motivation of this dissertation is to gain a better under-
standing of herd behaviour in the overall buying decision process and the role of so-
cial media therein. Furthermore, the influence of social media on herding in buying
decisions should be understood in the mobile communications industry.
1.2 Research Focus
Herd behaviour is an integral part of human behaviour, which also influences con-
sumers' purchasing decisions. In addition, the transparency and interaction possibili-
ties of social media seem to have an impact on herd behaviour, and hence on human
decision-making.
The phenomenon of herd behaviour can be investigated from different perspectives,

e.g. with regard to 'business to business' (B2B) or 'business to partner' (B2P) pur-
chasing decisions. The focus of the research project at hand is the 'business to con-
sumer' (B2C) perspective, i.e. consumer buying decisions.
Furthermore, the research focus of this dissertation is on the linkage of the topics:
herd behaviour, buying decision and social media. In other words, it is not the single
topics that are focused on, but the interaction of the topics. The initial focus is on a
general understanding of these topics based on existing literature. Then, these topics
are specifically examined in the mobile communications industry.
In particular, industries whose business model is based on contractual relationships

with customers have to deal with the entire buying process (pre-buying, buying, post-
buying) and in particular with the post-buying phase in order to avoid customers re-
gretting their buying decision and refraining from making a further purchase. Accord-
ingly, it is especially important for practitioners from these industries to pay attention
to the specific behaviour of consumers in the different buying phases. The mobile
communications industry is a good example of such an industry, as 75.3 % of the
customers hold a contract with a mobile provider (VuMA 2019). Despite the fact that
average revenue per user (ARPU) is continuously decreasing (Bundesnetzagentur
2019b), the German mobile communications market is a saturated market with a
5
Introduction
__________________________________________________________________________
penetration rate of 165% and several dozen mobile phone brands offered on different
distribution channels (Bundesnetzagentur 2019a). Consequently, customers can only
be attracted if they are lured away from the competition. Furthermore, the high level
of transparency provided by social media poses a particular challenge to industries
that offer commodity products, like the mobile communications industry, with a low
level of product or price differentiation (Li and Whalley 2002). In particular, if the
products are also experience goods (e.g. smartphone, mobile contract, value-added
services), which is difficult to evaluate prior to their adoption, consumers are moti-
vated to make extensive use of various information sources, hoping to make well-
informed decisions (Gu et al. 2012; Luo and Zhang 2013), and thus they are more re-
liant on the opinions of other consumers.
As described above, this is a market characterised by certain challenges (decreasing

ARPU, strong competitive intensity, lack of vertical and/or horizontal differentiation).
Particularly in this cluttered online environment, it becomes increasingly crucial for
mobile providers to develop an understanding of influencing factors on consumers'
decisions. The following Figure 3 summarises the research focus.
Out-of-Scope
In-Scope
• B2B perspective
General & German mobile communications industry specific
• B2P perspective
understanding of:
• Other industries
Herd behaviour • Other countries

• Herd behaviour only
Social Buying decision • Social media only
media process
(pre-buying, • Buying decisions only
buying, post-
buying)
from a B2C perspective
Figure 3: Research Focus
1.3 Objective of the Dissertation
In today's networked world, with all those who use the Internet and are connected via
social networks, information and opinions are exchanged very quickly. The increas-
6
Introduction
__________________________________________________________________________
ing number of social media makes it complicated for firms to understand the influ-
ence of different types of social media on consumer behaviour and how to deal with
this phenomenon (Hanna et al. 2011). Hence, the central goal of the present disser-
tation is to gain a better understanding of the interplay of social media, herd behav-
iour and consumers' buying decisions.
 First, the theoretical foundation should be set. This enables a general theoreti-
cal basis for this work to be established and further empirical studies to be
embedded in the respective research context. By means of a literature review,
an overview of the main topics addressed by scientific literature on herd be-
haviour in the consumer buying decision process is provided. Further aspects
that address the influence of social media on herd behaviour are also consid-
ered and elaborated.
 Second, with regards to the influence on consumer buying decisions and
therefore the importance for B2C firms, the elements content and social inter-
action seem to play a major role. Considering social interactions, there is
scarce research exploring deeper social interactions on different social media
platforms and specification with regards to the mobile communications indus-
try. Thus, the target is to identify how social interactions differ depending on
different social media platforms, and to cover mobile communications specifics
therein. The current work aims to elaborate on this issue and provide novel
and intriguing categories.
 Third, the widely spread social media usage and as a consequence a high
transparency of consumers' decisions and opinions poses a particular chal-
lenge to such industries as the mobile communications industry. Therefore, it
is crucial to understand what factors affect the influence of social media on
herd behaviour in the buying decision process. Hence, the present research
work elaborates more on these issues.
 Fourth, a further elaboration of individual buying decision phases and the
overall holistic consideration in the present dissertation can thus contribute to
the further theory development of herd behaviour. To investigate how herd
behaviour generally differs among the consumer buying decision could con-
tribute to ongoing theorising about herd behaviour. Further, particularly for
practitioners whose business model is based on contractual relationships and
loyal customers, as is the case in the mobile communications industry; it is
7
Introduction
__________________________________________________________________________
critical to success to understand how herd behaviour is affected in individual

buying decision phases and how the influence of different factors differ. Par-
ticularly, considering the numerous sources of social media (e.g. social net-
work, review/rating platforms, forums) that may greatly affect consumers' buy-
ing decisions.
By addressing the topics listed above, the objective of the present research work is to
contribute to a deeper understanding of herd behaviour in the individual buying deci-
sion phases by considering the social media context. Furthermore, the aim is to en-
rich existing Information Systems (IS) literature, provide a comprehensive basis for
further research, and generate valuable suggestions for practitioners in the mobile
communications industry.
8
Introduction
__________________________________________________________________________
1.4 Structure of the Dissertation
The structure of the dissertation at hand is based on different phases of the research
project, and comprises in total nine chapters (Figure 4).
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 4: Structure of the Dissertation
PART A: FOUNDATIONS
The first part of the dissertation lays the foundation for the studies conducted, and
consists of chapters one to three.
Chapter 1
This chapter introduces the studied phenomenon. The motivation is outlined, the re-
search focus deducted, and target of the doctoral research project presented. Fur-
thermore, the structure is specified as well as related works which have been already
published. The chapter closes with the deployed conventions.
9
Introduction
__________________________________________________________________________
Chapter 2
In this chapter the background of the research is provided, in order to explain the ba-
sic concepts and their application in the context of the dissertation at hand. There-
fore, key terms, theoretical concepts, and frameworks are outlined. First, herd behav-
iour with its underlying mechanics is introduced. Then, the 'buying decision process',
and finally the current understanding of 'social media' and respective theories and
frameworks are presented.
Chapter 3
This chapter explains the research design of the present dissertation. First, the aca-
demic discipline of IS research is introduced with its methodological paradigms, and
underlying philosophical perspectives. Next, the applied qualitative methods (non-
reactive data collection, expert interviews, qualitative content analysis), and quantita-
tive methods (online survey, and statistical analysis) are presented and described.
Finally, the research strategy is elucidated, for which the research questions (RQ) are
defined, the applied research paradigm briefly introduced, and the research approach
described.
PART B: CONDUCTED STUDIES
The second part of the dissertation presents details of the studies conducted. It starts
with the state-of-the-art research on the phenomenon studied. This is followed by
three empirical studies, including theoretical foundations, details on the methodical
approach, and the results of the studies.
Chapter 4
This chapter presents the results of the literature review on herd behaviour in buying
decisions, and the aspects of social media therein. It begins with a description of the
literature review approach, followed by the descriptive statistical results of the litera-
ture set identified and the qualitative assessment comprising the elaboration with re-
gards to herd behaviour in the buying process, and the aspects affecting the impact
of social media on herding in purchasing. A summary and an intermediate discussion
conclude the chapter.
Chapter 5
This chapter presents the results of an explorative study on social interactions on dif-
ferent social media platforms with regard to one mobile provider. First, the specific
10
Introduction
__________________________________________________________________________
objectives and the research questions are outlined and put into context with existing
research. Then, the theoretical background on social media classification and social
interactions on social media is discussed. The description of the methodical ap-
proach refers to the empirical field of social media, the non-reactive data collection
and qualitative content analysis. This is followed by the presentation of the evaluated
results. Finally, the chapter concludes with an interim discussion of the results.
Chapter 6
The study presented in this chapter attempts to identify the determining factors of
herd behaviour in buying decisions influenced by social media in the mobile commu-
nications industry. The factors are identified on the basis of a qualitative content
analysis, which is based on content from conducted expert interviews. At the begin-
ning, the objectives of the study and the research question are outlined and put into
the context of existing literature. A discussion of theoretical background on herd be-
haviour and the buying decision process in the context of social media follows. The
research methodology, including the empirical field, data collection, and data analysis
are displayed afterwards. Then, the evaluated results are presented and the interim
discussion of the study is presented.
Chapter 7
In this chapter, the most comprehensive study of the present dissertation is pre-
sented. In an empirical study, various factors influencing herd behaviour along the
buying decision process for mobile products are examined. First, the specific objec-
tives and the research question are outlined. Then, the theoretical background on the
topics of herd behaviour in buying decisions, information sources, product differentia-
tion, social media activity, and personality types, are discussed. Following this, hy-
potheses are developed, the research methodology outlined, including the procedure
and sample profile, research design, and data analysis. Next, the descriptive results
and the analytical results are presented. Finally, the chapter is concluded by provid-
ing an interim discussion of the results.
PART C: SUMMARY
The last part of the dissertation provides a comprehensive summary, including an

overall discussion and conclusion of the research work.
11
Introduction
__________________________________________________________________________
Chapter 8
This chapter summarises and discusses the key findings of the individual studies.
Specifically, it presents the contribution to recent research on herd behaviour in buy-
ing decisions, and describes how practitioners can profit from the results. Finally, the
limitations are outlined.
Chapter 9
This chapter concludes the research work by referring back to the research questions
posed. The assignment’s topic and its meaning are summed up once again and pro-
vides an outlook on potential future research opportunities in the area of herd behav-
iour in purchasing decisions in the context of social media.
1.5 Bibliographic Notes
Within the framework of the doctoral regulations, doctoral students are expected to
publish partial results and individual aspects of their work already during the qualifi-
cation phase of the doctorate.
A clear reference to these prior publications is intended to ensure transparency and

scientific honesty in this work. A corresponding reference can also be found in the in-
dividual chapters.
The results presented in the dissertation at hand are partly based on or refer to as-
pects of already published work.
The corresponding publication is:
Pavlovic, N. (2018). "Factors Affecting Herd Behaviour in Buying Decisions Influ-

enced by Online Communities," In Proceedings of the 24th Americas Conference on
Information Systems (AMCIS), New Orleans 2018
The publication (Pavlovic 2018) is considered in Study III and deals with factors af-
fecting herd behaviour in consumer buying decisions. The results are based on con-
ducted expert interviews. The core results as presented in the published work
(largely) remain, and are presented in chapter 6. Furthermore, individual parts of the
paper are also used in the remaining parts of the dissertation, e.g. in the abstract or
in chapter 2 or chapter 8.
12
Introduction
__________________________________________________________________________
1.6 Conventions
In the present dissertation the following conventions are defined:
 The structure of this research work is based on the DIN standard 1421. The
first level of outline is called a chapter. The second level of outline is called a
section. To simplify matters the subsection titles are also referred to as a sec-
tion and are followed by is referenced by the corresponding numbering.
 Figures and tables have been taken over as appropriate in the caption of the
illustration. Figures and tables without reference are own representations.
 Bibliographical references are based on the guidelines from MIS Quarterly.
13
Research Background
__________________________________________________________________________
Chapter Overview: In order to clarify the research objectives of the present disserta-
tion, key terms and theoretical concepts are introduced and defined in the following.
Therefore, section 2.1 introduces the term 'herd behaviour' and explains the underly-
ing mechanics. Subsequently, in section 2.2 the 'buying decision process' is intro-
duced and defined, and established models presented. The current understanding of
'social media' and respective theories and frameworks are presented in the last sec-
tion.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 5: Dissertation Overview - Chapter 2: Research Background
14
Research Background
__________________________________________________________________________
2.1 Herd Behaviour
In principle, there is a broad consensus on the concept of herd behaviour, but the
history of the concept in the various research areas shows great differences, both in
terms of the underlying assumptions and the mechanisms. Herd behaviour is an in-
tegral part of human behaviour, but the concept has attracted the attention of the
economic and social sciences, mainly through the analysis of stock markets and the
irrational 'exuberance' of investors (Raafat et al. 2009b).
2.1.1 Introduction of Terms
In the following, the phenomenon of herd behaviour is introduced and defined.
"Herding is a form of convergent social behaviour that can be broadly defined as the
alignment of the thoughts or behaviours of individuals in a group (herd) through local
interaction and without centralized coordination" Raafat et al. (2009b, p. 420). This
phenomenon is an influential and well documented feature of human behaviour
which has been studied in a number of research fields, particularly in social psychol-
ogy and economics. In the literature, the terms herding and herd behaviour are used
as synonyms, this is also applied in the present dissertation.
Early economists like Veblen (1899) or sociologists Simmer (1910) applied the phe-
nomenon of herd behaviour to sudden changes in consumer behaviour such as fash-
ions and fads. Further, herding was among the first topics investigated in social psy-
chology by Van Ginneken (1992). Nevertheless, herd behaviour the economic ap-
proach differs to the one in social psychology. As a result, researchers are con-
fronted with different ideas about herd behaviour. However, herd behaviour cannot
be fully understood from a single perspective. Although both disciplines address the
question of what herd behaviour is, the economic perspective is primarily concerned
with the long-term effects to be expected and examining the value of herd behaviour
to understand how much one can benefit from it. The motivations underlying herd
behaviour are considered in terms of the decisions they produce (Rook 2006). The
psychological perspective, however, takes into account the subjective value of herd
behaviour. It seeks to answer why and when people feel motivated to engage in col-
lective behaviour, to better understand the motivations underlying herd behaviour
with regard to the underlying process (Rook 2006).
Comprehensive literature reviews have also been carried out on the topic in various
research disciplines, although a large part is related to herding behaviour in financial
15
Research Background
__________________________________________________________________________
markets, e.g. (Bikhchandani and Sharma 2000; Hirshleifer and Hong Teoh 2003;
Spyrou 2013; Van Campenhout and Verhestraeten 2010). Table 1 provides a selec-
tive overview of herding definitions and mechanisms from different disciplines, thus
again illustrates the diversity of disciplines in which the phenomenon is addressed
from distinct levels of analysis.
Table 1: A Sampling of Herding Definitions and History
Author Discipline Definition Mechanism

(Year)
Le Bon Psychology the 'ancestral savagery of the The concept of collective

(1897) collective mind' and the 'loss hypnosis, in the sense of an
of self'. irrational and subconscious
social contagion.
Veblen Economics/ A comparison with similar The instinct for imitation.

(1899) Sociology people who are better than
oneself.
Tarde (1903) Psychology Collective hypnosis called The 'group mind' as an ex-
'social somnambulism'. Peo- planatory factor of group
ple get involved in crowd ac- psychology and herd be-
tions 'in mental unity'. haviour.
Simmel Sociology The consideration of social 'Impulse for sociability' and

(1910) networks in connection with branching into loosely
herd behaviour. woven networks.
Freud (1922) Psychiatry/ A group is considered an Through the herd instinct,

Medicine obeying herd that could not the individual loses his abil-
exist without a leader. ity to be critical and allows
himself to be influenced
emotionally. The simpler
the emotional impulse is,
the more likely it is to
spread within a group.
Allport (1924) Psychology Collective behaviour should According to non-irrational

be explained from the per- suggestiveness, individuals
16
Research Background
__________________________________________________________________________

(Year)
spective of individual psy- behave in a crowd because

chology.. they believe others share
their feelings and beliefs.
Keynes Economics Contagious 'animal spirits' New information is not effi-

(1936) moving the market. ciently processed by indi-
viduals, as they do not
know which information is
relevant. Usual behaviour
transforms easily into herd
behaviour.
Sherif (1936) Psychology An autokinetic effect is used The common frame of ref-
to demonstrate that individu- erence is social norms,
als use the judgements of which try to give meaning to
others to approximate the so- social reality with a rational
cial norms of the group. attempt.
Leibenstein Social Psy- Using the bandwagon meta- Social influence as underly-
(1950) chology phor: the degree to which the ing motivation; to adapt to
demand for a product is in- the people they want to be
creased by the fact that oth- associated with; to be fash-
ers also consume the same ionable or stylish.
product.
Penrose Genetics A comparison of the devel- Highlighted that both physi-

(1952) opment of behaviour during cal and psychological epi-
mass hysteria with the epi- demics depend on viru-
demiology of a disease. lence, transmission and re-
ceptivity.
Asch (1956) Psychology The convergence in the well- Through rational attempts
known line experiments re- by individuals, attempts are
veals the pressure to conform made to give meaning to
to a mistaken view. social reality in order to
share group perceptions.
17
Research Background
__________________________________________________________________________

(Year)
Rogers Sociology Diffusion is defined as a Diffusion of innovations and

(1962) mechanism whereby informa- analysis of social networks
tion is exchanged over time as two underlying mecha-
between the members of a nisms.
social system through certain
channels.
Moscovici et Social Psy- It is assumed that social in- Social conflicts lead to in-
al. (1969) chology fluence causes conformity, security and the readiness
which shapes cognitive un- to agree on opposing points
certainty about social reality of view.
in given situations.
Granovetter Sociology The strength of weak ties in- Social networks in diffusion
(1973) terpersonal ties are defined processes emphasise the
as information-carrying con- importance of weak ties.
nections between people. According to this, a larger
number of people are
reached and greater social
distances are overcome if
weak ties are used instead
of strong ones.
Katz and Economics Herd behaviour as a long- The advantage of a product

Shapiro term and stable network ef- from the user's point of view
(1985) fect. depends on the number of
other users who are on the
same 'network' as that user.
Banerjee Economics People might follow others The number of others con-
(1992) even if private information suming the good was taken
and individual motivations as evidence that the others
suggested doing something had access to better infor-
else. mation than oneself.
18
Research Background
__________________________________________________________________________

(Year)
Bikhchandani Economics Modelling shows that people This is based on presumed

et al. (1992) followed others, even when incomplete information and
private information and moti- rationality. The number of
vation suggested otherwise. others implementing a
measure is considered as
evidence of better informa-
tion, this leads to conformity
and 'informational cas-
cades' based on incomplete
information.
Fowler et al. Political Sci- Recent example of social Social networks

(2009) ence network methodology
Overview is taken from summaries provided by Van Campenhout and Verhestraeten

(2010), Raafat et al. (2009b, p. 421), and Rook (2006)
The history of herd behaviour clearly shows that similar ideas and explanations have
emerged in many fields, albeit with different emphases, demonstrating the interdisci-
plinary nature of the concept.
Banerjee (1992) and Bikhchandani et al. (1992) are probably most frequently cited by
economists with regards to herding (Parker and Prechter 2005). The widely spread
definition in economics, including IS, is the definition according to Banerjee (1992, p.
798), where the phenomenon is defined as the notion that “everyone does what eve-
ryone else is doing, even when their private information suggests doing something
quite different". Primary conditions which lead to the mechanism of herd behaviour
are the following: first, the presence of uncertainty about a decision, and second, the
observed actions of other individuals. In the case of both conditions, an informational
cascade, defined as individuals following the actions of others and ignoring their own
private information or signals, may occur (Banerjee 1992; Bikhchandani et al. 1992).
Accordingly, in the present dissertation, this definition is used as the underlying foun-
dation.
19
Research Background
__________________________________________________________________________
2.1.2 Applied Theories
As in the definition of herd behaviour, the underlying theories are based on diverse
approaches. Parker and Prechter (2005) give an overview and analyse in their litera-
ture review the wider diversity of theories (Table 2).
Table 2: Theories of Herding (Parker and Prechter 2005, pp. 1-2)
Theories of Herding Characteristics
Social psychological Imitation processes
Information theory Information cascades
Ethological Flocking, migrating birds, ant recruitment, etc
Econophysics Catastrophe theory, sandpile analogies, self-

organised criticality, etc.
Medical Disease and infection analogies: contagion, etc
Socionomic Unconscious, prerational
In the following, these theories are briefly described.
Social Psychological Theory of Herding
The main representative of this theory is Shiller et al. (1984). He challenged the as-
sumption of full rationality of investors (agents in economic situations). His descrip-
tion of the social dynamics in a stock market follows a combination of social enthusi-
asm, excessive optimism, and selective attention, and on waves of excessive opti-
mism and pessimism in market 'fads' (Shiller 1990; Shiller 2015). His 'fads and fash-
ions' model posits that “investors have overconfidence in a complex culture of intui-
tive judgments about expected future price changes, and an excessive willingness to
act on these judgments” (Shiller 2001, pp. 3-4). Another view, e.g. that followed by
Scharfstein and Stein (1990), is based on the assumption of 'imitation for social ad-
vantage' by assuming superior knowledge or skills, as an explanation for herding.
20
Research Background
__________________________________________________________________________
Information Theory
The most prominent representatives of this theory are: Banerjee (1992), Bikhchan-
dani et al. (1992), and Hirshleifer and Hong Teoh (2003). The essence of their theory
is based on an informational cascade. “An informational cascade occurs when it is
optimal for an individual, having observed the actions of those ahead of him, to follow
the behaviour of the preceding individual without regard to his own information"
(Bikhchandani et al. 1992, p. 994). This theory assumes utility-maximisation, and
also makes an assumption that neoclassical equilibrium theory is correct. Further-
more, it specifies a rational, conscious process where causality is exogenous. It is
one of the most influential among traditional economists (Parker and Prechter 2005).
Ethological Theory of Herding
Ethology is the study of animal behaviour and in the ethological theory, and it is used
as a source of metaphors and analogies for the herd model. The focus is on depict-
ing the animals' behaviour and its analogies in human herding behaviour, "rather than
a focus on a a single set of theoretical assumptions about the dynamics of herding"
(Parker and Prechter 2005, p. 3). As a representative of this theory, Kirman (1993)
refers to the process of recruitment observed in ant behaviour. His model does not
support the equilibrium theory of the neoclassical economy because there is no con-
vergence to a particular state. Instead, he explains that his ant recruitment model can
generate bubbles similar to the investor market; meanwhile, the traditional models
that include exogenous shocks cannot detect them (Kirman 1993).
Econophysics Theory of Herding
The econophysics theory has much in common with the information theory models of
herding (Sornette and Andersen 2002). It differs in that, econophysics attempts to
model the dynamics of the 'rational bubbles' created by herding endogenously,
whereas information theory models contain only exogenous causality (Parker and
Prechter 2005). Different versions of econophysical models describe both homoge-
neous and heterogeneous actors. The different studies differ in the question whether
the processes involved are conscious or not. Catastrophe theory, self-organised criti-
cality, and sandpile models are included in this theory (Parker and Prechter 2005).
21
Research Background
__________________________________________________________________________
Medical Model Theory of Herding
This herding theory has a long history, where market panics are described in terms of
social contagion (Kelly and O Grada 2000). Another body of literature draw its ex-
planatory power from primarily unconscious processes, which often contain the 'in-
fectiousness' of social mood (Levy and Nail 1993). Many studies that refer to the
medical model from the financial perspective define 'financial contagion' as the rapid
spread of from one market to another (Kyle and Xiong 2001).
Socionomic Theory of Herding
The socionomic theory of herding (Prechter 1999; 2001; 2003) is described by the
following characteristics: "unconscious, prerational herding behaviour that posits en-
dogenous dynamics that have evolved in homogeneous groups of humans in con-
texts of uncertainty, while eschewing the traditional economic assumptions of equilib-
rium and utility-maximization" (Parker and Prechter 2005, p. 4). It is characterised by
the ability to capture the process of decision-making under uncertainty in a way that
reflects the psychological reality of the individual's behaviour while providing a prob-
abilistic prediction of the formative path of development of the social whole.
Overall, it is evident that the literature reveals many theoretical approaches to herd-
ing. The dissertation at hand follows the economic herd model according to the in-
formation theory.
2.1.3 Established Frameworks
There is an extensive range of frameworks to describe herding and organise the dif-
ferent approaches. For example, Raafat et al. (2009b) have developed a conceptual
framework from a cognitive science perspective within which the different ap-
proaches and models can be described. It is based on the transmission mechanism
that propagates a certain thought or behaviour from one agent to another, and the
patterns of connections between agents. Additionally, they differentiate between two
major forms of transmission: automatic contagion and rational deliberation.
Hirshleifer and Hong Teoh (2003) have developed a framework that provides a com-
prehensive overview of approaches, as well as a taxonomy of the various sources of
convergence or divergence that are caused by actual interactions between individu-
als. Herding/dispersing includes any behavioural similarity/dissimilarity caused by the
interaction of individuals (Hirshleifer and Hong Teoh 2003). Their framework (Figure
22
Research Background
__________________________________________________________________________
6) describes a double hierarchy of means of convergence. Herding/dispersing is the

most comprehensive category, at the top of the hierarchy. The observational hierar-
chy is represented by the four rectangles (A, B, C, D), which describes the informa-
tional sources of herding/dispersing. The oval forms depict the payoff interaction hi-
erarchy (I, II, III), which provides a different (though not mutually exclusive) perspec-
tive on herding or dispersing (Hirshleifer and Hong Teoh 2003).
I. A. Herding / Dispersing
B. Observational Influence
C. Rational Observational Learning
D. Informational Cascades
II. Payoff and
Network
Externalities
III. Reputational
Herding and
Dispersion
Figure 6: Herding Framework according to Hirshleifer and Hong Teoh (2003, p. 28)
In the following the two hierarchies are briefly described based on the description
provided by Hirshleifer and Hong Teoh (2003, pp. 28-30):
Observational Hierarchy:
A. Herding/dispersing: observation of others can lead to dispersing instead of

herding, e.g. in the case of opposing preferences
B. Observational influence: dependence of behaviour upon the observed be-
haviour of others, or the results of their behaviour; may be imperfectly rational
23
Research Background
__________________________________________________________________________
C. Rational observational learning: observational influence resulting from ra-

tional Bayesian conclusions from information reflected in the behaviour of oth-
ers, or the results of their behaviour
D. Informational cascades: observational learning, based on the observation of
others' actions, payoffs, or conversations; thereby, the observation is so infor-
mative that an individual’s action does not depend on his own private signal
Payoff interaction hierarchy (based on the description by Hirshleifer and Hong Teoh
(2003, pp. 29-30)):
I. Herding/dispersing: as in the observational hierarchy

II. Payoff and network externalities: involving convergence or divergence of
behaviour arising from the fact that an individual’s action affects the payoffs to
others of taking that action
III. Reputational herding and dispersion: it is about convergence or divergence
of behaviour based on the attempt of one person to maintain a good reputa-
tion with another observer; such a desire for good reputation can lead to pay-
off interactions, so that III becomes a subset of II (Scharfstein and Stein 1990)
A growing body of literature has highlighted the importance of category D: the infor-
mation cascades in economics and the social sciences (Cipriani and Guarino 2009;
Sunstein 2005), which describes a state in which imitation is certain to occur. Al-
though, the mechanism is a simple form of social interaction it offers a crucial benefit:
"it allows an individual to exploit information possessed by others about the environ-
ment" (Hirshleifer and Hong Teoh 2003, p. 28). In the modern networked world with
its numerous communication systems, information cascades are predicted regardless
of the plenitude of available information and despite the common interest of the ac-
tors (Raafat et al. 2009b).
2.2 Consumer Buying Decision Process
The investigation of consumer decision-making processes has been the focus of in-
terest in consumer behaviour for over 30 years (see, e.g., Bettman (1979), Howard
and Sheth (1969)). Rapid technological change has led to a large number of new
products and a shortened product life cycle. In addition, new communication media
such as the Internet and social media have made enormous amounts of information
24
Research Background
__________________________________________________________________________
about possible options available. It will therefore continue to be important to under-

stand consumer decisions (Foscht and Swoboda 2011).
Consumer behaviour comprises a broad definition which includes the process and
activities that people undertake in searching for, selecting, buying, using, evaluating
and disposing of products and services in order to satisfy their needs and desires.
This also includes the decision-making processes that precede and follow these ac-
tions (Belch and Belch 2003). Until the mid-1960s, consumer behaviour was a rela-
tively new field, which emerged from other disciplines like economics, marketing and
behavioural science (Engel et al. 1995). Thereby, concepts from other research dis-
ciplines, such as psychology (study of the individual), social psychology (study of the
functioning of an individual in groups), and economics (study of spending patterns in
society) were incorporated. Last but not least, due to the increasing spread of the
Internet, the IS research area has also been used to study the use of new technolo-
gies and their effects (Smith and Rupp 2003).
One of the fundamental issues in consumer behaviour is the way consumers take
decisions, i.e. the decision-making process (Moon 2004). Early studies focused
mainly on buying behaviour, while studies after the 1960s included a wider range of
activities (Engel et al. 1995). Using the information processing approach, the buying
decision process of consumers can be explained (Bettman et al. 1998; Howard and
Sheth 1969). In the first step, the consumer finds the necessary information, evalu-
ates it and takes a decision. The buying decision process is shaped by the consumer
himself and by the context of the decision-making process. It may therefore vary from
person to person, from decision to decision and from context to context (Bettman et
al. 1998). It should be mentioned that in the literature both terms 'purchase' and 'buy-
ing' are used as synonyms (e.g. Howard and Sheth (1969), Engel et al. (1995));
meanwhile, in this dissertation the term 'buying' is mainly used.
25
Research Background
__________________________________________________________________________
Theories of consumer decision-making have been developed over time from various
research disciplines and have evolved steadily. Essential theories are presented in
the following.
Rational Choice Theory
The first approaches for the investigation of consumer decisions were based on ra-
tional decision theories known as economic views, which assume that individuals act
completely rational to maximise their utility in a buying situation. This view is based
on a rational decision-maker, having perfect information, with fully defined prefer-
ences that do not depend on specific options or methods and that have a clear set of
choices. The following premises are taken as a foundation (Bettman et al. 1998;
Foscht and Swoboda 2011):
 Full knowledge of the structure of one's own preferences

 Utility maximisation assumption, i.e. the assumption of rational behaviour
 Complete market transparency
 Unlimited capacity of information processing
 No temporal, material or geographical preferences
 No influence by other persons
 No influence by previous buying decisions
"This approach to studying consumer decisions, often attributed to economists and

called rational choice theory, has contributed greatly to the prediction of consumer
decisions" (Bettman et al. 1998, p. 187).
It is clear, however that such an expectation is unrealistic. In most cases, consumers

do not have access to all the information, they do not have the time for such an ex-
tensive process and they are not qualified or motivated enough to make the 'perfect'
decision. Usually they are not prepared to engage in extensive decision-making ac-
tivities and will instead settle for a 'satisfactory' decision even though it is not the op-
timal choice (Solomon 2010). Accordingly, the theory of rational choice has its limita-
tions, and it is not able to explain generally observed, less 'rational' decision-making
behaviour (Bettman et al. 1998).
As with many economic theories, the broad criticism is directed at the assumptions
and less at the often formally logical conclusions drawn. There is general agreement
26
Research Background
__________________________________________________________________________
that the approaches have relatively little explanatory power, due to assumptions
about buyer behaviour that are not very realistic. These are not characterised by:
buying according to rational premises, complete information, the abstraction of be-
havioural–scientific explanatory variables (e.g. cultural, social, psychological deter-
minants of buyer behaviour) or the reduction of marketing instruments to price
(Foscht and Swoboda 2011). Regardless of its criticism, this approach has made a
remarkable contribution to the prediction of consumer decisions, an aspect that
should not be neglected (Bettman et al. 1998).
Information Processing Theory
Researchers such as Bettman (1979) have argued that rational choice theory is in-
complete and flawed as an approach for understanding how consumers actually
make decisions. Therefore, the problem of rational choice theory described above led
to the development of a new generation of theories on consumer behaviour, which
are based on an information processing approach for buying decision making, and
endorses bounded rationality (Simon 1955). Simon (1955) developed the 'satisficing'
model, a cognitive heuristic according to which the behaviour is described as choos-
ing through the available alternatives until an acceptability threshold is met. It ac-
knowledges the notion that decision-makers are limited in their ability to process in-
formation. These limitations include limited memory and computing capabilities. In
addition, decision-makers are characterised by perceptions that are more sensitive to
change than to absolute magnitudes, and by a decreasing sensitivity to changes in
stimuli (Tversky and Kahneman 1991). Therefore, Tversky and Kahneman (1983)
and Ariely (2008), among others, have shown several reproducible ways in which
human judgements and decisions differ from the theory of rational choice. The expla-
nation in human differences in decision-making is based on heuristics (mental abbre-
viations that provide quick estimates of the possibility of uncertain events). An exam-
ple of such heuristics is the tendency to judge 'the frequency or likelihood' of an event
according to the extent to which the event 'resembles the typical case' (Tversky and
Kahneman 1983). In general, behaviour is determined by the interaction between the
characteristics of the human information processing system and the characteristics of
the task environment (Simon 1990).
The information processing approach has been developed in the field of cognitive
psychology and is based on the fact that the human being is understood as an infor-
27
Research Background
__________________________________________________________________________
mation processing system, which means that the human mind is like a computer or
an information processor — rather than merely responding to stimuli (Larsen 2015).
Miller (1956) has developed two theoretical ideas that are fundamental to the frame-
work of information processing and cognitive psychology in general. The so-called
'chunking' is the first concept and the capacity of 'short-term memory'. According to
his idea the 'short-term memory' can only hold five to nine pieces of information
(seven plus or minus two), whereby a chunk must be a meaningful unit. These can
be, for example, numbers, words, chess positions or faces of persons. The concept
of 'chunking' and the limited capacity of 'short-term memory' have become a basic
element of all subsequent theories of memory (Larsen 2015).
Atkinson and Shiffrin (1968) presented a further developed information processing

model that illustrates the theory of the human memory. They used the model to show
that the human memory can be broken into three sub-sections: 'sensory register',
'short-term memory' and 'long-term memory' (Figure 7).
Rehearsal
Encoding
External stimulus Attention
Sensory Short-Term Long-Term
Register Memory Memory
Retrieval
Forgotten Forgotten
Figure 7: Information Processing Model according to Atkinson and Shiffrin (1968)
'Sensory register' is the part of memory that receives all the information that a person
senses. These memories are usually unconscious and last for a very short amount of
time. As the senses are receiving constantly a large amount of information, the 'sen-
sory register' acts like a filter by focusing on what is important and forgetting what is
unnecessary. Only information that is considered relevant progresses into 'short-term
memory (Atkinson and Shiffrin 1968).
28
Research Background
__________________________________________________________________________
'Short-term memory' is also known as working memory, and can be described as the
part where new information is stored temporarily until it is either forgotten or trans-
ferred into long-term memory. Information is held in 'short-term memory' through
maintenance or extensive rehearsal. Maintenance refers to repetition, while elabora-
tion refers to the organisation of information (such as chunking or chronology) (Atkin-
son and Shiffrin 1968).
'Long-term memory' includes various types of information (e.g. declarative, proce-

dural, imagery), and is the part of memory that has unlimited capacity and can man-
age information indefinitely. However, how well the information is organised is crucial
when it comes to proper encoding (elaboration process in transferring to 'long-term
memory') and retrieval process (scanning memory and transferring into 'short-term
memory') (Atkinson and Shiffrin 1968).
One of the key factors for effectively coding information is to ensure that the material
is meaningful and that prior knowledge is activated. Strategies to support coding in-
clude chunking (breaking down information into manageable pieces), rehearsal, im-
ages, mnemonics, schema activation, and the level of processing (Larsen 2015).
Behavioural Theories
In addition to the theories described, 'Rational Choice Theory and 'Information Proc-
essing Theory', behavioural science approaches are also applied.
Behaviourism draws only on observable variables as a basis for explaining behav-

iour. It is based exclusively on the observation of stimuli and reactions. The behav-
iouristic research paradigm is known as the stimulus–response model, the core
statement of which is: when an existing stimulus (S) meets an organism, a certain re-
action is to be expected with a certain probability (Nelson 1975).
However, behaviourism theory ignores how an external stimulus affects the inside of
a consumer, i.e. how psychological processes take place in the organism. In con-
trast, neo-behaviourism abandons the black box view of the organism and also uses
statements about unobservable internal processes to explain behaviour (Suppes
1993). In these stimulus–organism–response (S-O-R) models, two classes of vari-
ables are distinguished: observable and intervening variables. The intervening vari-
ables structure the contents of the black box, which allows an explanation of the psy-
chological effect of individual marketing measures. However, observable phenomena
(indicators) are needed to measure the intervening variables. Their quality (objectiv-
29
Research Background
__________________________________________________________________________
ity, reliability and validity) must be critically questioned each time (Foscht and
Swoboda 2011).
2.2.3 Established Models
In principle, consumer behaviour could only be investigated using experimental ap-

proaches and without decision process models. However, models offer the advan-
tage of explaining these processes and visually representing the way in which vari-
ables and circumstances relate to each other by linking causes and effects. A model
simplifies reality, thus reducing complexity, and illustrates the phases that the con-
sumer goes through when making a buying decision, as well as post-buying behav-
iour. Hence, models facilitate the understanding of the differences in the specific de-
cision-making phases (Engel et al. 1995).
In the following, well-known and established models for the buying decision process
are presented and briefly discussed.
Grand Models
Models with a wide scope are often referred to as 'grand models' (Kassarjian 1982).
The often cited and comprehensive model of Howard and Sheth (1969) tries to ex-
plain the buying decision process for a certain brand from a range of alternative
brands. It integrates different input variables ('stimuli'), which are composed of mar-
keting activities of the company and the influences of the consumer's social environ-
ment. The model indicates that these influencing factors are internalised by the con-
sumer before affecting the decision-making process. If this information contains in-
consistencies from the consumer's point of view, e.g. if symbolic information (price,
quality, etc.) deviates, this can lead to a renewed search behaviour or increased at-
tention. The 'perception constructs' are followed by the 'learning constructs'. The
'brand knowledge' describes as a learning construct the knowledge of the existence
of the characteristics of brands. Conversely, 'decision criteria' serve to evaluate the
alternatives in consideration of the motives. The 'attitude' assigns the possibility of
'motive fulfilment' to the trademarks and depending on the perceived 'confidence' the
'intention to buy' is formed or a new search behaviour occurs. If all expectations and
wishes are fulfilled by the 'buying', the 'attitude' towards the brand is stabilised or a
perceived certainty about the correctness of the action is achieved.
30
Research Background
__________________________________________________________________________
The peculiarity of the Howard and Sheth's model at the time of publication was the
way in which different variables are linked together and their linkage is developed
during the process. However, the developmental linkages and the correlation of all
constructs make it difficult to test this model. In addition, the unobservable nature of
its constructs hinders the simplicity of measurement (Foscht and Swoboda 2011) .
Inputs
Perceptual Constructs Learning Constructs Outputs
Stimuli Display
Significative
Intention Purchase
a. Quality
to buy
b. Price
c. Distinctive
Overt search Confidence
d. Service
e. Availability
Intention
Symbolic Attitude
a. Quality Stimulus Attitude
b. Price ambiguity
c. Distinctive
d. Services Decision
e. Availability criteria Brand
knowledge
Attention
Brand
Motives knowledge
Social
a. Family Attention
b. Reference Perceptual
groups bias
c. Social class Satisfaction
Figure 8: Model of Buyer Behaviour according to Howard and Sheth (1969, p. 30)
There is much similarity between the Engel-Kollat-Blackwell’s model (Engel et al.

1995) and the model of Howard and Sheth (1969). The main difference lies in the
representation and linkage of the constructs, since the model of Engel-Kollat-
Blackwell is a phase model. It divides the buying decision into three main compo-
nents (decision, information processing and evaluation process) or, in more detail,
into several successive process phases: 'problem recognition', 'search', 'alternative
evaluation', 'purchase' and 'outcome'.
The decision-making process can be described as follows:
Triggered by a perceived need, which is initiated by activating motives and stimuli in-
fluencing the consumer, the decision-making process starts. This is followed by the
information search and processing. The information is compared with the consumer's
own convictions, opinions and behavioural intentions, whereupon it forms the basis of
31
Research Background
__________________________________________________________________________
the evaluation process of different product alternatives and decision-making proc-

esses. Individual consumer characteristics and external environmental influences,
such as cultural norms and values, play a determining role. After the decision and the
subsequent purchase, the consumer comes to a post-purchase alternative evalua-
tion, which can lead to satisfaction or dissatisfaction. This experience is retained by
the consumer and serves as a basis for future purchases (Bänsch 2017).
Input Information Decision Variables Influencing

Processing Process Decision Process
Problem
recognition
Internal Individual
Exposure Search Characteristics:
search
Motives
Values
Attention
M Alternative
Stimuli E evaluation
Marketing Comprehension M
dominated, Perception O Social Influences
other R Purchase
Yielding/ Y Culture
Acceptance reference Group
Outcomes
Situational
Retention Influences
Post-purchase
External rating
search
Dissatisfaction Satisfaction
Desinvestition
Figure 9: Model of Consumer Behaviour according to Engel et al. (1995, p. 95)
Although the large models are highly complex models with a detailed focus on the re-
lationships between the factors and the stages in the decision-making process. Nev-
ertheless, their ground-breaking work has led to further research to identify the con-
textual and individual factors that influence consumers' decision-making processes,
which should not be ignored (Bänsch 2017).
SR/SOR Models
SR and SOR models are rooted in behaviourism theory. Its research paradigm finds
expression in SR models (Figure 10), with the assumption of the following patterns:
When a certain stimulus (S), e.g. an attractively presented product, meets an organ-
ism, the response (R), e.g. the purchase of a product, is observed. The behaviour of
the consumer is only examined on the basis of the input and output variables of the
black box. Hence, the psychological processes in the organism are excluded. In this
32
Research Background
__________________________________________________________________________
way, what makes a consumer buy a product and others not buy it cannot be ex-
plained, even those both consumers were confronted with the same stimuli.
Thus, the SR models are not sufficient to explain such complex processes as con-
sumer buying behaviour (Foscht and Swoboda 2011).
Response (R)
Stimuli (S) Black Box
Environment
Marketing Budget
Purchase
Target Group
Marketing Mix (Buyer‘s Mind) Quantity of sales
Product, Price, Place, Promotion
Competition
Unobservable behaviour/
observable observable
process and conditions
Figure 10: SR-Model - Schematic Diagram (Foscht and Swoboda 2011)
The observation of external behaviour (R) must be supplemented by the internal be-
haviour analysis. Thus, the neo-behaviouristic SOR models abandon the black box
approach and consider the 'inner behaviour'. The SR model is extended and two
variable categories are applied: observable variables and intervening variables. In
the SOR model, these two variables are linked together. The observable variables
are the stimuli (S) and the observable reactions (R). The intervening variables are
employed to describe and specify the black box. The system of intervening variables
consists of: activating processes (e.g. emotions, motivation, attitudes) and cognitive
processes (e.g. perception, learning, memory). According to the current state of
knowledge, the intervening variables form the basis for research on consumer pur-
chase behaviour (Kroeber-Riel et al. 2009).
33
Research Background
__________________________________________________________________________
Stimuli (S) Organism (O) Response (R)
Marketing mix Emotional responses Cognitive valuation

• Product
• Price • Emotion • Perception • Brand
• Place • Motivation • Learning selection
• Promotion • Memory • Quantity of
sales
External Attitude • Amount of
influences money
• Cultural
• Social
• Economic
• Technological Involvment Reference group Culture
Directly observable Indirect observable Directly observable
Figure 11: SOR Model according to Kroeber-Riel et al. (2009, p. 51)
Buying Decision Process Model
The buying process is understood as the whole process, from the emergence of a
certain need, the various types of decision-making with information acquisition and
processing, product selection, shopping behaviour, use and eventual disposal of the
product, including the increase in consumer product experience (Kuß and Tomczak
2007). A widespread tool to gain better understanding of customers and their behav-
iour is a five-stage buying decision process model, comprising need recognition, in-
formation search, evaluation of alternatives, buying decision and post-buying behav-
iour (Armstrong et al. 2009).
According to Foscht and Swoboda (2011), a holistic view makes sense if at least
three phases are distinguished (Figure 12). They propose the phases pre-buying
phase, buying phase and post-buying phase. A differentiation into phases is advan-
tageous, because the process of a transaction as well as the comprehensive cus-
tomer relationship can be analysed. This results in the possibility of a phase-specific
analysis of the buying behaviour and the phase-specific use of marketing instruments
based on it. In an ideal-typical process, the recognition of a need by the consumer
begins with the 'pre-buying phase'. Depending on the buying decision, a more or less
intensive search for information then begins, whereupon the evaluation of considered
alternatives is triggered. During the 'buying phase', after the identification of the opti-
mal alternatives and the determination of the satisfactory ones, a buying intention is
formed and the buying is processed. The 'post-buying' phase, which is decisive for
34
Research Background
__________________________________________________________________________
the further course of the customer relationship, begins with the consumption or use of
the product. Under certain circumstances, a complaint is made and finally the dis-
posal is carried out. Ideally, this phase ends with satisfaction, which leads to the be-
ginning of a new 'pre-buying phase'.
Structuring the buying process in three phases is an ideal-typical division, especially

with regard to the sub-processes, which in reality are not always available in the de-
picted form. However, consumers do not necessarily pass through all stages, e.g. in
routine purchases, consumers often skip or reverse some of these stages (Armstrong
et al. 2009). However, the three-phase division proves its worth in so far as it allows
the differentiated explanation of different behaviour patterns and the explanation of
the effect of certain marketing measures (Foscht and Swoboda 2011).
Pre-Buying Phase Buying Phase Post-Buying Phase
Internal
search for Complaint
Information
Identification
Problem Purchase Consumption
Evaluation of
recognition decision / utilisation
alternatives Removal
External
Development
search for
of intention
information
Evaluation
Figure 12: The three Phases of the Buying Decision Process according to Foscht and
Swoboda (2011, p. 186)
2.3 Social Media
The phenomenon of social media has only developed within the last two decades. It
is covered in a wide range of research fields (e.g. marketing, economics, IS), aca-
demics and practitioners have studied and examined the many sides of social media,
whereby they have largely focused on explaining new concepts that makeup its
foundations, and further, explored the impact on consumer behaviour by integrating it
in company’s strategy (Paquette 2013). Nevertheless, social media research today
remains rather fragmented and does not yet facilitate a general understanding of the
phenomenon (Kapoor et al. 2018; Richter et al. 2011). This chapter introduces and
defines key terms and theoretical concepts used in social media.
35
Research Background
__________________________________________________________________________
"There is not one commonly accepted scholarly (or colloquial) definition of social me-
dia. In part, this lack of integration is due to the fast-changing nature of social media"
(Bayer et al. 2020, p. 473). Nevertheless, the understanding of social media of Kap-
lan and Haenlein (2010, p. 61) is still widely spread today as “a group of Internet-
based applications that build on the ideological and technological foundations of Web
2.0, and allow the creation and exchange of user-generated-content”.
Researchers have conducted literature reviews with different emphasis and perspec-
tives to explore the subject, e.g. Richter et al. (2011) concentrate their review on the
implications for enterprise 2.0, the goal of the literature review of Pérez-Latre et al.
(2012) is to identify interesting research areas in the field, Paquette (2013) rather fo-
cus on social media as a marketing tool, Ngai et al. (2015) focus in their literature re-
view on social media research, Alves et al. (2016) concentrate on social media mar-
keting, and Kapoor et al. (2018) probably offer the most comprehensive literature re-
view from the IS field perspective. "In studying the existing literature on social media,
it becomes apparent that the authors in this field have not focussed on defining social
media" (Kapoor et al. 2018, p. 536). They further state that of all the studies included
in their review, only a handful of studies have come close to defining or clarifying the
concept of social media.
Table 3 provides a selective overview of social media definitions from IS literature.
Table 3: A Sampling of Social Media Definitions from IS Perspective (Kapoor et al.

2018)
Author Definitions
(Year)
Tang et al. User-generated media, which is a source of "online information created,

(2012, p. 44) initiated, circulated, and used by consumers intent on educating each
other about products, brands, services, personalities, and issues".
Xu and Zhang Social media is a set of interned-based technologies/applications, which

(2013) are aimed at promoting the creation, modification, update and exchange
of user-generated content, while establishing new links between the con-
tent creators themselves.
36
Research Background
__________________________________________________________________________
Author Definitions
(Year)
Bharati et al. A technology "not focussed on transactions but on collaboration and

(2014, p. 258) communication across groups both inside and outside the firm".
Miranda et al. Social media is "mainly conceived of as a medium wherein ordinary peo-
(2016, p. 304) ple in ordinary social networks (as opposed to professional journalists)
can create user-generated news".
Kapoor et al. "Social media is made up of various user-driven platforms that facilitate
(2018, p. 536) diffusion of compelling content, dialogue creation, and communication to
a broader audience. It is essentially a digital space created by the people
and for the people, and provides an environment that is conducive for in-
teractions and networking to occur at different levels (for instance, per-
sonal, professional, business, marketing, political, and societal)".
All of the above descriptions, including Kaplan and Haenlein's (2010) definition,
clearly view social media as a means of communication tool supported by Internet-
based technologies for disseminating information. Most of them recognise the high
concentration of user-generated content on such platforms.
There are many papers using other terms interchangeably with social media, e.g. 'vir-
tual communities', 'online communities', 'blogs', 'Web 2.0', 'social networking sites',
and 'social computing' (Ngai et al. 2015). Thereby, social networking sites in particu-
lar are used as synonym. However, "social media is the environment in which social
networking takes place and has altered the way in which consumers gather informa-
tion and make buying decisions" (Paquette 2013, p. 3).
When referring to social media applications like blogs and microblogs (e.g. Twitter),
social networking sites (e.g. Facebook, LinkedIn), virtual worlds (e.g. Second Life),
collaborative projects (e.g. Wikipedia), content community sites (e.g., YouTube, Pin-
terest), and sites dedicated to feedback (e.g. online forums, review- and rating plat-
forms) are included (Maas et al. 2014; Mangold and Faulds 2009).
Social media enable firms to communicate with their customers and also allows cus-
tomers to communicate with each other (Mangold and Faulds 2009). Furthermore,
these applications are driven by user-generated content and are highly influential in a
37
Research Background
__________________________________________________________________________
variety of areas, from buying and selling behaviour, entrepreneurship, political ques-
tions to risk capitalism (Greenwood and Gopal 2015).
A large number of theories and models are used in the extant social media research.
Thereby, three groups of theories can be obtained: 'personal behaviour theories', 'so-
cial behaviour theories', and 'mass communication theories' (Ngai et al. 2015). Each
group is briefly described below.
Personal Behaviour Theories
The first group 'personal behaviour theories' of accepted theories in social media re-
search aims to explain human behaviour at the level of the individual. The group of
includes 15 theories (Table 4), and several of the most important theories are briefly
explained (these are highlighted in the overview).
Table 4: Overview of Personal Behaviour Theories (Ngai et al. 2015)
Theory Group Assigned Theories
Personal Behaviour  Attribution theory

 Elaboration likelihood model
 Existence, relatedness, growth theory
 Expectation and disconfirmation paradigm
 Goal-directed behaviour model
 Hofstede’s theory of cultural difference
 Personality traits
 Psychological choice model
 Risk perception theory
 Social cognitive theory
 Switching behaviour
 Task-technology fit model
 Technology acceptance model (TAM)
 Theory of reasoned action (TRA)
 Theory of planned behaviour (TPB)
38
Research Background
__________________________________________________________________________
'Personality traits' are often regarded as the fundamental theories to explain traits
that influence the behaviour of individuals. Psychologists Gordon Allport and Henry
Odbert (1936) laid the foundation for the personality traits used today. They took
about 4,500 terms from Webster's New International Dictionary that described behav-
ioural types or personality traits. Many of these terms were grouped into overarching
factors, so that later work concentrated on creating more concise lists of characteris-
tics. McCrae and Costa (2008) developed five-factor model (Big Five) of personality
traits as openness, conscientiousness, extraversion, tolerability, and neuroticism.
The Big Five were later proven by a large number of studies and are today interna-
tionally considered the universal standard model in personality research (Asendorpf
and Neyer 2012).
Davis (1989) developed the 'technology acceptance model' (TAM). TAM postulates
that the perceived usefulness and perceived ease of use is used to study people's at-
titudes towards adoption, using new technologies. Perceived usefulness is also con-
sidered to be directly influenced by perceived usability (Larsen 2015). Many re-
searchers have used the model to investigate similar scenarios in relation to different
social media technologies, e.g. Casaló et al. (2011), Evans et al. (2014), and Wirtz
and Göttel (2016).
The 'theory of reasoned action' (TRA) was developed by Ajzen and Fishbein (1973;
1977). TRA states that individual behaviour is driven by behavioural intentions, where
behavioural intentions are a function of an individual's attitude to behaviour and the
subjective norms surrounding the execution of the behaviour. Attitudes toward be-
haviour are defined as the individual's positive or negative feelings about the execu-
tion of behaviour (Larsen 2015). This theory, which is similar to the situation where
people voluntarily participate and engage in social media activities, is often used in
social media research (see, e.g., Kim et al. (2015b), and Hsu and Lin (2008)).
TRA was subsequently developed by Ajzen and Kuhl (1985) into 'theory of planned
behaviour' (TPB). Since TRA brought some limitations, such as a considerable risk of
confusion between attitudes and standards or the assumption that one acts freely
and without restriction, TPB is trying to remove this restriction. Thus TPB suggests
that perceived behavioural control is used to mitigate the effects of subjective norms
and attitudes on behaviour (Larsen 2015). With regards to social media, TPB is used
to predict users’ behaviour from intention to action (Baker and White 2010).
39
Research Background
__________________________________________________________________________
Social Behaviour Theories
The second group of theories: 'social behaviour theories' includes 13 assigned theo-
ries (see
Table 5) that focus on individuals’ behaviour toward social media in a social context.
Table 5: Overview of Social Behaviour Theories (Ngai et al. 2015)
Social Behaviour  Cognitive map

 Effectuation process
 Involvement theory
 Justice theory
 Social capital theory
 Social exchange theory
 Social identity theory
 Social influence theory
 Social interaction theory
 Social loafing theory
 Social network analysis
 Social power
 Social ties
Social aspects theory is a collective term that comprises: 'social capital theory', 'so-
cial identity theory', social influence theory', 'social interaction theory', and 'social ties
theory'. These theories have been widely used to investigate the attitudes, intentions
and actions of users in relation to the adoption or use of social media (Ngai et al.
2015).
Latané et al. (1979) introduced the term 'social loafing theory' which reflect the fact
that people make less effort in groups than when they are alone. For example, Shiue
et al. (2010combined this theory with social ties and social media to suggest that the
degree of individual contribution may be minimal even though social media is based
on collective effort.
'Social power' is based on the five power bases (reward, coercion, legitimacy, referee
and expert power) developed by French et al. (1959). In the context of social media,
40
Research Background
__________________________________________________________________________
Wei (2009) discuss knowledge production of social media as critical indicator of so-
cial power.
Mass Communication Theories
The last group of applied theories relates to 'mass communication'. According to that,
mass communication exerts a considerable influence on the behaviour of people.
Three types of mass communication are assigned to this group. Two essential ones
are briefly explained below.
Table 6: Overview of Mass Behaviour Theories (Ngai et al. 2015)
Mass Communication  Media richness theory

 Para-social interaction (PSI)
 Uses and gratifications theory (UGT)
'Para-social Interaction' (PSI) theory, originally comes from the television and film in-
dustry to study the effects of celebrities on consumer behaviour. Eighmey and
McCord (1998) developed it further in their study of website visitation rates, linking
PSI to websites and their visitors. In the context of social media, it is utilised to inves-
tigate user behaviour in relation to brand attitudes and buying intentions (Colliander
and Dahlén 2011).
Another theory assigned to 'mass communication' is 'uses and gratifications theory',

which tries to understand customers' behaviour (Eighmey and McCord 1998). With
regards to social media UGT particularly investigates why people use social media,
and how it satisfies customers’ needs (e.g. Chen (2011), and Whiting and Williams
(2013)).
2.3.3 Established Frameworks
In the following, well-known and established models for social media are presented
and briefly explained.
Classification Framework of Social Media
Kaplan and Haenlein (2010) developed a classification scheme widely applied and
cited to this day . It is based on a set of theories in the field of media research (social
41
Research Background
__________________________________________________________________________
presence, media richness) and social processes (self-presentation, self-disclosure),

the two key elements of social media. This framework allows to classify the different
social media types.
According to Short et al. (1976), 'social presence' is defined as the acoustic, visual,
and physical contact that can be achieved; depending on the media, the degree of
social presence can differ. It can be influenced by the intimacy (interpersonal vs. me-
diated) and immediacy (asynchronous vs. synchronous) of the medium. The higher
the social presence, the greater the social influence that the communication partners
have on the behaviour of the other.
The second applied theory is 'media richness' according to Daft and Lengel (1986).
The theory is based on the assumption that the goal of all communication is to re-
solve ambiguity and reduce uncertainty. According to this, media differ in the amount
of information (richness) they can transmit in a given time interval, and that therefore
some media are more effective than others at resolving ambiguity and reducing un-
certainty.
Kaplan and Haenlein (2010) have applied these theories to the context of social me-
dia, and assume that a first classification of social media types can be made drawn
on the degree of 'social presence' it provides and the 'media richness' of the medium.
Furthermore, the concept of 'self-presentation' is applied. According to this, in any

kind of social interaction, people have the desire to control the impressions other
people have of them (Goffman 1959). This is done primarily with the aim of influenc-
ing others to obtain rewards and to create an image that is consistent with personal
identity.
'Self-presentation' usually takes place through 'self-disclosure', i.e. through the con-
scious or unconscious disclosure of personal information (e.g. thoughts, feelings,
preferences, dislikes) that is consistent with the image one wishes to convey (Kaplan
and Haenlein 2010). "More broadly, profiles help to reveal how self-presentation
functions in an environment in which, contrary to face-to-face interaction, versions of
the self can endure over time" (Bayer et al. 2020, p. 478). When applied to the con-
text of social media, Kaplan and Haenlein (2010) believes that a second classification
can be made based on the degree of self-disclosure and the type of self-
presentation.
42
Research Background
__________________________________________________________________________
The combination of the previously described dimensions leads to a classification

scheme of social media, which is visualised in Figure 13.
High
Self-presentation /
Low High
Self-disclosure
Low
Social presence / Media richness
Figure 13: Classification Scheme of Social Media according to Kaplan and Haenlein (2010)
S-O-R Framework for Social Media
"Social media are fundamentally different from any traditional or other online media
because of their social network structure and egalitarian nature. These differences
require a distinct measurement approach as a prerequisite for proper analysis and
subsequent management" (Peters et al. 2013, p. 281).
Therefore, Peters et al. (2013) have developed a holistic framework covering the
most important elements of social media, based on theories from marketing, psy-
chology and sociology. They draw on the logic of the S-O-R paradigm (see section
2.2.3). Thereby, the 'stimuli' part represents common marketing instruments, such as
advertising or pricing, social media stands for the 'organism', and 'response' covers
marketing outcomes like success metrics (e.g. customer lifetime value, brand aware-
ness).
The four cornerstones of social media for the S-O-R framework are 'motives', 'con-
tent', 'network structure', and 'social interactions'. First, the communication in social
media is driven by 'motives'. Second, individuals communicate with each other and
produce user-generated 'content'. Third, the 'network structure' (combination of all
dyadic bonds) forms the environment for each individual as well as for the social me-
43
Research Background
__________________________________________________________________________
dium as a whole. Fourth, during communication, content is not only produced, but
also modified, shared or simply consumed. Thus, actors share different types of 'so-
cial interactions' with each other and can assume different social roles (Peters et al.
2013). Within each social medium, all four elements interact continuously, changing
and strengthening each other as in a living organism. The process of message al-
teration in a social network is highly non-linear, as individuals can participate in sev-
eral social media simultaneously (Hennig-Thurau et al. 2010). The combination of all
four elements (motives, content, network structure, and social interactions) with their
different aspects results in the framework shown in Figure 14.
Stimulus Marketing Input
Social Value
S
Motives
O
Cultural Intellectual
C Value Value
I
A
L
Organism
Expressing
M Sharing
E Social Size
Quality
D Interactions Connections
Networking Network
I Content
Gaming
A Valence
Structure
Volume Distribution
Segmentation
Response Marketing Outcomes
Figure 14: S-O-R Framework for Social Media according to Peters et al. (2013)
44
Research Design
__________________________________________________________________________
3 Research Design
Chapter overview: This chapter introduces the research design of the present dis-
sertation. First, in section 3.1 the academic discipline of information systems re-
search is introduced with its method paradigms, and underlying philosophical per-
spectives. Then, the applied methods in the dissertation are presented and described
in section 3.2. Finally, the research strategy is outlined in section 3.3, containing the
research questions, research paradigm, and research approach.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 15: Dissertation Overview - Chapter 3: Research Design
45
Research Design
__________________________________________________________________________
3.1 IS Research Paradigms and Methods
Research paradigms and methods are a characteristic element of a scientific disci-

pline. As a set of instruments, their spectrum defines the researcher's framework of
action and is closely linked to the researcher's scientific self-conception (Schreiner et
al. 2015).
However, it is also possible that there are different orientations within a discipline, as
it is the case in Wirtschaftsinformatik (WI)/information systems (IS). While the (Ger-
man-speaking) WI has its focus more in the field of constructive methods (e.g. the
creation and evaluation of prototypes), the design of solutions for practically oriented
problems in the context of information and communication technology (ICT), as well
as in the integrated consideration of economic and ICT-related aspects (Mertens
2010; Österle et al. 2011). In contrast, the Anglo-American counterpart of IS research
is associated primarily with research in behavioural studies focusing on the applica-
tion of information technology (IT) within an organisation to meet organisational and
managerial needs (Bankhofer et al. 2008).
Nevertheless, there has been an increase in the openness of the WI for behavioural
science methods (Ayanso et al. 2007). At the same time, design-oriented research
also seems to be establishing itself in the IS community (Gregor and Hevner 2013).
Further, the use of multi-method designs (e.g. quantitative and qualitative methods,
linkage of different method classes, linkage of behavioural science and design-
oriented methods) seems to be established, particularly in WI research. This is repre-
sentative of a certain degree of openness, which appreciates methodological diver-
sity (Schreiner et al. 2015).
For the present dissertation, the differentiation of WI and IS research focus plays a
subordinate role. Therefore, following this chapter, the research-related discipline is
referred to as IS and refers to both WI and IS.
3.1.1 IS Research Paradigm
According to Kuhn (1962), paradigms are concrete problem solutions accepted by

the professional community. Paradigms enable researchers to start a project with
certain assumptions about, how they will learn and what they will learn during their
investigation, and thus to establish a knowledge claim (Creswell 2009). From a phi-
losophical perspective, a paradigm is primarily characterised by the following compo-
nents (Khazanchi and Munkvold 2003):
46
Research Design
__________________________________________________________________________
 Ontology: Theory of existence

 Epistemology: Theory of knowledge
 Methodology: Procedure of knowledge generation
The number of research paradigms has multiplied over the last decades, in the IS lit-
erature three research paradigms are widely embraced: the critical research, the in-
terpretive research, and the positivist research. They are based on different beliefs
regarding the sources and development of knowledge, the nature of physical and so-
cial reality and the relationship between theory and practice (Klein and Myers 1999;
Orlikowski and Baroudi 1991).
Critical Research
"IS research can be classified as critical if the main task is seen as being one of so-
cial critique, whereby the restrictive and alienating conditions of the status quo are
brought to light" (Klein and Myers 1999, p. 69). It is assumed that people can act
consciously to change their social and economic conditions. However, their ability to
improve their conditions is limited by various forms of social, cultural and political
power as well as by natural laws and resource restrictions. Thereby, researchers take
a critical viewpoint and try to understand the power structures behind the accepted
interpretations in order to propose improvements in the conditions of human exis-
tence (Myers and Klein 2011).
Interpretive Research
"Interpretive studies assume that people create and associate their own subjective
and intersubjective meanings as they interact with the world around them" (Or-
likowski and Baroudi 1991, p. 5). Interpretive research does not prescribe dependent
and independent variables, but concentrates on the complexity of human sense in a
given situation (Kaplan and Maxwell 2005). It attempts to understand phenomena
through the meanings assigned to them by people (Orlikowski and Baroudi 1991). In-
terpretive research is "aimed at producing an understanding of the context of the in-
formation system, and the process whereby the information system influences and is
influenced by the context" (Walsham 1993, p. 4).
Positivist Research
"Positivist studies are premised on the existence of a priori fixed relationships within
phenomena which are typically investigated with structured instrumentation" (Or-
47
Research Design
__________________________________________________________________________
likowski and Baroudi 1991, p. 5). In general, in an effort to increase predictive under-
standing of phenomena, positivists attempt primarily to test theory. In the case of evi-
dence for formal statements, quantifiable measurements of variables, testing hy-
potheses or induction of inference on a phenomenon gave, the IS research can be
classified as positivist (Myers and Klein 2011).
The latter perspective dominated IS research in the 1990s (Orlikowski and Baroudi
1991). Meanwhile, the application of epistemological viewpoints has become more
balanced and different paradigms and the diversity of epistemological perspectives
are openly accepted (Keutel et al. 2014; Lee and Baskerville 2012).
3.1.2 IS Research Methods
In order to capture the spectrum of methods used by WI and thus to concretise the
hitherto rather unclear self-perception, Wilde and Hess (2007) carried out an empiri-
cal study. This confirmed the presumed design-oriented focus of the WI. However, it
also showed that the actively applied methods are much more diverse and that be-
havioural science instruments are also used. Moreover, a trend towards a more for-
malised research process as well as towards empirical studies could be identified
(Schreiner et al. 2015). Since the database on which this study is based ends in
2006, Schreiner et al. (2015) updated the original analysis of Wilde and Hess (2007)
on the applied research methods in the observation period 2007 to 2012, and
showed that the WI community has continued to apply a broad range of research
methods.
According to Wilde and Hess (2007) and Schreiner et al. (2015), a two-dimensional
portfolio representation is used to put the individual methods of the spectrum into re-
lation to each other (Figure 16). The first axis is the ‘level of formalisation’ of a
method. According to this, the more precisely a method specifies the cognitive proc-
ess and its individual steps, the more formalised the underlying system of rules.
Thus, the 'level of formalisation' is a spectrum of lower to higher formalised methods,
which can be graded according to the degree of concretisation of the respective
method rule system. The second axis 'paradigm' is based on the idea of a spectrum
that indicates the extent to which a method relies on elements of behavioural science
or constructive research methods. On the one hand, the analysis of social reality with
regard to the provision and use of IT solutions is at the centre of the behavioural sci-
ence paradigm. On the other hand, the design science paradigm focuses on the de-
48
Research Design
__________________________________________________________________________
velopment or enhancement of useful IT solutions by creating and evaluating various

artefacts in the form of concepts, models, development methods, tools, modelling
languages or systems.
The depicted three research method clusters are: design-oriented methods, empirical
methods, and formal-analytical methods. Thereby, constructive research methods
dominated the research published in the German journal Wirtschaftsinformatik. Nev-
ertheless, as stated earlier, the authors point to a trend of the portion of behavioural
studies increasing. A short introduction to the different methods can be found in Ap-
pendix A.1.
There is also a dominance of high versus low formalised methods (75% to 25%). Fur-
thermore, a look at the portfolio shows that almost half of it (49%) is made up of de-
sign-oriented and 15% of formal-analytical methods, while empirical methods are
clearly represented with 26% (Figure 16).
Laboratory Formal Formal

Experiment Analytical Deductive
Field Methods
Experiment (15%) Simulation
Quantitative
High Cross-Sectional
Reference
Empirical Methods
Level of Modelling
(26%) Conceptual
Formali- Deductive
sation Grounded
Theory
Prototyping
Qualitative
Low Cross-Sectional Case
Design Oriented
Study
Action Methods
Research (49%)
Argumentative
Deductive
n=108
Behavioural Constructive
Paradigm
Figure 16: Method Profile of WI in the years 2007-2012 according to
Schreiner et al. (2015, p. 9) and Wilde and Hess (2007, p. 184)
49
Research Design
__________________________________________________________________________
3.2 Applied Research Methodologies
As shown in the previous section, quantitative/qualitative cross-sectional methods

are an integral part of the IS method portfolio and are associated with the behavioural
research paradigm (cf. section 3.1.2).
These two methods combine survey techniques such as questionnaires, interviews,

the Delphi method, content analysis etc. into two aggregates. They comprise a single
survey of several individuals, which is then coded and evaluated quantitatively or
qualitatively. The result is a cross-sectional picture across the sample participants,
which usually allows conclusions to be drawn about the population (Schreiner et al.
2015). These applied methods are described in the subsequent sections.
3.2.1 Qualitative Methods
Qualitative research has been developed in the social sciences in an effort to enable
researchers to study social and cultural phenomena. "Qualitative approaches to data
collection, analysis, interpretation, and report writing differ from the traditional, quanti-
tative approaches" (Creswell 2009, p. 17). The basis for qualitative procedures are
for example: purposeful sampling, the collection of open data, participant observa-
tion, interviews, questionnaires, documents, the analysis of text or pictures, the pres-
entation of information in figures and tables, and the personal interpretation of the re-
sults (Creswell 2009). Further, the quality criteria of quantitative research (objectivity,
reliability, validity) cannot simply be adopted. Qualitative validity is understood to in-
dicate that researchers check the accuracy of results using specific procedures, while
qualitative reliability indicates that the researcher's approach is consistent across dif-
ferent researchers and different projects (Gibbs 2007). Additionally, validity is one of
the strengths of qualitative research; it is based on determining whether the results
are accurate from the perspective of the researcher, the participant or the readers of
a report (Creswell and Miller 2000). Nevertheless, the researcher should actively in-
corporate validity strategies into their work, such as triangulation of different data
source, 'thick' descriptions to convey findings, or presenting negative or discrepant in-
formation. Reliability procedures in qualitative research recommend documenting as
many steps in the procedure as possible, as well as reviewing developed codes and
comparing results by different researchers independently derived from each other.
The term qualitative generalisation is only used in a limited way in qualitative re-
search, because the intention of this form of investigation is not to generalise the re-
50
Research Design
__________________________________________________________________________
sults to persons, places or facilities. Particularity rather than generalisability is the dis-
tinguishing characteristic of qualitative research (Greene and Caracelli 1997).
3.2.1.1 Non-reactive Data Collection on the Internet
"Online research methods have come of age as the permeation of everyday life by in-
formation and communication technologies has grown ever more ubiquitous" (Lee et
al. 2017, p. 2). The connotation online research is understood as the Internet as the
object and method, with strong methodological reference, academic, practical and
commercial-oriented (Welker and Wünsch 2010). Not only for questionnaires the
Internet offers new possibilities, e.g. by using email, chat rooms, blogs, and social
media. In addition, the possibility is offered to observe human unaffected conversa-
tions and use a high amount of accessible non-reactive and unobtrusive data
(Janetzko 2017). The availability of the material in a digital form facilitated the access
to the data. The generated data allows rich insights into individuals’ perspectives, in-
terpretations and constructions of meaning and thus is also less affected by issues
relating to sample representativeness (Hewson 2017). Observation without the inter-
vention of the researcher (non-participating) and analysis of archival records, log-
files, conversations, publicly accessible texts, and user interactions are the main
types of non-reactive data collection on the Internet. The defining criterion for non-
reactive data collection is not a feature of the method as such, but non-awareness of
the data collection process on the part of the subjects under investigation (Janetzko
2017). A challenge that such data brings with it is of course the sample itself, it can
only ever be a so-called 'snap shot'. The representativeness of the total population
cannot be given (Rasmussen 2017).
In using online information, researchers have to adopt a new way of observation by
watching texts on a screen, adapt recording procedures and strengthening their skills
in interpreting textual data (Creswell and Poth 2016). Further, researchers are likely
to encounter ethical issues in Internet-related research (Eynon et al. 2017). Conse-
quently, researchers are responsible for ensuring the confidentiality of data at all
stages of the process, regardless of whether data collection is participating or non-
participating, in the transmission and storage of data (Nosek et al. 2002).
Observational protocols as recording procedure method are used for recording in-
formation while observing. Depending on the field, the researcher must first develop
a scheme tailored to the field to structure descriptive notes. The following elements
can be part of such an observation protocol: impressions, portraits of the participants,
51
Research Design
__________________________________________________________________________
a reconstruction of dialogue, appearance, general description of the atmosphere,

place, and the date of the field setting where the observation takes place (Creswell
2009; Flick et al. 2012).
3.2.1.2 Expert Interviews
Expert interviews belong to the qualitative interviews, i.e. the researcher conducts
personal interviews with the participants, interviews the participants by telephone, or
conducts focus group interviews. These interviews are unstructured or semi-
structured interviews with generally open questions aimed at eliciting the views and
opinions of participants (Creswell 2009). In general, experts have exclusive knowl-
edge and expert interviews offer the advantage of quickly achieving good results.
The more important aspect, however, is the use of experts as 'crystallization points'
for specific knowledge, which becomes effective in practice and thus guides action
for other actors, making this a valuable research method (Bogner et al. 2009; Flick et
al. 2012).
According to Bogner et al. (2014), when choosing expert interviews as research

method, it must first be clarified which knowledge is of interest, as different dimen-
sions of expert knowledge can be accessed, namely technical, process and interpre-
tative knowledge. Technical knowledge is defined as data, facts, 'relevant informa-
tion', facts. Respondents have privileged access to certain information that research-
ers do not have. Process knowledge includes insight into action flows, interactions,
organisational structures, events etc. in which the respondents are or were involved.
Accordingly, process knowledge is less specialist knowledge in the narrower sense,
but rather a form of experiential knowledge. Finally, the third form of knowledge, the
interpretive knowledge, contains the subjective relevance, views, interpretations,
sense drafts and explanatory patterns of the experts. Usually all dimensions are cov-
ered in expert interviews (Bogner et al. 2014).
The preparation of the data collection procedure also includes the development of
the interview questions. Therefore, a multi-step procedure is recommended that in-
cludes the compilation of the research questions and existing literature. Then the de-
velopment of a general catalogue of questions, and finally the translation into inter-
view questions, which are then tested in a pre-test (Gläser and Laudel 2010).
Expert interviews are generally audio recordings, for which consent of the inter-
viewee must be obtained and the anonymisation of the content ensured. Usually this
52
Research Design
__________________________________________________________________________
audio material is then written down (complete transcription) to serve as a basis for
further evaluation. For some systematising expert interviews it may be sufficient to
create interview transcripts based on the recording (Bogner et al. 2014).
3.2.1.3 Qualitative Content Analysis
Content analysis has successively proven to be a useful tool for systematic text
analysis in the course of the 20th century, and has emerged as an established em-
pirical research method (Mayring 2000). Initially, the content analysis was limited to
the quantitative evaluation, i.e. simple frequency analyses of certain text components
or words. The description of the individual analysis steps as well as the systematic
and theory-based approach make the results intersubjectively comprehensible (Früh
2011; Mayring 2015). Qualitative content analysis is applied in various scientific dis-
ciplines, and also recognised and established as a qualitative evaluation tool in IS re-
search (Harwood and Garry 2003; Myers 1997).
The focus of the content-analytical investigation is the formation of a category sys-

tem. Thereby the coding of the data material is central to the formation of the cate-
gory system. This generally involves the assignment of categories to relevant text
passages or the classification of text characteristics (Kuckartz 2007). Thus, coding is
a procedure to identify relevant text passages and mark them with the appropriate
codes. The codes are usually identical with the individual categories or subcatego-
ries. The categories can be developed either directly and without theoretical prelimi-
nary considerations from the data material (inductive procedure), whose categories
describe the object of investigation as completely, distinctly and exclusively as possi-
ble. Or the structuring of the material (deductive procedure) on the basis of a previ-
ously defined category system (Früh 2011; Mayring 2015). The category system is
divided into main and subcategories. In practice however, both methods are often
combined (Kuckartz 2014).
In the literature, a number of procedural instructions for rule-based content analysis

of qualitative data can be found, see e.g. Früh (2011), Krippendorff (2004), Kuckartz
(2014), or Mayring (2015). Thereby, the developed analytical method by Mayring
(2015) is a frequently used model, especially in German-language research (Schreier
2014). Basically, Mayring (2015) distinguishes between three basic forms of qualita-
tive content analysis:
53
Research Design
__________________________________________________________________________
 Summary: The material is reduced and the essential contents are preserved.
The inductively developed category system is thus a representation of the
base material.
 Structuring: The material is structured using pre-defined order criteria ana-
lysed. Thus, an estimation of the material is based on the deductively defined
category system.
 Explication: Additional material is added for individual, questionable text com-
ponents to broaden the understanding.
The following central requirements apply to all procedures, which guarantee an inter-
subjective examination (Mayring 2015):
 The situation of origin and the context of the material are specified
 The text analysis orientates adhere to pre-defined rules or a process model
 The current state of research is included in the analysis since structuring theo-
ries or theoretical concepts are used
 The results of the content analysis are verified by quality criteria regarding re-
liability and validity.
Depending on the object of investigation, a content-analytical process model is to be

set up, taking into account the aforementioned requirements (Mayring 2015). The
procedure model defines the individual analysis steps in order to ensure intersubjec-
tive reproducibility.
The following procedure model (inductive procedure) is used in Study II (chapter 5)

and Study III (chapter 6).
First, the research question was the starting point of the evaluation and provided the
direction and selection criterion of the analysis; the abstraction level was also de-
fined.
In the next step, the unit of the analysis of the coding process was specified. Based
on the works by Mayring (2015) and Morris (1994), clear meaningful elements in the
text were defined as code units.
In the third step, the research material was reviewed to inductively derive first catego-
ries from the material. These first categories were coded when the text segments re-
vealed insights regarding the research question. The categories were built as close
54
Research Design
__________________________________________________________________________
as possible from the text material and respective text passages assigned to the cate-
gories.
In the fourth step, after about 30% of the material was processed, a revision of the
categories was conducted to check whether the categories came close to the goal of
the analysis to answer the research question, and whether the selection criterion and
the level of abstraction were judiciously chosen. Where needed, the criterion and
level of abstraction were revised accordingly.
In the fifth step, the abstraction level was further raised and the categories arranged
according to main categories. During the coding process, memos were written to
document interesting findings and capture the relationships between different catego-
ries. For the revised categories, a codebook with anchor examples was created.
In a final step, a reliability check was accomplished by means of intercoder reliability

(Mayring 2015). There are a number of formulae for calculating intercoder reliability,
such as the simple coefficient of overlaps or the more sophisticated Krippendorff's α
(Krippendorff 2004).The applied reliability coefficient is one of the best-known and
most widespread for determining intercoder reliability: the Holsti formula 3 (Holsti
1969). The Holsti coefficient is expressed as a number between 0 and 1, depending
on the complexity of the construct coefficients, at least 0.8 is acceptable (Rössler and
Geise 2013). Therefore, a second researcher familiar with the research area is intro-
duced to the category system and the researcher code randomly selected data from
the material.
3.2.2 Quantitative Methods
Quantitative research has its origin in sociology, where there was a need to base pol-
icy on demographic and economic data (Stigler 1986). Long before the emergence of
the modern social sciences, researchers were already striving to get closer to their
research objectives by using measurements, tests and mathematical quantifications.
(Bortz and Döring 2007). The application of quantitative methods is therefore charac-
terised by the collection of data in numerical form and the conversion of the charac-
teristic value into numerical format, where data are not naturally available in numeri-
3
Coefficient of reliability = 2m / (n1+n2), where m is the number of judgments on which both of the
coders agree, and n1 and n2 are the numbers of judgments made by coder 1 and coder 2,
respectively.
55
Research Design
__________________________________________________________________________
cal form. The measured values collected in this way are usually evaluated with the
help of statistical methods (Creswell 2009).
Compared to qualitative methods, quantitative methods can be distinguished by the

following characteristics: clearly structured procedure, and clearly defined and de-
scribed analysis methods. The collection and evaluation of data does not require in-
terpretation by the researcher. The results obtained are therefore intersubjectively
comprehensible, and the investigations carried out can be replicated. The quality of
the research results and the quality of the conclusions drawn can therefore be as-
sessed by outsiders and checked in comparative studies (Schwaiger and Meyer
2011).
In this context, well accepted data collection methods are: survey methods, labora-
tory experiments, formal methods, and numeric methods. In the following, the applied
methods are presented and described in more detail.
3.2.2.1 Survey
"A survey design provides a quantitative or numeric description of trends, attitudes,

or opinions of a population by studying a sample of that population. From sample re-
sults, the researcher generalises or makes claims about the population" (Creswell
2009, p. 137). Surveys usually follow a standard format for which there are numerous
examples in the literature. Thereby, the following components should be considered.
The first step is to develop the survey design. This includes, first checking the pur-
pose of a survey and the reasons for its selection for the proposed study. In addition,
it should be specified whether the survey is a cross-sectional survey, in which data
are collected at a specific point in time, or whether it is a longitudinal survey (Fowler
2013). Finally, the form of data collection is specified, distinguishing four types: self-
administered questionnaires, interviews, structured record reviews, and structured
observations (Fink 2002). Data collection may also include the creation of a web-
based or Internet survey and its online management (Sue and Ritter 2012). Regard-
less of the form of data collection, a justification for the procedure should be pro-
vided, including the strengths and weaknesses of the procedure.
The second step includes the specification of the characteristics of the population
and the sampling procedure. According to Creswell (2009), the following essential
aspects should be considered:
56
Research Design
__________________________________________________________________________
 Identification of the population in the study

 Identification of the selection process, e.g. random sample, convenience sam-
ple
 Identification of, if the study foresees stratification of the population prior to
sample selection
 Define the methods for sample selection, e.g. choose individuals using a ran-
dom numbers table
 Indication of the number of people in the sample
Then, it is necessary to develop and use suitable measuring instruments. The use of
suitable scales is of paramount importance in the research process. Their quality has
a decisive influence on the quality of the data collected and the research results
achieved. The criteria of objectivity, reliability and validity are usually used to assess
the quality of a scale (Herrmann and Homburg 2000). Only where these criteria are
met can quantitative research methods exploit their advantages such as comparabil-
ity, validity and replicability of the results. Wherever possible, already validated
scales should be used. If validated scales do not yet exist, measuring instruments
must be developed and tested beforehand.
Then, the survey instrument has to be determined. As part of a rigorous data collec-
tion process, the researcher must also provide detailed information about the survey
instrument used in the study. Finally, the variables should be related to the specific
questions or hypotheses of the instrument.
3.2.2.2 Statistical Analysis
Before starting with the analysis of the data, the response bias (effect of nonre-
sponses) should be identified and the data cleared accordingly (Fowler 2013).
In general, the analysis of quantitative studies starts with descriptive statistics for all
independent and dependent variables in the study. "This analysis should indicate the
means, standard deviations, and range of scores for these variables" (Creswell 2009,
p. 143).
An essential characteristic of quantitative methods is the applicability of statistical

methods. These allow the exact assessment of the quality of the generated state-
ments and enable inferential statistical verifications as to whether an effect is ran-
domly present in the investigated sample or whether it can be assumed that the ef-
fect also exists in the population (Schwaiger and Meyer 2011).
57
Research Design
__________________________________________________________________________
In order to achieve such a quantitative assessment of the research results obtained,

the hypotheses investigated must first be expressed in mathematical-exact form.
Once the research hypotheses have been precisely formulated in this way, statistical
tests can be applied (Bortz and Döring 2007).
In order to determine the appropriate statistical test (e.g. to answer the research
question or test the hypothesis), the following factors have to be considered (Cres-
well 2009):
 Choose according to the type of research question (e.g. relate variables to

each other or compare groups), the number of independent and dependent
variables and the number of variables controlled.
 Consideration of whether the variables on an instrument should be measured
as a continuous score or a categorical score.
 Consideration whether the values from the sample could be normally distrib-
uted or not.
The elements combined results in the selection of a number of common statistical

tests (e.g. t-test, analysis of variance and covariance, Mann–Whitney U test, Chi-
square test, Pearson test) (Creswell 2009; Schwaiger and Meyer 2011).
3.3 Research Strategy
In this section, the research strategy of the present dissertation is outlined. First, the
research goals are presented and the research questions defined (3.3.1). Next, the
applied research paradigm is briefly introduced (3.3.2). Finally, the research ap-
proach is described in section 3.3.3.
3.3.1 Research Objectives and Questions
The dissertation at hand aims to shed light on the phenomenon of herd behaviour in
consumers purchasing decisions and the aspects influencing the impact of social
media using the mobile communications industry as an example. Therefore, the fol-
lowing research questions are addressed.
Many scientific disciplines such as marketing, psychology, economics, information

and knowledge management, and information systems show a steady increase in in-
terest studying the phenomenon of herd behaviour in purchasing decisions (e.g.
(Huang and Chen 2008), Chen et al. (2011), Kuan et al. (2014), and Liu et al.
(2017)). Given the development of social media and with the easy availability of in-
58
Research Design
__________________________________________________________________________
formation about other users’ choices in recent years, it is also not surprising that this
aspect is also being considered in research in this context; see e.g. Nakayama et al.
(2010), Lee et al. (2015), and Li and Wu (2018). The target of this section is to pre-
sent the current state of research on the subject. In order to provide this theoretical
background a systematic literature review was conducted. The literature review aims
to understand the topics addressed in literature regarding herd behaviour in consum-
ers' buying decisions so as the role of social media within this phenomenon.
Study I, in which the literature review was conducted, aims to answer the following
research questions:
RQ 1a What are the main topics addressed by scientific literature on herd

behaviour in the consumer buying decision process?
RQ 1b What aspects are addressed by scientific literature on the influence of so-

cial media on herd behaviour in buying decisions?
The relative effectiveness of different social media types and content characteristics
on buying process has been explored by many researchers in different industries
(Maas et al. 2014; Park et al. 2014; Powers et al. 2012; Zhang et al. 2014). Consider-
ing social interactions, Peters et al. (2013) proposed four social interaction categories
(sharing, gaming, expressing, and networking); however, there is scarce research
exploring deeper social interactions on different social media platforms. Thus, the
aim of the empirical pre-study (Study II) is to elaborate on this issue and provide
novel and intriguing categories on social interactions in the mobile communications
context. Building on these thoughts and the industry context Study II aims to answer
the following research questions:
RQ 2a How do social interactions differ depending on different social media plat-

forms?
RQ 2b How can social interactions be specified in relation to the mobile

communications industry?
The growing use social media and different social media types makes it complicated
for firms to understand their influence on consumer behaviour and how to deal with
59
Research Design
__________________________________________________________________________
this phenomenon (Hanna et al. 2011). The high level of transparency provided by
online platforms poses a particular challenge to industries that offer commodity prod-
ucts to consumers, as they cannot offer a high level of product or price differentiation.
The mobile communications industry is chosen as the empirical context for a com-
modity market (Li and Whalley 2002). Hence, it is important to understand the inter-
play of online communities, herd behaviour and buying decision.
Building on these thoughts, Study III aims to answer the following research question:
RQ 3 What factors affect the influence of social media on herd behaviour in buy-
ing decisions?
First, it is necessary not only to identify herding on buying intention but to understand
the influence on the holistic view of buying decision process: in the pre-buying phase,
when consumers actively search for information and evaluation of alternatives (Kuß
and Tomczak 2007).
Second, it is necessary to understand the effectiveness of different on herding along

the buying decision process. Study IV utilises cues from different information
sources, measures product involvement, social media activity, and investigates per-
sonality types, to examine herd behaviour in purchasing mobile communications
products.
Building on these thoughts, Study IV aims to answer the following research question:
RQ 4a How does herd behaviour generally differ in the consumer buying decision
phases?
RQ 4b How do different factors influence herd behaviour in the individual

buying decision phases?
Addressing these research questions contributes to a deeper understanding of herd

behaviour in consumer purchasing decisions and the aspects that influence the im-
pact of social media in the mobile communications industry.
60
Research Design
__________________________________________________________________________
3.3.2 Research Paradigm
In addition to the positioning in the methodological IS profile in section 3.2, the cho-
sen research perspective should also be disclosed in order to understand the under-
lying research paradigm. In IS research there is a fundamental distinction between
critical, interpretative, and positivistic research perspectives (cf. section 3.1.1). Study
I sets the theoretical foundation of the dissertation at hand, while Study II and Study
III are based on an interpretative approach. "Interpretive studies assume that people
create and associate their own subjective and intersubjective meanings as they in-
teract with the world around them. Interpretive researchers thus attempt to under-
stand phenomena through accessing the meanings that participants assign to them"
(Orlikowski and Baroudi 1991, p. 5). This means that within the framework of Study II
and Study III a detailed understanding is obtained via the problem domain. It is as-
sumed that the assessments are of a subjective nature and are considered in isola-
tion by means of Study II and Study III; hence, no generally valid statements can be
made.
For this reason, a positivistic research perspective is adopted in the follow-up study
(Study IV). According to Orlikowski and Baroudi (1991), positivist studies assume the
existence of a priori defined relations within phenomena that are typically studied with
structured instruments. This type of study is primarily intended to test the theory, in
an attempt to improve the predictive understanding of the phenomena. It is therefore
assumed that the observations in Study IV are objective and that the presented hy-
potheses can therefore be generally validated on the basis of the perceptions.
The combination of interpretative and positivistic research perspectives within a re-

search project is not unusual and is recommended in scientific practice for mutual
supplementation (Lee and Baskerville 2012).
3.3.3 Research Approach
In principle, it is evident that in IS research there is a trend towards the use of multi-
ple methods, which is defined by the combination of qualitative and quantitative
methods in a research framework (Mingers 2001).
In the present dissertation a research framework is proposed that includes quantita-

tive and qualitative research methods, with a predominance of qualitative ones.
Table 7 depicts the research methods applied in the different phases. Thereby, a
cross (x) indicates that the respective method is applied within the corresponding re-
61
Research Design
__________________________________________________________________________
search sequence. Taking into account the research questions and the research con-
text, it is believed that these methods are appropriate for the respective study. Ac-
cording to the typology of Mingers (2001), the presented multi-method research
framework follows the 'QUAL → qual' and 'QUAN → quan' design type; here, '→' in-
dicates a sequential application process, 'QUAL' indicates a qualitative-driven re-
search design, and 'QUAN' indicates a quantitative-driven design.
Table 7: Research Methods applied in Each Research Sequence
Research Method
Qualitative Quantitative
Research Research
Coding Procedure
Non-reactive Data
Expert Interview
Sequence Design
Online Survey
Collection
Statistical
Literature
Analysis
Review
Study I QUAL x
Study II QUAL → qual x x
Study III QUAL → qual x x
Study IV QUAN → quan x x
Considering that the review of prior relevant literature is an essential feature of any
academic project, Study I provides the results of a comprehensive literature review,
following the guidelines of Webster and Watson (2002). Accordingly, the state-of-the-
art research on the phenomenon herd behaviour in buying decisions considering the
aspects of social media are presented.
Study II is positioned as the first empirical study, and follows an explorative approach
with the aim to gain data from unaffected conversations to allow a deeper under-
standing of people's interactions the research process applies non-reactive data col-
lection from different social media sources (Aggarwal and Zhai 2012; Fielding et al.
2017; Hewson 2017). The analysis of the generated content is carried out based on
the qualitative content analysis.
62
Research Design
__________________________________________________________________________
Study III strives to understand which factors affect the influence of online communi-
ties on herd behaviour in purchasing decisions. In order to utilise the considerable
knowledge of practitioners, the research question was approached through explor-
ative expert interviews. An essential goal of expert interviews is to provide new sup-
porting ideas for further research (Bogner et al. 2014); this approach is also pursued
in this study. The conducted expert interviews were followed by qualitative content
analysis according to the approach of Mayring (2015).
Study IV, a quantitative cross-sectional study, examines how different aspects in the
course of the buying decision process affect the decisions of the test persons. To test
the developed hypotheses, an online self-administered and voluntary survey was
conducted. To be able to analyse different influencing factors on herd behaviour
along the buying decision process, a split-sample approach was used, where partici-
pants were randomly assigned to three comparison groups. The conducted survey is
followed by a statistical analysis.
63
Study I: State-of-the-Art Analysis
__________________________________________________________________________

Chapter overview: In order to develop a comprehensive understanding of the cur-
rent state of research on herd behaviour in buying decisions, and hence to anchor
the empirical studies, a comprehensive literature was conducted. Section 4.1 pro-
vides the objectives of the literature review and a description of the approach applied
to identify the relevant literature. The first part of the results is presented in section
4.2 in the form of descriptive statistics of the literature set identified. Subsequently,
section 4.3 comprises a qualitative assessment. Here, the main topics of the litera-
ture identified are elaborated with regards to herd behaviour in the buying process,
and the aspects affecting the impact of social media on herding in purchasing. A
summary and interim discussion rounds up the chapter in section 4.4.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 17: Dissertation Overview - Chapter 4: State-of-the-Art Analysis
64
__________________________________________________________________________
4.1 Literature Review Approach
The target of this section is to present the current state of research on the subject of
this dissertation. In order to provide this theoretical background a systematic litera-
ture review was conducted. The literature review aims to understand the topics ad-
dressed in literature regarding herd behaviour in consumers' buying decisions so as
the role of social media within this phenomenon. Therefore, the literature review tar-
gets to answer the following research questions:


For an objective review of a field of research, systematic literature reviews are rec-
ommended in the methodological literature. These are characterised by the fact that
the search, selection and evaluation process is disclosed and the results are repro-
ducible and comprehensible (Okoli 2015; Webster and Watson 2002). Following the
guidelines for a systematic literature review according to Webster and Watson
(2002), the theoretical foundation is provided and gives the context of the research
question. The approach is structured in four steps (Figure 18).
A
Definition of the • Definition and selection of search outlets
source material • Definition of requirements
B
Identification and • Definition of search terms
selection of relevant • Definition of boundaries (e.g. level of analysis temporal and
publications contextual limitations )
C
Backward and • Review of citations within identified articles
forward search • Identification of articles citing identified key articles
D
• Descriptive quantitative analysis
Analysis • Qualitative analysis
Figure 18: Steps of Literature Review according to Webster and Watson (2002)
65
__________________________________________________________________________
A. Determination of the Source Material
To answer the research questions and to contribute to the development of the disser-
tation at hand, the Senior Scholars' Basket of Journals4 of Association of Information
Systems (AIS), the German journal Wirtschaftsinformatik as well as leading IS con-
ferences were selected as search options. In order to exclude the possibility that
relevant publications from relevant conferences or journals remain unrecorded, rele-
vant conferences and journals are additionally searched that are not listed in the
VHB-JOURQUAL5 ranking. Since the existing research questions relate to other re-
search streams besides IS, as marketing and psychology, scientific databases
namely Business Source Complete (EBSCO) and ABI/Inform were also used. Web-
ster and Watson (2002) also recommend to look outside the IS discipline since IS is
an interdisciplinary field that extends to other disciplines. To ensure a high standard
and exclude irrelevant contributions the analysis was restricted to peer-reviewed arti-
cles.
Table 8: Overview of Research Sources
Source cate- Name of source

gory
IS Journals - European Journal of Information Systems (EJIS)

- Information Systems Journal (ISJ)
- Information Systems Research (ISR)
- Journal of Association for Information Systems (JAIS)
- Journal of Information Technology (JIT)
- Journal of Management Information Systems (JMIS)
- Journal of Strategic Information Systems (JSIS)
- Management Information Systems Quarterly (MISQ)
- Wirtschaftsinformatik (WI)
4
The College of Senior Scholars encourages colleagues, as well as deans and department chairs,
to treat a "basket" of eight journals as top journals in the IS field.
https://aisnet.org/page/SeniorScholarBasket
5
VHB ranking: http://vhbonline.org/en/service/jourqual/vhb-jourqual-3/
66
__________________________________________________________________________
Source cate- Name of source

gory
IS Confer- - Americas Conference on Information Systems (AMCIS)

ences - European Conference on Information Systems (ECIS)
- International Conference on Information Systems (ICIS)
- Hawaii International Conference on System Sciences
(HICSS)
- Pacific Asia Conference on Information Systems (PACIS)
- Wirtschaftsinformatik (WI)
- Mediterranean Conference on Information Systems (MCIS)
Databases - Business Source Complete (EBSCOhost)

- ABI INFORM Complete
B. Identification and Selection of Relevant Publications
Before starting with the retrieval of relevant literature several test queries and result
verifications were carried out. Furthermore, the search was limited to the years 2008–
2018 so that a comprehensive set of literature could be obtained and to ensure a fo-
cus on newer research. The defined search parameters were defined and relatively
broad search terms were chosen. These were 'herd behaviour', 'buying decision' and
'social media'. Since there are synonyms used across the papers, corresponding
synonyms (e.g. herding, purchasing, virtual communities, online communities) were
also applied in the search process.
As there was a very high number of hits (62,179), a combination of search terms
should refine the results and increase relevance to the present research questions.
The identification and selection of relevant publications followed a three-stage filter-

ing process. First, the search terms were applied to the title, abstract, and keywords,
and a pairwise combination of search terms resulted in 8,574 hits of peer-reviewed
articles. These included overlaps of the IS journals, IS conferences and two scientific
databases for the selected time period. This was followed by a pre-selection of rele-
vant publications were roughly screened and sorted by reading title, keywords, and
abstract. In this step, 123 publications that promise an answer to the research ques-
tions in the preliminary sample were selected. In the last step, a full-text reading was
67
__________________________________________________________________________
performed of the identified publications, resulting in 43 relevant articles. Table 9

shows the results of the applied reduction process and the outcome of each step.
Table 9: Reduction Process of the Literature Review
Sources 1. Systematic key 2. Selection of rele- 3. Full-text reading

word search hits in vant publications af-
abstract, title, or ter the first screen-
keywords ing
IS Journals 389 49 11
IS Conferences 253 23 10
Databases 7,932 51 22
Total 8,574 123 43
To ensure a stringent and consistent selection of articles the described filtering proc-
ess considered the following key questions and the definition of exclusions as a
guideline while reading title, abstract and key words:
 Does the article focus on herd behaviour in the consumer buying decisions? or
 Does the article focus on the impact of social media on herd behaviour in buy-
ing decisions? or
 Does the article consider different aspects that influence herd behaviour in the
buying decision process?
In addition to the justification for including scientific articles, it is important to describe

which aspects are excluded from literature research (Webster and Watson 2002).
Therefore, the following exclusions were applied:
 Missing multi-dimension:
o studies focusing only on the phenomenon herd behaviour itself without
a linkage to consumers' individual buying decisions; see e.g. Morone
and Samanidou (2008), Veeraraghavan and Debo (2011), and Raafat
et al. (2009a).
o studies focusing on social media itself without a linkage to herd behav-
iour (e.g. motivation of using social networks, sustainability of social
68
__________________________________________________________________________
media) were excluded; see e.g. Salehan et al. (2017), and Gee-Woo et
al. (2015).
 Level of perspective: B2B, B2P purchasing decisions were excluded; see e.g.
Greenwood and Gopal (2017), Abdalla Mikhaeil and George (2016), and Luo
and Lin (2013).
 Level of investigation: studies focusing on industry or firms decisions were ex-
cluded (e.g. herd behaviour in the financial industry); see e.g. Ackert et al.
(2008),and Ali and Kartik (2012).
C. Backward and Forward Search

According to the recommendation of Webster and Watson (2002) a backward and
forward search was conducted. Backward search means that the citations of the
identified articles which were considered to be very close to the research goals of the
dissertation at hand were reviewed. The proximity of a publication was determined
through the qualitative evaluation while reading the entire text. Subsequently, a for-
ward search was carried out using the search database Google Scholar. Finally, a
backward and forward search were conducted to set the foundation for the subse-
quent analysis of the searched literature and revealed five additional relevant arti-
cles. Including this final step, 48 articles were subject to analysis.
D. Analysis Approach
A literature review generally provides a large amount of data. In order to be able to

synthesise this amount of knowledge meaningfully and to present the results trans-
parently, Webster and Watson (2002) propose the following: "A literature review is
concept-centric and thus, concepts determine the organizing framework" (Webster
and Watson 2002, p. 16). This step involves combining the facts extracted from the
studies according to common criteria using appropriate techniques, whether quantita-
tive, qualitative, or both (Okoli 2015).
Therefore, section 4.2 first provides a quantitative analysis (descriptive statistics) of

the literature set identified. Specifically, articles were categorised along their associ-
ated research streams, methods and theoretical lenses. Then, in section 4.3 the re-
sults of the qualitative analysis of the literature set are presented. The qualitative
analysis is guided by the research questions with the aim to thematically classify and
describe the main topics addressed (4.3.1) and to evaluate the different aspects in-
69
__________________________________________________________________________
fluencing the impact of social media on herd behaviour in purchasing decisions

(4.3.2), which is of particular interest for the research goals pursued here.
To identify the main topics (RQ 1a), several potential categories were iterated in vari-
ous runs, while constantly challenging them against the set of literature identified and
assigned keywords that emerged from multiple readings. After having reached a
saturation point, in the sense of a good balance between the correspondence be-
tween the topics of one category and the demarcation between the topics of different
categories, three categories remained, which were used to structure the topics and
derive subcategories. This method revealed the following categories: 'information
type', 'information source', and 'technology adoption', each with subcategories. With
regards to the aspects influencing the impact of social media on herd behaviour in
purchasing decisions (RQ 1b) the same approach to identify the main aspects was
applied and five clusters identified. The identified clusters are 'popularity', 'valence',
'information quality', 'homophily', and 'product type'. An additional cluster was defined
as 'others' as topics were not assignable to the other five categories.
The analysed literature set and their categorisation are summarised in a concept ma-
trix according to Webster and Watson (2002) and is provided in Appendix B.1.
70
__________________________________________________________________________
4.2 Results of the Quantitative Analysis
The identified literature so far showed that the topic became significantly more rele-
vant over time, the almost fourfold increase from 2008 to 2018 may be interpreted as
an increasing interest among researches in the topics. Especially, from 2014 on-
wards the number of hits has risen even further. Figure 19 shows the first overview of
the temporal classification of the identified publications with regards to RQ 1a and
RQ 1b.
Number of publications
8
8
7 7
7
6 6
6
4
3 3 3
3
2 2
2
1
1
0
2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Figure 19: Temporal Classification of Identified Publications
For the purpose of clarity and to support the following analysis with regard to the dis-
tribution of topics over time, the publications have been divided into two different time
periods. The first period includes the years 2008 to 2013; the second period, in which
the topics seem to have gained in importance, covers the years 2014 to 2018.
The analysis of the industries within the scope of the literature review revealed that
out of the 48 articles only one article deals with the mobile communications industry
(Godinho de Matos et al. 2014). The remaining articles either chose the industry con-
text e-commerce, software industry, music industry, retail or no industry. Therefore, in
the following analyses, Study I deals with the research questions in general, without
specifically referring to the mobile communications industry.
The following sections provide descriptive statistics of the literature identified con-
cerning their assigned research streams, methods and theoretical lenses applied, as
71
__________________________________________________________________________
well as an overview of the categories with regards to the research questions (RQ 1a,
RQ 1b).
4.2.1 Contributing Research Streams
The classification of the publications into research areas shows that the topic of herd
behaviour in the buying decision process and the aspects of social media are re-
flected in five research areas: 'IS', 'Economics', 'Marketing', 'Information & Knowledge
Management' (I&KM), and 'Psychology' (Figure 20). 'IS' research is dominant, which
indicates that the topic seems to be of particular interest there. The increase could be
explained by the rise of social media and the interest of researchers from the IS field
to study their influence. This is followed by the research areas 'Economics' and 'Mar-
keting' where the topic also seems to be of great interest. In comparison, the re-
search areas 'I&KM' and 'Psychology' seem to be less focused on this topic. The
breakdown into the two time periods reveals an increase in articles during the time
period of 2014 to 2018 compared to the preceding years. The increase is reflected in
all research streams, with the exception of 'Psychology', where researchers pub-
lished only in the first time period of the investigated timeframe.
The assignment to the research streams results from an investigation of the sources.
Conferences are usually clearly assigned to a research area, and to classify these ar-
ticles the journal websites were examined. These usually provide a classification of
their main research objectives and topics.
Number of publications
26 2008 - 2013
26
24 2014 - 2018
22 7
20
18
16
14
12
10
10 19 2
8 7
6 2
4 8 3
5 1 2
2
2
0
IS Economics Marketing I&KM Psychology
Figure 20: Contributing Research Streams of the Literature Set Investigated
72
__________________________________________________________________________
4.2.2 Applied Methods
With regard to the methods applied, the analysis of the literature set revealed the fol-
lowing seven classifications: 'survey', 'experiment', 'mathematical modelling', 'quanti-
tative analysis', 'literature review', 'conceptual', and 'multiple case study' (Figure 21).
The analysis shows that more than one method is used in some studies. Almost one
third of the applied methods in the articles are based on 'surveys', with most of them
carried out online. Furthermore, 'experiments' are also widespread as an applied
method, with different types of experiments being used, such as laboratory experi-
ments (e.g. Chen (2008)), natural experiments (e.g. Chen et al. (2011)), neurosci-
ence experiments (e.g. Purnawirawan et al. (2014), or online quasi experiments (e.g.
Dewan et al. (2017)). 'Mathematical modelling', which is mainly based on crawled or
download data (e.g. Amy Wenxuan and Li (2018), and Lee et al. (2015)), is also very
common. This is followed by the method 'quantitative analyses', which are mainly
based on statistical analysis of retrieved data, e.g. data extractions from a website
(e.g. Duan et al. (2009)). Meanwhile, 'literature review', 'conceptualisation', and 'multi-
case study' are employed less often.
A closer look into the two time periods shows that almost all methods are more
strongly represented in the second time period. It is noticeable that 'mathematical
modelling' is not represented at all in the first time period, but in the second time win-
dow it is the third most strongly represented method. This may possibly be explained
by the fact that most mathematical models are based on digital trace data and that
the technical possibilities for collecting such data have been greatly expanded in re-
cent years.
73
__________________________________________________________________________
Number of applied methods

18
16 2008 - 2013
16 2014 - 2018
14
14
6
12 11
6
10
8
8
3
6
10
4 8
5 2 2
2 1
1
1
0
Survey Experiment Mathematical Quantitative Literature Conceptual Multiple
modelling analysis review isation case study
Figure 21: Applied Methods of Literature Set Investigated
4.2.3 Applied Theoretical Lenses
In order to provide an overview of the applied theories, the theoretical foundations of

the identified publications were analysed. Articles which not explicitly applied a theory
were excluded from the analysis. As a result, 34 of the 48 publications (71%) use at
least one theory. Figure 22 gives an overview of the identified theories. In total,
seven theories were recognised and have been applied several times. In addition,
theories that occur only once in the analysed literature set were grouped together to
the category 'others'. Most theories can be found in the IS Theory collection of Larsen
(2015) or other encyclopaedias, e.g. the international encyclopaedia of social and
behavioural sciences (Baltes and Smelser 2004). A short introduction to each of the
theories referenced is provided in Appendix B.2. The seven theories in descending
frequency are:
 'Information cascade theory' describes the conditions which lead to the

mechanism of herd behaviour; the presence of uncertainty about a decision,
and the observed actions of other individuals (Banerjee 1992; Bikhchandani et
al. 1992).
74
__________________________________________________________________________
 'Social influence theory' describes that an individual’s attitudes, beliefs, and

subsequent actions or behaviours are influenced by referent others (Kelman
1958).
 'Herding theory' describes the phenomenon where “everyone does what eve-
ryone else is doing, even when their private information suggests doing some-
thing quite different" Banerjee (1992, p. 798).
 'Social learning theory' is a category of learning theories based on the belief
that human behaviour is ascertained by a combination of three aspects: cogni-
tive factors, environmental influences and behaviour (Bandura 1969).
 'Dual-process theory' describes the fact that thoughts can arise in two different
ways or as the result of two different processes, thereby the two processes
can consist of an implicit unconscious process and an explicit conscious proc-
ess (Chaiken and Trope 1999).
 'Attribution theory' describes the process by which people associate causes
with events and the results they experience (Kelley and Michela 1980).
 'Adoption theory' deals with the decisions an individual makes to accept or re-
ject a particular innovation (Straub 2009).
 Under 'others' the following six theories are summarised (in in alphabetical or-
der): 'information economics theory', 'information processing theory', 'long tail
theory', 'media richness theory', 'prospect theory', and 'social network theory'.
In some articles more than one theoretical lens is applied, frequently 'social learning
theory' is used in combination with 'information cascade theory' or 'herding theory';
see e.g. Chen et al. (2011), and Salazar et al. (2013).
Looking at the theories applied over the course of time, it is noticeable on the one
hand that the variety of theories applied is increasing in the second time period. In
addition, the examination under the theory lenses: 'social influence theory' and 'herd-
ing theory' seems to be more frequent in the years 2014–2018.
75
__________________________________________________________________________
Number of applied theories
2008 - 2013
2014 - 2018
10
8
7 7
6 6
6 2
1
4 2
4
4
1 3
5 5 1 2 2
2 4
3 3
2
0
Information Social Herding Others Social Dual- Attribution Adoption
cascade influence theory learning process theory theory
theory theory theory theory
Figure 22: Applied Theoretical Lenses of Literature Set Investigated
4.2.4 Overview of identified Key Topics
In relation to the first research question (RQ 1a), a total of three main topic clusters
could be identified: 'information type', 'information source', and 'technology adoption'.
Each article is assigned to at least one or two (10 occurrences) key topic clusters,
whereby the combination of the two topic clusters 'information type' and 'information
source' occur significantly more frequently (8 occurrences). Figure 23 gives an over-
view of the distribution of topics over the two time periods.
It is striking that the treatment of the topic 'information type' in literature has more
than doubled in the period between 2014 and 2018. The topic 'information source'
could also be found dominantly in the second time period, while 'technology adoption'
has been researched in the first time period.
76
__________________________________________________________________________
Number of main topics addressed
35
32 2008 - 2013
30 2014 - 2018
9
25
20 19
15 8
23
10
5 11 4
3
0 1
Information Type Information Source Technology Adoption
Figure 23: Key Topics Identified in Investigated Literature Set
4.2.5 Overview of Aspects affecting the Social Media Impact
With regards to the second research question (RQ 1b), a total of six topic clusters
could be summarised: 'popularity', 'valence', 'information quality', 'product type', 'ho-
mophily', and 'others'. Topics which only rarely occur in the literature set and could
not be assigned to any of the other categories are grouped together in 'others'.
Each article is assigned to at least one or two (14 occurrences) key topic clusters,
one article is assigned to three topics (Liu and Karahanna 2017). The combination of
the topic clusters 'popularity' with 'valence' or 'information quality' (6 occurrences) oc-
curs most frequently. Figure 24 gives an overview of the distribution of topics over
the two time periods.
A closer look at the development along the two time periods shows that all topics are
researched more frequently in the second time period. In addition, two topics stand
out in particular: 'information quality' is treated five times more frequently between
2014 and 2018 than in the previous period. Further, 'product type' is discussed more
than three times as often in the second time period.
77
__________________________________________________________________________
Number of aspects affecting the impact of social media

16
15
14 2008 - 2013
2014 - 2018
5 12
12
2
10
10
9 9
3
8 2
4
6
6
10 10 2
4
7 7
5
2 4
0
Popularity Information Others Valence Product Type Homohily
Quality
Figure 24: Identified Aspects affecting the Social Media Impact
4.3 Results of the Qualitative Analysis of the Literature Set
The following sections present the results of the qualitative analysis of the literature
set. First, the topics are described which have been identified in the literature on the
subject of herd behaviour in the buying decision process. Then, aspects influencing
the impact of social media on herd behaviour in purchasing decisions are described.
4.3.1 Key Topics regarding Herd Behaviour in Buying Decisions
In relation to the first research question (RQ 1a), a total of three main topic clusters
could be identified: 'information type' (32 occurrences), 'information source' (19 oc-
currences), and 'technology adoption' (4 occurrences). In each category respective
subcategories are outlined. The analysis of the identified literature results in the fol-
lowing assignment under 'information types': 'word of mouth' (WOM), 'eWOM', 'sys-
tem recommendations', and 'buy/sales information'. Further, the following assignment
to 'information source': 'peer groups', 'experts', and 'crowd'. Under 'technology adop-
tion' no differentiation of further topics was identified. In the following, the categories
are briefly described and the most-discussed questions in the identified literature are
summarised.
78
__________________________________________________________________________
4.3.1.1 Information Type
With regard to research on the topic of herd behaviour in the buying decision proc-
ess, the most common types of information a briefly described and the main ques-
tions from the analysed literature set is outlined. In the present cluster different infor-
mation types are summarised, these are: 'WOM', 'eWOM', 'system recommenda-
tions', and 'buy/sales information'.
WOM
Traditional WOM is the interpersonal communication between consumers and de-

fined as the one-to-one and face-to-face exchange of information. The process of
conversations between people that know each other about a particular product, ser-
vice, firm or a brand usually because they think it is good and want to encourage
other people to try or in case of a negative experience to discourage someone from
doing something. The significance of WOM communication is well documented and
as it has proven to play a major role in consumer buying decisions and choices by
reducing consumer uncertainties regarding product quality (Arndt 1967; Katz and
Lazarsfeld 1955). As consumers in general trust WOM more than they trust advertis-
ings, it is so influential (Sen and Lerman 2007). Furthermore, it is independent of
marketers’ selling intents and is thus considered to be more trustworthy and credible
(Bickart and Schindler 2001).
eWOM
The increased proliferation of social media has expanded traditional interpersonal

communication in the form of eWOM (Chen et al. 2011; Godes et al. 2005). Further-
more, eWOM has several unique characteristics as it often occurs between strangers
or have little or no relationship with one another (Sen and Lerman 2007). The given
anonymity through the Internet is considered as an advantage of eWOM, since it
supports users to share opinions more comfortably. At the same time, anonymity al-
lows users to more comfortably share their opinions without revealing their identities
(Goldsmith and Horowitz 2006). Therefore, anonymity is also seen as one of the im-
portant factors that ensures higher volume of eWOM (Chatterjee 2001). Conse-
quently, there is a greater likelihood that users find other consumers with product ex-
pertise on the eWOM platforms and use this information during their buying decision
process (Lee and Youn 2009). eWOM helps consumers to evaluate offerings through
the lens of previous customers, and this benefits both consumer and sellers. There-
79
__________________________________________________________________________
fore, this information type has become key for consumers to obtain product or service
information and reduce their uncertainties regarding price, quality, etc. (Jabr and
Zheng 2014). Nevertheless, eWOM is still more popular among younger consumers
than among their older counterparts (Nakayama et al. 2010).
System Recommendations
System recommendations are becoming integral to support customers evaluating

competing offerings and entice them into considering alternative products or services
(Jabr and Zheng 2014). They are used for all major types of products and services,
such as books, movies, music, news, and television, etc. "They are commonplace at
major online firms, such as Amazon, Netflix, and Apple’s iTunes store, and they have
a strong influence on what consumers buy and view" (Fleder et al. 2010, p. 2). These
systems also help firms convert browsers into buyers, cross-sell products, and in-
crease customer loyalty via a customised searching experience (Fleder et al. 2010).
The value for consumers of system recommendations is personalisation according to
each user's taste, i.e. personalised newspapers or a personalised mobile offer ac-
cording to one's usage pattern (Fleder et al. 2010). But system recommendations are
not only of value for consumers but also for the industry, their importance is reflected
in the willingness of firms to pay significant sums to acquire and improve these sys-
tems (Jabr and Zheng 2014).
Buy/Sales Information
With advances in online marketing technology, retailers are able to provide product
information by showing past sales information to potential customers and thereby in-
crease product awareness, which may increase demand individually or collectively (Li
and Wu 2018). For example, Amazon.com displays a product’s sales rank in its de-
scription. The number of people who have bought a product or service is a basic in-
formation that is technically easy to implement but with great effectiveness (Li and
Wu 2018). Previous sales volume information reveals the actions of other consum-
ers, so supposedly increase certainty about product quality, inducing consumer to
herd, without including the reasons behind them (Bikhchandani et al. 1992;
Bikhchandani et al. 1998; Zhang and Liu 2012). Thereby, previous sales information
may be caused by randomness in early consumer choices, in this case herding does
not lead necessarily to buy the 'best' product (Simonsohn and Ariely 2008; Zhang
2010).
80
__________________________________________________________________________
Addressed Questions in the Literature Set
Most articles are dealing with the questions which information type has a greater im-
pact on herd behaviour within purchasing decisions.
For example, the following questions are frequently found in the literature: How do
consumers use these different 'information types' when making buying decisions?
Are these sources equally important? How do different 'information types' moderate
the effect of herd behaviour?
Several authors compare the impact of two commonly used types of information to
induce purchasing: 'buy or sales information' (the number of people who have bought
a deal) and 'like' information (e.g. the number of Facebook friends who 'liked' a deal)
(Kuan et al. 2014; Li and Wu 2018). They found that integrating these two types of in-
formation in the social media strategy facilitates herding and are complements that
generate sales (Li and Wu 2018).
Jabr and Zheng (2014) studied the joint impact of 'eWOM' in the form of user reviews
and 'system recommendations' on online sales. Based on a dataset collected from
Amazon.com consisting of books, sales ranks, recommendations, and reviews, they
found that 'system recommendations' lead to competitive pressure among products
and that those with higher centrality within the resulting network of referrals are asso-
ciated with higher sales. However, they also found that these sales gains are ham-
pered by improvements in the reviews of competing products.
In the study by Liu et al. (2017), it was investigated how two types of market-
generated information, namely, online reviews (eWOM) and past sales volume infor-
mation, jointly affect consumer buying decisions. The authors offered insights into
precisely why online reviews may not be beneficial; for example, when the signals
delivered from online reviews are inconsistent with the sales volume information. Fur-
ther, they found that the impact of the 'information type' depends on consumers’ prior
beliefs and the product.
Tseng et al. (2017) found that in an e-commerce context, review voters will discount
their own information when faced with clear and strong signals from previous voters.
Thus, they will herd in the previous voters’ voting direction. Conversely, review voters
will make their own judgement when confronted with weak signals from previous vot-
ers.
81
__________________________________________________________________________
4.3.1.2 Information Sources
The second cluster with regards to topics addressed in literature to herding in buying
decisions, the following different 'information sources' are summarised: 'peer groups',
'experts', and 'crowd'. Afterwards, their description and the addressed questions in
the literature set are outlined.
As described in the previous section interpersonal communication in the form of

eWOM plays a significant role in consumers' purchasing decision process. Neverthe-
less, there are different sources for eWOM and research has shown that the source
of the information plays an important role in consumers’ attitude and behavioural in-
tention formation (Purnawirawan et al. 2014).
Peer Groups
As described in the previous section interpersonal communication in the form of

eWOM plays a significant role in consumers' purchasing decision process. According
to Nielsen’s Global Trust in Advertising Survey (Nielsen 2015) 78% of consumers in
Europe say that recommendations from 'peer groups' are the most trusted sources of
information when making consumption decisions, followed by 60% of consumers who
say that they trust consumer opinions posted online. A 'peer group' is usually the
group of people who know each other, e.g. colleagues, family and friends, acquaint-
ances (Salazar et al. 2013). In this sense, the information provided by peer groups
with a higher degree of social proximity ('proximity influence') appears to have a
stronger weight in consumer's decision process (Salazar et al. 2013).
Companies are well aware of this phenomenon. Thus, they often actively reward
peer influence, e.g. offering a discount to seeds who bring friends who may also buy
the product (Godinho de Matos et al. 2014).
Godinho de Matos et al. (2014) provide empirical evidence on how information from
friends may shape the consumption of costly goods, using the example of iPhone
3Gs. The propensity of a subscriber to adopt increases with the percentage of friends
who have already adopted.
Salazar et al. (2013) investigate in their study the social influence that peer groups
may exert in the decision to choose for environmentally friendly products rather than
conventional ones. Their experimental setting, corroborated the relevance of peer ef-
fects, and clear evidence for herd behaviour.
82
__________________________________________________________________________
Experts
In addition to eWOM, consumers often rely on the opinions of well-known experts or

recommendations from well-known published publications, especially when buying
complex, durable, expensive products or services (e.g. financial products, home, car
and insurance) (Nakayama et al. 2010). Expert sources (e.g. sources who are con-
sidered as knowledgeable and competent) are assumed to possess the knowledge
and ability to provide accurate information, thus consumers tend to seek advice from
experts rather than non-experts (Pornpitakpan 2004). Tseng et al. (2017) state in
their paper: "reviews written by expert reviewers are more vivid, thus, more likely to
be voted". Purnawirawan et al. (2014) investigated in their experiment the role of ex-
perts and conflicting review information for buying intention. They found out that a re-
view is more likely to be discounted when it comes from an expert and/or when the
content of the review is not coherent with the rest of the reviews. Nevertheless, there
are few studies that assessed the relative impact of experts with other information
sources.
Crowd
Besides reviews submitted by friends or other peers, there are in addition to those
submitted by the rest of the online community (stranger-networks), the so-called
'crowd'. In general, the crowd consists of a great number of people with heterogene-
ous preferences, experiences and know-how, generating the wisdom of the crowd
(Surowiecki 2004). Therefore, online reviews generated by the 'crowd' are as well an
important source for prospective buyers to observe, and learn from the mass experi-
ences and behaviour of other users (Lee et al. 2015; Zhang et al. 2015).
The rise of social media has greatly increased the visibility of actions by both strang-
ers and friends. A good example is Amazon.com, they provide information about
other consumers’ choices on its website (Chen et al. 2011). One can observe all pre-
vious consumer purchasing decisions in the section "What do customers ultimately
buy after looking at this item?" on the product page. These new opportunities provide
consumers with more capabilities to observe the actions of others before making
their own decisions. At the same time, firms gain unprecedented capacity to strategi-
cally use different information sources among consumers.
83
__________________________________________________________________________
Addressed Questions in the Literature Set
Specifically, the identified articles focus on the questions which source of information
ultimately has the greatest influence on consumers decision behaviour. But the pic-
ture does not seem so unambiguous. The researchers present a differentiated pic-
ture here, one that seems to depend on other factors.
Chen (2008) shows in his study examining the relative strength of an expert opinion
versus crowd opinion in influencing book choices of consumers at an online book-
store. Their results show that recommendations of the crowd influenced subject
choices more than recommendations of an expert.
Nakayama et al. (2010) compared the perceived importance of different information

sources (self-evaluation, peers, crowd and experts) for different product types and
showed a differentiated picture depending on the product type and demographics of
the subjects, e.g. the importance of expert opinion over other consumers' opinion in-
creased by the age of consumer. Moreover, this perceived importance was more sig-
nificant for men than for women for experience goods. Overall, however, the expert
opinion wins out over recommendations from peers or other consumers.
Zhang et al. (2015) investigated the effectiveness of the 'information sources':

'friends' and 'crowd' for consumers and firms. They discovered that providing con-
sumers with access to friends’ purchase history is not always beneficial, particularly
for firms with superior quality products and where useful crowd information about the
product is available. In this case, the independent crowd information is more effective
in helping consumers to learn the true product quality.
The question of how reviews and ratings are influenced was examined by Lee et al.
(2015) in particular to find out whether the source ('crowd' or 'friends') influences the
rating of a later user differently in the context of movies. They found evidence in
herding induced by friends and crowd ratings, but the influence varies based on
whether a movie is popular or not. While user ratings always herd with friends rating,
this is not the case for crowd ratings where the herding behaviour is only obvious for
popular movies.
Tsao and Hsieh (2015) use the differentiation between corporate and independent
eWOM. Accordingly, review posts, e.g. those posted on enterprise websites, are allo-
cated to the group corporate. In contrast, public online forums or social networks
where people with similar interests and without the intention or performing monetary
84
__________________________________________________________________________
transactions interact with each other, are called independent. Consumers also regard
the eWOM from independent sources like peer groups or other consumers as more
credible, professional, and objective in the absence of apparent control by firms.
Therefore, independent sources are believed to have greater influence on herd be-
haviour than corporate sources like enterprise websites (Cheung et al. 2009; Kiecker
and Cowles 2002; Truong and Simmons 2010).
4.3.1.3 Technology Adoption
Another topic that existing literature addresses in the context of herd behaviour in the
buying decision process is 'technology adoption'.
Technologies belong to the most complex artefacts humans build, and since their im-
pact can take years to be realised, the benefits of adoption decisions are often uncer-
tain (Brooks Jr 1995). New technologies are constantly being introduced for users
and firms, the market is flooded with IT products, and users and firms have neither
the time nor the information to make well-informed adoption decisions. This provides
a perfect environment for information cascades and herding, which may lead to the
rejection of technically superior products in the case of doubt (Duan et al. 2009).
Usually, technology adoption does not occur by accident, "all of the information an
individual has about the adoption decision is consolidated and expressed in his or
her behaviour" (Walden and Browne 2009, p. 32). Irrespective of the amount or
source of information, consumers behaviour consolidates and expresses the informa-
tion he has available. Because of the complexity and uncertainty of technology deci-
sions, adopters make it easier for themselves by observing and utilising the behav-
iour of others (Walden and Browne 2009). In addition, technology adoption is often
sequential, so that later potential users can see the decisions (adopt or not adopt) of
earlier potential users. It is precisely this complexity and uncertainty that makes this
so appealing for researchers.
Walden and Browne (2009) explore the following questions: When people have little
information available, do people simply follow each other and essentially ignore the
small amount of information they personally have? How quickly, if at all, do potential
adopters start following the decisions of others? They found that decision-makers
more quickly converge on the correct decision because both their own private signals
and the inferred signals of others are more reliable. The probability of being right in-
creases the more decisions are made rather than a certain permanent state of af-
85
__________________________________________________________________________
fairs. The decision-making process can be slowed down if extreme signals threaten
to reverse the cascade. In this case, the decision also depends on a stronger bal-
ance between costs and benefits.
Duan et al. (2009) empirically examine informational cascades in the context of

online software adoption. Their results demonstrate "that online users’ choices of
software products exhibit distinct jumps and drops with changes in download ranking"
(Duan et al. 2009, p. 23). In addition, they found that user ratings have a positive in-
fluence on the adoption of low-ranked products in particular, rather than an influence
on the adoption of the most popular product.
Furthermore, extant research deal with the phase after the technology has been
adopted: the post-adoption stage (Sun 2013; Zou et al. 2015). Since users are faced
with trade-offs when deciding between different technologies, they anticipate the re-
sults of each option and choose the one with highest expected value. In the post-
adoption phase, adopters verify their decision quality against other options to see
what might have been (Zeelenberg and Pieters 2007). They regret their choice if they
believe that they would have made a better decision with another technology. There-
fore, the comparison with foregone options influences the assessment of the chosen
option (Zeelenberg and Pieters 1999).
Zou et al. (2015) shed light on post-adoption regret, which considers both the chosen
and foregone technologies. They investigated how herd behaviour induces regret
and how regret affects user satisfaction, as well as the subsequent continuance and
switching intention. They found that discounting own information and negative dis-
confirmation have a significant impact on post-adoption regret. In addition, post-
adoption regret in relation to foregone technologies influences user's evaluation and
thus their satisfaction on their chosen technology and their intention to continue. On
the other hand, independent from internal comparisons, regret will induce users to
switch to foregone technologies as an external drive even if users are satisfied with
the chosen technology.
Findings from the Sun (2013) study suggest that discounting own information and
imitating others when adopting a new technology are provoked primarily by the ob-
servation of the decisions of prior users and perceptions of uncertainty concerning
the adoption of new technology. They also point to that consumers "tend to adjust
86
__________________________________________________________________________
their beliefs when herding and also to revive their discounted initial beliefs to modify
their beliefs about the technology at the post-adoptive stage" (Sun 2013, p. 1013).
4.3.2 Aspects affecting the Impact of Social Media on Herding in Buying

Decisions
As outlined in the previous section there is clear evidence that different 'information
types' and 'information sources' have a moderating role on herd behaviour within the
buying decision process. Thereby, social media plays a prominent role. Despite this
evidence, it is not clear yet whereof this influence is affected. In order to shed more
light on this, the answer to the second research question (RQ 1b) is dealt with in the
next chapter. As a result, a total of six topic clusters could be named: 'popularity' (15
occurrences), 'information quality' (12 occurrences), 'valence' (9 occurrences), 'prod-
uct type' (9 occurrences), 'homophily' (6 occurrences), and 'others' (10 occurrences)
(see Figure 24). Under 'others' are the topics of 'consumer characteristics', 'strength
of ties', and 'social media usage'. In the following, these categories are briefly de-
scribed and the addressed questions from the literature are presented.
4.3.2.1 Popularity
As outlined in the previous section, the new technological possibilities, especially

those driven by social media with new types of real-time information, offer multiple
ways to observe the decisions and behaviour of others. Moreover, there is evidence
of clear influence on the own decision behaviour. However, it remains unclear on
which components this influence depends.
People tend to believe what most others believe, even though these beliefs may not
be true (Deutsch and Gerard 1955). Therefore, herding behaviour occurs driven by
the volume on information on social media, where consumers monitor the comments
of others regarding specific topics and use them as a basis for their own decisions
(Huang and Chen 2008).
Existing research addresses as one factor the 'popularity' of information (Chen 2008;
Chen et al. 2011; Dewan et al. 2017; Kuan et al. 2014; Liu et al. 2016). Popularity re-
fers to the total number of favourites from the community as a whole, which means
this is about the volume of information (e.g. number of likes, sales volume) (Dewan et
al. 2017). The volume of shared and available information plays an informative role
by increasing the degree of consumer awareness and the number of informed con-
sumers in the market and can therefore increase sales (Liu 2006).
87
__________________________________________________________________________
Since popularity influence is an important lever (Chen et al. 2011; Duan et al. 2009),
there are multiple dimensions of popularity that can be used to emphasise the effec-
tiveness of this information. Examples of these are best seller lists or rankings, the
number of 'likes', preferences, a number of times a product was recommended (De-
wan et al. 2017). Existing research sheds light on this to find out how the influence of
social media on herd behaviour in consumer's decisions are influenced by the popu-
larity of information.
To study popularity influence, Dewan et al. (2017) exploited an experiment in the

music context to investigate on the music consumption choices. They used popularity
feature in the experiment which reflected the total number of favourites for a song.
They allowed users to observe all other users’ music favouriting behaviour (anony-
mously) in the aggregate on a social networking platform. Clear evidence was found
that popularity influence is statistically and economically significant. Accordingly,
popularity information increases the number of listens for the average song by 12%,
and a 21% for narrow-appeal music. Further, they expect herd behaviour leading to
inequality in consumption (popular songs getting more popular, while unpopular
songs getting more unpopular) and to unpredictability of outcomes ('good' songs may
not become popular, while 'bad' songs may become viral hits).
Kuan et al. (2014) conducted a study in the group-buying context (such as Groupon).
They investigated the effects popularity by the number of people who have bought a
deal ('buy' information) and have used a joint approach of neuroscience and tradi-
tional self-information. Their results revealed a significant influence of popularity on
the buying intention but with asymmetric influence of positive and negative 'buy' in-
formation. While, positive 'buy' information (the deal was bought by more people than
expected) does not seem to increase the buying intention. However, negative 'buy'
information (the deal was bought by fewer people than expected) has a negative in-
fluence on buying intention.
Based on the results of a similar experiment and comparable results, Chen et al.
(2011) make a clear recommendation to report popularity statistics for mass-market
products, to make the best possible use of social media to influence consumer deci-
sions and consequently increase product sales.
88
__________________________________________________________________________
4.3.2.1 Information Quality
Besides the aspect of 'popularity', 'information quality' also seems to play decisive
role in the effect social media has on herd behaviour in the buying decision process.
There is empirical evidence of the influence of 'information quality' on behavioural in-

tention, which confirms the view that the content of online reviews plays an important
role in consumer decision-making (Pavlou and Dimoka 2006). People who read re-
views will carefully consider the opinions of other consumers about a product they
are considering to buy. People who do not read reviews will look for other clues that
signal the quality of a message. Prior studies have found that the quality of argu-
ments positively influences the information adoption (Zhang and Watts 2008) or the
buying intention (Park et al. 2007; Zhang et al. 2014). Quality is determined by vari-
ous characteristics, such as perceived informativeness, persuasiveness, relevance,
objectiveness, understandability, and sufficiency (Liu 2006; Park et al. 2007; Zhang
et al. 2014). Hirschmeier et al. (2016) compared different 'information quality' dimen-
sions against each other and discovered that believability and accuracy are the most
important dimensions for consumers throughout the buying decision process, while
completeness, timeliness, and amount of data were least important. Further, in many
studies, the length of a message was considered a proxy for information quality, with
longer messages being expected to provide high quality information on product char-
acteristics (Bosman et al. 2013; Cheng and Ho 2015; Huang et al. 2015; Mudambi
and Schuff 2010).
For example, Shen et al. (2014) used an existing book review site to investigate the
influence of information quality (and credence) on information adoption by incorporat-
ing the perspective of herd behaviour. Their results illustrate two major findings. First,
the quality of the arguments and the credibility of the source predict the adoption of
online reviews. Second, there is strong empirical support for the impact of herd fac-
tors, i.e. discounting one's own information and imitating others, represents a signifi-
cant influence on the acceptance of online reviews.
Zhang et al. (2014) found that argument quality is one of the key determinants of
consumers’ willingness to buying products. Meanwhile, Hirschmeier et al. (2016) in-
vestigated the perceived importance of information quality of UGC along the purchas-
ing decision process. The results showed that information quality requirements are
stable throughout the process and that users strongly prefer accuracy and credibility
89
__________________________________________________________________________
when using aggregated UGC from reviews, as opposed to timeliness, completeness

and quantity of data, which were rated as the least important information quality di-
mensions.
Matute et al. (2016) explored the characteristics of eWOM such as quality, volume,
and credibility as antecedents of online repurchase intentions. The results show that
only the quality of EWOM has a positive direct influence on consumers' intention to
buy again, while quantity has a negative influence. The credibility and quality of
EWOM also indirectly influence repurchase intentions through trust in the online mer-
chant.
Lastly, Tsao and Hsieh (2015) investigated how eWOM quality influences eWOM
credibility and buying intention. The results show that positive ratings of higher quality
can increase the credibility and thus buying intention of eWOM. Positive online re-
views with more detail, such as more photos and descriptions with greater specificity
and objectivity, allow readers to better perceive the value of the information conveyed
and therefore reduce uncertainty. This strengthens the users' confidence in the re-
view and increases their willingness to follow the proposed recommendations.
4.3.2.2 Valence
Besides information 'popularity' and 'information quality', there are additional metrics
proposed by researchers such as 'valence' (i.e. positive of negative reviews or rat-
ings) that affect the impact of social media on herding in buying decisions (Jabr and
Zheng 2014). The 'valence' of a review is an indicator that reflects the customers' atti-
tude towards a product (Liu 2006). Thereby, the 'valence' of the review in a rating
system is expressed in a form of star ratings that reflect the popularity and quality of
a product (Sundar et al. 2008). Extant research is ambiguous regarding the effects of
ratings (King et al. 2014), but the majority of literature to date have found that higher
star ratings are associated with more favourable impressions of products, increasing
the buying intention, e.g. for books, cell phones or movies (Chen 2008; Kim et al.
2015a; Sundar et al. 2008).
Customer online reviews can be both positive and negative. The literature indicates
that positive information can increase revenue by attracting new customers, whereas
negative information is considered a form of customer complaining behaviour (Huang
and Farn 2009). Therefore, it is often suggested in the literature that website opera-
tors should be extremely cautious about how consumers exert negative effects on
90
__________________________________________________________________________
their businesses through social media (Chiou and Cheng 2003). In addition, previous
studies of the literature on impression formation have also shown that people place
greater emphasis on negative information when comparing negative with positive in-
formation in product evaluation (Fiske 1980; Skowronski and Carlston 1989).
Lee and Youn (2009) examine the moderating role of the valence of eWOM in the re-
lationship between platforms and consumer product evaluation. The effect of eWOM
platforms on the willingness of consumers to recommend the product to friends was
found only if the review was positive. However, if the review was negative, there were
negative effects on consumers' willingness to recommend the product to friends in-
dependently of the eWOM platform.
Purnawirawan et al. (2014) investigated how consumers cope with online reviews
that are in conflict with each other (positive and negative). The results reveal a strong
presence of a consensus heuristic: a positive valuation balance creates a signifi-
cantly better review impression and a better intention to follow others in buy deci-
sions than a negative valuation balance.
Sun (2012) examines the informative role of product reviews in her study. In this con-
text, a highly positive average rating indicates a high product quality, while a high
variance of ratings is associated with a niche product, i.e. some consumers rate it
very positively and others very negatively. Due to the informative role of reviews, a
higher variance would correspond to a higher demand if the average review is low.
Similar results are shown by the work of Meiseberg (2014), according to which the
impact of the valuations on niche sales are striking. While positive reviews increase
the sale of poorly sold works, negative reviews, contrary to expectations, also in-
crease sales. In the case of niche products, which are usually found only through tar-
geted searches for correspondingly special interests, critical reviews lose their bite
('all publicity is good publicity') while positive ones obviously confirm the supposed at-
tractiveness of a good for consumers.
Furthermore, several researchers (e.g. Kunst and Vatrapu (2018), Kim et al. (2018))
have further refined the element 'valence' by including the direction of valence:
sidedness. While 'valence' provides an overall evaluation from a negative to a posi-
tive spectrum, advantages and disadvantages are also explicitly pointed out when
considering sidedness. Kim et al. (2018) deal with the question of whether two-sided
messages (advantages and disadvantages) have a different influence on the prob-
91
__________________________________________________________________________
ability of buying than one-sided messages (showing advantages or disadvantages).

They found a negative impact of the presence of cons. This implies that signalling
negative aspects of a product has a greater influence than signalling positive as-
pects.
4.3.2.1 Product Type
Extant research addresses another aspect related to the influence of social media on
herding. The question is whether the impact of herd behaviour differs between differ-
ent 'product types'. In the literature, various concepts are employed to differentiate
product types. For example, the differentiation of search and experience goods
based on how difficult it is to ascertain product quality before consumption (Hong and
Pavlou 2014; Nelson 1974). Experience goods are those whose product quality is dif-
ficult to determine before consumption, such as restaurants, hotels, and services.
Therefore, the descriptive information on the product side is often insufficient to de-
termine the quality of experience goods, and consumers look for external quality sig-
nals (e.g. ratings or reviews from other buyers or sales figures) for further evaluation.
In contrast, the quality of search goods (e.g. furniture, cars) is easier to determine be-
fore buying. The description on the product page often has sufficient information to
enable consumers to assess the product quality, additional signals are not necessar-
ily helpful (Huang et al. 2009; Nelson 1974). Additionally, there is a third type based
on quality assessment, whereby this is difficult to determine even after long use (e.g.
effectiveness of vitamins to health); this type of product is called credence goods
(Dulleck et al. 2011). In this case, eWOM or sales information play an even higher
role to consumers to reduce uncertainty about the product (Nakayama et al. 2010).
Although the categorisation of search and experience goods remains relevant and
widely accepted (Huang et al. 2009), researchers do not agree on their categorisa-
tions for products outside Nelson's original product list. The Internet is blurring the
boundaries between search and experience goods, and consumers use the opportu-
nity to read about the experiences of others and compare and exchange information
at low cost (Klein 1998; Weathers et al. 2007). Products can be described as existing
along a continuum, from pure search goods to pure experience goods.
Another perspective to investigate the product type differences is the extent of prod-
uct involvement, which relates more to the individual perception of consumers.
Zaichkowsky (1985) defines product involvement through the perceived relevance of
92
__________________________________________________________________________
the product by a person based on the inherent needs, values and interests. Product
involvement refers to the level of interest or attention that a consumer pays to a
product (Cohen 1983; Mitchell 1979).
With regard to the influence of social media on herd behaviour depending on the
product type, all product classifications are represented in the literature, but with an
emphasis on the concept of search and experience goods.
Previous research has shown that that consumers tend to make decisions based on
their private knowledge when evaluating search products; in contrast, consumers
would place more emphasis on the information derived from observing the actions of
others when evaluating experience products (Xiao and Benbasat 2011). For exam-
ple, Li and Wu (2018) found that herding is more salient for experience goods than
for search goods. Liu et al. (2016) had similar findings that demonstrate the informa-
tion cascades effect as more significant for online purchase of experience goods than
for search goods.
Nakayama et al. (2010) conducted a survey experiment to find out the influence of
different sources of information on the buying decision of different types of products
(search, experience, credence). The results indicate that the influence of eWOM
seems to be almost equal for search and experience goods, but different for the pur-
chase of credibility goods. Whereas expert opinions are more important than eWOM
for search and experience goods, but equal with eWOM for credence goods.
The study by Tsao and Hsieh (2015) investigated how the influence on consumers'
buying decision varies with the source of eWOM and the product type. The results
reveal that the persuasiveness of eWOM depends strongly on the product type. Re-
views that refer to credence goods are usually more convincing than those that refer
to search goods. Thus, consumers seem to rely more on opinions and recommenda-
tions when searching for products whose characteristics and quality are difficult to
evaluate. Consumers tend to rely on eWOM when searching for credence goods and
to strengthen their trust, especially when it comes to new or unknown brands.
Chen (2008) argues in his study that books, which he defines as low-involvement
products, and where consumers spend less time and cognitive efforts, are more likely
to elicit herd behaviour. Meanwhile, Gu et al. (2012) examine in their work the rela-
tive impact of different WOM sources on retailer sales for digital cameras as high-
involvement products.
93
__________________________________________________________________________
4.3.2.2 Homophily
Besides the previously outlined evidence from literature that several aspects influ-
ence consumer's decision decision-making behaviour towards herding and thus in-
fluences their purchasing decisions, 'homophily' also plays a role, particularly in the
context of social media.
"Similarity breeds connection. This principle — the homophily principle — structures

network ties of every type, including marriage, friendship, work, advice, support, in-
formation transfer, exchange, comembership, and other types of relationship"
(McPherson et al. 2001, p. 415). Hence, the concept of homophily, or similarity, is
central to the formation of a person’s social networks. Two aspects are decisive here:
status similarity, including the major socio-demographic characteristics, and value
similarity, representing the internal attitude, which determines the behaviour of a per-
son (McPherson et al. 2001). Therefore, homophily is different to pure peer influence
which does not consider the similarity aspect as such (Godinho de Matos et al.
2014). Peer-to-peer influence refers to the process through that a user can get his
network friends to make similar decisions, while the homophily process is described
by the fact that similarities between network contacts lead to similar decisions or imi-
tation takes place without any causal influence (Dewan et al. 2017).
Social networks open up new opportunities to obtain information about other people.
In fact, social observation has been identified as one of the most common social
network behaviours, which also includes research on profiles of social network con-
tacts (Joinson 2008). Thereby, profile information and status updates help determin-
ing the degree of similarity. Further, social networking contacts already possess
some degree of similarity, either status or value similarity, since users add acquaint-
ances to their social network (Tsai and Reis 2009). Hence, contacts from social net-
works are typically more similar than people in an online forum or a shop assistant
would be (Thelwall 2009). This similarity among users and their social network leads
to development of trust, as the other person seems familiar, and it is assumed the
person thinks, acts, and behaves in the same way. Hence, this phenomenon facili-
tates reduction of perceived risk and uncertainty, and therefore consumers are more
likely to trust in the recommendation from similar people (Matook et al. 2015).
Researchers address this aspect in their research and differentiate between the gen-
eral influence of social media on consumer decisions, and influence from 'similar'
94
__________________________________________________________________________
people. For example, Zhang et al. (2015) investigated in their study different types of
network with a different degree of preference heterogeneity. Friend-networks with low
heterogeneity, and on the other hand stranger-networks with higher heterogeneity.
Social networks allow consumers to easily observe the behaviours of both friends
and strangers. Their results show that, because of similar preferences in a friend
network, herding is more likely to happen there than in a stranger network. Whereas,
social networks composed of more strangers will be more likely to decide based on
the true product quality.
4.3.2.3 Others
In addition to the aspects outlined above for answering RQ1b, other factors are also
dealt with occasionally in the literature set, which are presented in the following.
Consumer Characteristics/Personality Type
Extant research discovered that people with different levels of expertise can be per-
suaded differently by external information. For example, Park and Kim (2008) found
that consumer expertise plays an important moderating role in determining the im-
pact of eWOM content on consumer buying decisions. Consumers with little expertise
are more likely to rely on external signals to process information and make buying
decisions. Brucks (1985) pointed out that a number of studies have found a negative
relationship between the amount of experience and the degree to which an individual
performs an external information search. Consumer involvement refers to the per-
sonal relevance of the product or service. Previous studies have already empirically
demonstrated the role of consumer participation in moderating the effect of eWOM
on consumer decision-making (Doh and Hwang 2009; Park and Kim 2008).
For example, Cheung et al. (2012) examined how consumer expertise and consumer
involvement, moderate consumer buying decision process, using panel data from a
forum. The results show that both consumer expertise and consumer involvement
play an important moderating role but in the opposite direction: while consumer ex-
pertise has a negative moderating effect, consumer involvement has a positive mod-
erating effect. Compared to less experienced consumers, those with a higher level of
expertise are less likely to be influenced by the opinions or actions of others. More-
over, the higher the involvement of the consumer in a brand community, the more
likely they are to follow the opinions of other members. These consumers also spend
95
__________________________________________________________________________
more time within a brand community and thus have more opportunity to observe the
behaviour and opinions of others.
Chen (2008) emphasise at the end of their article that it would be worth examining
the influence of personality traits on herd behaviour.
Strength of Ties
Another aspect that is addressed in the literature set affecting the influence of social
media, is the strength of ties, also described as friendship quality. Friendship plays a
crucial role in social commerce in today's societies (Li et al. 2018). Social media sup-
port social interaction and information exchange between members of social net-
works (either firms or individuals, or both) and help them buy and sell products and
services online (Xia 2013).
On a social networking platform (e.g. Facebook, LinkedIn) every consumer can form
a social network with three levels of tie strength: strong, weak and potential. "Tie
strength is a multidimensional construct that represents the strength of dyadic inter-
personal relationships in the context of social networks, including closeness, inti-
macy, support, and association" (Shen et al. 2016, p. 2266). Kavanaugh et al. (2005)
define strong ties as members with a high degree of trust and intensive daily contacts
within an exclusive community. In contrast, weak ties are characterised by a low de-
gree of trust, less information exchange. Grayson (2007) provides a similar classifica-
tion: close friend, acquaintance, not knowing the person at all, which corresponds to
the three levels of attachment. All these bonds maintain different levels of friendship
quality.
Li et al. (2018) confirm with empirical evidence that friendship quality has positively
effect on buying intention. Good friends with a strong tie have a higher probability to
influence their friends, particularly with high-priced or risky products. However,
friends with a weak tie are not as attractive as strangers who have good user ratings.
Simplified, it can be said that people are more likely to follow the information that
comes from reliable strangers than from simple friends.
Kwahk and Kim (2017) elucidate in their study that the strength of social ties encour-
ages members to openly share their shopping experiences and product reviews.
When more trusted members use the social media of e-commerce sites, consumers
are more likely to rely on the product information. Their results show that frequent in-
teractions between members enhance the strength of the relationships between them
96
__________________________________________________________________________
and therefore shared information is more accepted, so online retailers are trying to
encourage consumer participation in social media.
Social Media Activity
Several authors also shed light on the impact of social media usage on consumers'
decision behaviour. For example, Vithayathil et al. (2017) investigate how the use of
popular social media is associated with shopping preferences. They found that the
frequency of usage of the different social networks (Facebook, Skype, Twitter,
LinkedIn) was positively associated with the customers' preference regarding their
favourite store.
Matook et al. (2015) consider in their study the posting behaviour of users, and inves-
tigate how this aspect influences the intention to act on a recommendation. Postings
in social networks can be of different types: e.g. status updates, comments, product
recommendations. These postings are crucial for the functioning and maintaining the
attractiveness of social networks. However, there is also the phenomenon of exces-
sive posting, in which case the frequency exceeds the normal level and irrelevant in-
formation is also posted. Matook et al. (2015) found out that exactly this behaviour of
excessive posting negatively affects trust in recommendations. Thus, they conclude
that less excessive posting behaviour may be more effective to build trust and influ-
ence others with regard to their product choices.
Li and Wu (2018) provide a few alternative explanations what drives herding driven
by social media, e.g. the saliency effect which can create biases. Consumers often
choose the prominently displayed products (Cai et al. 2009), especially if the entire
selection set is not observable. Instead of deciding on the basis of information such
as eWOM or sales figures, consumers can choose a product solely because of its sa-
liency (Cai et al. 2009). Another factor could be overcrowding which can inhibit herd
behaviour. For example, high sales volume my indicate overcrowding, which can re-
duce product quality, especially when the expiration date for the product is coming up
(Horton 2014). Further, providing past sales information may also fail to trigger a
herding response when consumer preferences are dissimilar to those of typical
mass-market consumers.
97
__________________________________________________________________________
4.4 Summary and Interim Discussion
In order to provide a deeper understanding on the state of research with regard to

herd behaviour in consumers' buying decisions so as the role of social media within
this phenomenon a systematic literature review was conducted. This followed the
guidelines for a systematic literature review according to Webster and Watson (2002)
along a four step approach. Thereby, 48 scholarly articles were subject to analysis.
The derived descriptive statistics results are presented in section 4.2, followed by the
qualitative analysis of the existing academic body of knowledge along the two re-
search questions (RQ1a, RQ1b). The respective results are presented in section 4.3.
In the following, the main insights gathered from the previous sections are summa-
rised and discussed.
First, the quantitative analysis of the identified publications in the timeframe from
2008 to 2018 has revealed the following:
 A closer look at the course of time shows showed that the topic became in-
creasingly more relevant over time, an almost fourfold increase from 2008 to
2018 may be interpreted as an increasing interest among researches in the
topics. Especially, from 2014 onwards the number of hits has risen even fur-
ther.
 The classification of the publications into research areas shows that the topic
is reflected in five research areas, where 'IS' research is dominant with ~54%,
and is followed by 'Economics' (~ 21%) and 'Marketing' (15%).
 Analysing the methods applied, the results revealed seven classifications:
'survey', 'experiment', 'mathematical modelling', 'quantitative analysis', 'litera-
ture review', 'conceptual', and 'multiple case study', with 'survey' (~30%) and
'experiments' (~26%) as the most popular methods, followed by an increased
popularity of 'mathematical modelling' (~20%) particularly in the time period of
2014–2018.
 The developed overview of the applied theories shows in total seven recog-
nised theories, which are applied several times. Thereby, the top three theo-
ries are: 'information cascade theory' (19%), 'social influence theory' (19%), ,
and 'herding theory' (16%). Under 'others' theories are summarised which oc-
cur only once in the analysed literature set. In the second time frame the vari-
ety of applied theories has almost doubled.
98
__________________________________________________________________________
 The analysis in the scope of the first research question (RQ 1a) revealed a to-
tal of three main topic clusters: 'information type', 'information source', 'tech-
nology adoption'. Thereby, the treatment of the topic 'information type' in the
literature has more than doubled in the period between 2014 and 2018, while
the other two topic blocks 'information source' and 'technology adoption' have
been researched in the literature consistently over both time periods.
 The results with regards to the second research question (RQ 1b), summa-
rises a total of six topic clusters: 'popularity', 'information quality', 'valence', ,
'product type', 'homophily', and 'others', with 'popularity' and 'information qual-
ity' being the most frequently discussed topics, especially in the second time
period.
The following Figure 25 summarises the results of RQ 1a, and RQ 1b.
Addressed Topics in Literature (2008 – 2018) with regard to

Herd Behaviour in Buying Decisions
General Topics: Influencing aspects of Social

Media:
• Information type
• WOM • Popularity
• eWOM
• System recommendation • Information quality
• Buy-/sales information • Valence
• Information source • Product type
• Peer groups • Homophily
• Experts
• Crowd • Others
• Technology adoption
Figure 25: Summary of the Qualitative Literature Review Results
Then, the qualitative analysis of the literature set with regards to the first research
question (RQ 1a) revealed the following results:
 A total of three main topic clusters could be identified: 'information type', 'in-
formation source', and 'technology adoption'.
99
__________________________________________________________________________
o In the first cluster different 'information types' are summarised, these

are: 'WOM', 'eWOM', 'system recommendations', and 'buy/sales infor-
mation'. In this regard, most of the research deals with the questions
which information type has a greater impact on herd behaviour within
purchasing decisions. To this purpose, researchers compare the impact
of the different Information types', study their joint impact on consumer
buying decisions, and elaborate when users discount their own informa-
tion when faced with signals from others.
o The second cluster 'information source' subsumes: 'peer groups', 'ex-
perts', and 'crowd'. Extant research focus on the questions which
source of information ultimately has the greatest influence on consum-
ers decision behaviour. But the picture does not seem so unambiguous.
The researchers present a differentiated picture here, which seems to
depend on other factors, e.g. product type, popularity, and demograph-
ics.
o The third cluster 'technology adoption' deals primarily with the complex-
ity and uncertainty of technology decisions in the context of observing
and utilising the behaviour of others. Furthermore, the extant research
deal with the phase after the technology has been adopted, and therein
the effect of herd behaviour.
Finally, the second part of the qualitative analysis of the 48 articles with regards to
the second research questions (RQ1b) revealed the following results:
 As a result, a total of six topic clusters could be summarised: 'popularity', 'in-

formation quality', 'valence', 'product type', 'homophily', and 'others'.
o Popularity refers to the total number of favourites from the community
as a whole, which means this is about the volume of information (e.g.
number of likes, sales volume). Existing research deals with the ques-
tion how the influence of social media on herd behaviour in consumer's
decisions are influenced by the popularity of information.
o There is empirical evidence of the influence of 'information quality', de-
termined by various characteristics, on behavioural intention. Therefore,
researchers investigate the influence of information quality on consum-
ers' decisions by incorporating the perspective of herd behaviour. Most
results show that information quality reduce uncertainty, and thus, in-
100
__________________________________________________________________________
creases the willingness of users to follow the proposed recommenda-

tions of others.
o Researchers discuss how 'valence' (i.e. positive of negative reviews or
ratings) affects the impact of social media on herding in buying deci-
sions. The extant research is ambiguous regarding the effects of rat-
ings, but the majority of literature to date have found that higher star rat-
ings are associated with more favourable impressions of products, and
increased buying intentions.
o The reviewed studies address the difference in 'product type' as an-
other aspect related to the influence of social media on herding.
Thereby, various concepts to differentiate product types are used in lit-
erature but that the type has an influence is proven by different re-
searchers.
o 'Homophily' is described by the fact that similarities between network
contacts lead to similar decisions or imitation takes place without any
causal influence. Thus, researchers address this aspect and differenti-
ate between the general influence of social media on consumer deci-
sions, and influence from 'similar' people. They show that herding is
more likely to happen there than in a stranger network.
o Under 'others', the topics 'consumer characteristics', 'strength of ties',
and 'social media usage' are summarised. The influence of these topics
on the effect of social media on herd behaviour in purchasing decisions
is only mentioned sporadically in the literature and therefore does not
seem to have been extensively researched yet.
Overall, the results of the literature review show important aspects with regards to
herd behaviour in the purchasing decision and the role of social media therein. For
example, Li and Wu (2018) or Jabr and Zheng (2014) studying the impact of different
'information types' on herding. Sun (2012) investigating the influence of 'valence' or
Shen et al. (2014) the one of 'information quality'. Nevertheless, most studies ad-
dress only single aspects, e.g. the impact of the 'information type', 'information
source', and 'popularity'. In addition, the entire buying decision process is only very
rarely considered (e.g. Hirschmeier et al. (2016)). Instead, most studies focus on the
buying intention (see, e.g., Kwahk and Ge (2012), Purnawirawan et al. (2014), Zhang
et al. (2014), and Matook et al. (2015)).
101
__________________________________________________________________________
Furthermore, the existing research does not cover some aspects in detail, e.g. social
media usage or consumers characteristics and their influence in the discussed con-
text of herd behaviour in purchasing decisions. Some researchers state that further
research on these topics would be valuable. In summary, it can be concluded that
single aspects concerning herd behaviour in the buying decision process have been
researched. However, there is a missing comprehensive view that brings together the
different aspects along the entire buying decision process. Especially considering the
aspects related to the influence of social media therein.
102
Study II: Social Interactions on Different Social Media Platforms
__________________________________________________________________________
5 Study II: Social Interactions on Different Social Media

Platforms
Chapter overview: This chapter presents the results of an explorative study on so-
cial interactions on different social media platforms in the context of the mobile com-
munications industry. In section 5.1, the specific objectives and the research ques-
tions are outlined and put into context with existing research on the discussed topic.
Section 5.2 discusses the theoretical background on social media classification and
social interactions on social media. The research methodology, including the empiri-
cal field, data collection, and data analysis, are presented in section 5.3. The evalu-
ated results are presented in section 5.4. Finally, section 0 concludes this study, and
provides an interim discussion of the results.
PART A: Foundations
1 Introduction
3 Research Design
Study II: Social Interactions on Different Social Media

5 Platforms

6 Media
PART C: Summary
9 Conclusion
Figure 26: Dissertation Overview - Chapter 5: Study II
103
__________________________________________________________________________
5.1 Objectives of the Study
In the time of 'always on' with everyone using the Internet and being connected
through social networks, information and opinions are shared very quickly. Young
people especially are using social media to a high extent, whereby more than 50% of
them do so daily or several times a week (mpfs 2018). The increased proliferation of
social media has expanded traditional interpersonal communication in the form of
electronic word of mouth (eWOM) (Chen et al. 2011; Godes et al. 2005). The given
anonymity through the Internet is considered as an advantage of eWOM, since it
supports users to share opinions more comfortably. Given that people rely on the
opinions of others, individual behaviour may be influenced by other members of the
community (Zhou 2011).
Furthermore, social media is changing communication and behaviour patterns and

confronting firms with completely new requirements. For one, there are new opportu-
nities for firms to use social media to foster knowledge sharing within the organisa-
tion and thus to benefit from people's metaknowledge (knowledge of 'who knows
what' and 'who knows whom') (Leonardi 2015; Majchrzak et al. 2013). Additionally,
information asymmetries between suppliers and consumers disappear or are re-
versed, giving networked consumers power (Akehurst 2008). Firms and organisa-
tions have realised that the growth and popularity of social media and their influence
on consumers' decisions have become a powerful tool for marketers. They can pro-
mote their products and consumers share their opinions. Therefore, it becomes in-
creasingly crucial for firms to develop an understanding of social media platforms and
the interactions on them so that they can select the most appropriate platforms and
consider them as a vital factor in their social media strategies (Hanna et al. 2011). It
is relevant to understand the differences between various platform types and what
people do there. The high level of transparency provided by social media platforms
poses a particular challenge to industries that offer commodity products, as they
cannot offer a high level of product or price differentiation (Li and Whalley 2002). In
this study, the mobile communications industry is chosen as the empirical context.
Social media is a medium that is masspersonal, which means that they allow users to
broadcast interpersonal communication to many people (Bayer et al. 2020).
Furthermore, it may have different degrees of interactivity, and to understand this it is
crucial to know the contingency, context and structure, goals, sequences of actions
104
__________________________________________________________________________
and reactions, and characteristics of the medium (Stewart and Pavlou 2002). Kaplan
and Haenlein (2010) describe social media as a group of Internet-based applications
allowing the creation and exchange of user-generated content (UGC). They have
developed a classification scheme along two axes: social presence and self-
presentation. Peters et al. (2013) enhanced this structuring and developed a holistic
framework, which is essentially based on four elements: motives, network structure,
content, and social interactions.
With regard to the influence on consumer buying decisions and therefore the impor-
tance for B2C firms, the elements content and social interaction seem to play a major
role. Thereby, the relative effectiveness of different social media types and content
characteristics on the buying process has been explored by many researchers in dif-
ferent industries, such as in electronics, insurance, automobiles, and groceries (Maas
et al. 2014; Park et al. 2014; Powers et al. 2012; Zhang et al. 2014). The major influ-
ence on the buying process seems to come from rating/review platforms, forums, and
social networking sites (Maas et al. 2014). Considering social interactions, Peters et
al. (2013) proposed four social interaction categories (sharing, gaming, expressing,
and networking); however, there is scarce research exploring deeper social interac-
tions on different social media platforms. Thus, the aim of the current study is to
elaborate on this issue and provide novel and intriguing categories to facilitate future
studies on social interactions. Building on these thoughts and the mobile communica-
tions industry context, the present study aims to answer the following research ques-
tions:
RQ 2a How do social interactions differ depending on different social media plat-

forms?

In order to approach these research questions, this study will present results based
on data of three weeks from three different social media platform types (social net-
work, forum, review/rating platform). The considered content of all three platforms re-
lated to telecommunication topics and thereby to one provider. Already more than 50
years ago, Webb et al. (1966) identified that machine readable and archived data of-
105
__________________________________________________________________________
fer the possibility to use 'unobtrusive' or 'non-reactive' instruments. People are usu-
ally not aware when their online traces are investigated, whereby one can assume by
using non-reactive data collection that people’s behaviour and conversations are
natural and not affected by the data collection procedure (Fritsche and Linneweber
2006; Webb et al. 2000). Then, the analysis of the archival data followed the qualita-
tive content analysis approach according to Mayring (2015). Since the research
questions aim to produce novel and intriguing categories the inductive procedure is
used in the study at hand.
5.2 Theoretical Foundation
5.2.1 Social Media Classification
Online communities have become very popular for research since Web 2.0 technolo-
gies have triggered their rapid development (Preece and Maloney-Kirchmar 2005).
Online communities (or virtual communities) are studied by different disciplines, and
therefore definitions, classifications, and characteristics differ. Already in 1993, online
communities were described as 'social aggregations' that arise when people discuss
publicly and build personal relationships online (Rheingold 1993). In order to provide
a common understanding for information systems (IS), Lee et al. (2003) compared
nine of the most popular existing definitions and defined an online community as a
“cyberspace supported by computer-based information technology, centred upon
communication and interaction of participants to generate member-driven contents,
resulting in a relationship being built” (Lee et al. 2003, p.57).
In 2005, a new type of online community arose, namely, social networks. These are
online communities that take advantage of new and improved social computing tech-
nology for interaction and multimedia sharing (Iriberri and Leroy 2009). In 2019, there
are over 3 billion social media users in the world — about 40% of the global popula-
tion (DataReportal 2020a). Like social networks, social media is an evolution of
online communities. Social media combines two related concepts: Web 2.0 and
UGC. Web 2.0 represents the technological foundation and UGC is published con-
tent that is publicly accessible. The content needs to be created outside of profes-
sional routines and represents a certain effort (Kaplan and Haenlein 2010). Accord-
ing to Kaplan and Haenlein (2010, p. 61), "social media is a group of Internet-based
applications that build on the ideological and technological foundations of Web 2.0,
and that allow the creation and exchange of User Generated Content". Mangold and
106
__________________________________________________________________________
Faulds (2009), Kaplan and Haenlein (2010), and Maas et al. (2014) subsume various
types under social media using the two dimensions: self-presentation/self-disclosure
and social presence/media-richness. This framework is introduced in chapter 2.3.3.
Figure 27 shows the corresponding classification of different social media types in
the framework.
High
Social
Virtual social
networking
sites worlds
Self-presentation / Self-disclosure
(e.g., (e.g., Second

Blogs Lif e)
Facebook)
Rating/Review
platf orms Forums
(e.g., Check
24)
Content Virtual game

Collaborative worlds
projects communities (e.g., World of
(e.g., (e.g., Warcraf t)
Wikipedia) YouTube)
Low
Social presence / Media richness
Low High
Figure 27: Classification of Social Media according to Kaplan and Haenlein (2010) and Maas
et al. (2014)
5.2.2 Social Interactions on Social Media
To further develop the classification according to Kaplan and Haenlein (2010), Peters
et al. (2013) proposed a holistic framework that covers the major elements of social
media: motives, network structure, content, and social interactions. Their social
media framework is described in the following.
First, the communication of each user is driven by specific motives like social, cul-
tural, or intellectual value. Second, the network structure that defines the underlying
infrastructure (size connections, distributions, segmentation) for each social medium.
Third, actors communicate with each other and produce user-generated content
(UGC). Thereby, content may have three sufficiently distinct aspects: valence sub-
suming tonality (e.g. positive, negative) and emotions, volume described by counts
107
__________________________________________________________________________
and volumes, and quality characterised by interactivity and vividness. Thus, content
is an input as well as an output to social interactions. Finally, users do not only pro-
duce content as they communicate, but they also interact with each other in a variety
of ways, as they modify, share, comment or simply consume the content.
There is a considerable amount of research examining the different aspects of con-

tent (valence, volume, quality). Valence is mostly associated to review-/and rating
platforms, as star ratings reflect a positive or negative average picture and so the
popularity of a product (Peters et al. 2013). The majority of previous studies have
found that that a higher average star rating is associated with more favourable im-
pressions of products, which increases buying intention (Chen 2008; Kim et al.
2015a). Volume represents the amount of information, and in the social media con-
text, is often associated with social networks and the number of likes (Peters et al.
2013). Volume plays an informative role by increasing the degree of consumer
awareness and is therefore often researched with regards to consumption choices.
For instance, Dewan et al. (2017) explore how the influence affect music consump-
tion choices. Quality is mostly associated with forums since they allow a high level of
interactivity and information sharing to specific topics (Peters et al. 2013). Previous
research has found that argument quality positively influences information adoption
buying intention (Park et al. 2007; Zhang et al. 2014). For instance, Zhang et al.
(2014) found that argument quality is one of the key determinants of consumers’ will-
ingness to buy products.
With respect to social interactions, they are positioned on the intersection of the other
three elements of the social media framework (motives, network structure, content)
(cf. section 2.3.3) and are therefore of special interest for researchers (Peters et al.
2013). It seems to be a given that consumers tend to be influenced by their social in-
teractions with others when they make buying decisions because they can learn from
and be affected by other consumers' opinions and/or others’ actual buying decisions
(Babić Rosario et al. 2016; Chen et al. 2011; Godes et al. 2005). Furthermore, a
suggestion of characteristics of social interactions (e.g., solidarity, antagonistic,
mixed, intensity, extension, duration, organisation) can be found, but no explicit clas-
sifications. Therefore, Peters et al. (2013) concluded based on extensive literature
review and practitioner analyses that there are four social interactions dominant in
social media: sharing, gaming, expressing, and networking. They encourage further
research on social interaction categories and how this link to content type to social
108
__________________________________________________________________________
media type. Little additional literature can be found on elaborating social interaction
categories, and it seems to be an underexplored research field. Thus, the aim of the
current paper is to elaborate on an understanding of social interactions connected to
different content types represented by three different social media platforms in rela-
tion to the mobile communications industry.
5.3 Methodical Approach
Social media can be studied with various reactive and non-reactive methods. Among
the reactive methods, online surveys are very common, while non-reactive methods
include observation, content analysis, log file analysis and text mining (Aggarwal and
Zhai 2012; Fielding et al. 2017; Hewson 2017). The defining criterion for non-reactive
data collection is not a feature of the method as such, but non-awareness of the data
collection process on the part of the subjects under investigation (Janetzko 2017). In
order to gain data from unaffected conversations to allow a deeper understanding of
people's interactions the research process followed an explorative approach.
5.3.1 Empirical Field
Online research has established itself as an innovative field and although there is no
agreed terminology, "scientists have reached a common ground on the most impor-
tant methods and instruments of online research" (Welker et al. 2014, p.14). The
connotation online research refers to the Internet as the object and method, with
strong methodological reference, academic, practical and commercial-oriented
(Welker and Wünsch 2010). The generated data allows rich insights into individuals’
perspectives, interpretations, and constructions of meaning, and thus is also less af-
fected by issues relating to self-reported biases like social desirability biases
(Hewson 2017).
Social media platforms offer the possibility to observe human unaffected conversa-
tions and use a high amount of accessible non-reactive and unobtrusive data
(Janetzko 2017). Since in the present study the mobile communications industry is
selected as the empirical context, the content is related to telecommunication topics.
To be able to infer something about the differences in people's interactions, elements
like the related organisation (o2) and the time period of the retrieved data were kept
constant.
Based on the findings of previous research, social networking sites, forums, and rat-
ing/review platforms have the greatest influence on purchasing decisions (Kim et al.
109
__________________________________________________________________________
2018; Kwahk and Kim 2017; Maas et al. 2014; Zhang et al. 2014) As previously dis-
cussed, there is a considerable amount of research examining the impact of social
media regarding different aspects of content such as valence, volume, and quality.
Furthermore, Peters et al. (2013) have outlined the relevance of further research on
social interaction categories and how this link to content type to social media type.
Therefore, one representative was selected for each category with regards to social
media type and content characteristic as defined by Peters et al. (2013). 'Check24' is
one of the most popular rating/review platforms (Lammenett 2017) in Germany, giv-
ing consumers the opportunity rate their experiences with services and products (va-
lence). 'Facebook' as one of the most popular social networking sites (mpfs 2018)
and its content is associated with volume. Finally, one of the largest independent
telecommunication forums 'Inside-Handy' covers the content aspect of quality.
5.3.2 Data Collection
After having selected the empirical field, the data was collected. The availability of
the material in a digital form facilitated the access to the data. To collect data from
Facebook, the author used their own existing account, but registration was necessary
to access the Inside-Handy forum and Check24. To overcome the volatility and dy-
namics of online content, researchers recommend to either use a static snapshot or
analyse data from a certain time period (Souza de and Preece 2004). The latter has
been applied for the study at hand, whereby data has been retrieved on a daily basis
for the same time period from all three platforms as a snapshot.
The data collection had the goal of providing the data for the content analysis. For
this purpose, user postings across three weeks were collected. On Facebook, the
chronicle from the o2 page was used as the source. On the forum Inside-Handy, the
top ten topics discussed on o2 were used as the data source, and on Check24 rat-
ings and comments related to the provider o2. Welker and Wünsch (2010) summa-
rise different specifics that should be considered for 'online content' and were taken
into account within this study. One is that it needs to be defined how to deal with me-
diality or multimedia. Hence, auditory and visual content was excluded as the basis
for the analysis is written textual content. Specific attention should be paid to the non-
linearity and hypertextuality of online content: to counteract this condition, the col-
lected data excluded content from further link references. Furthermore, to avoid and
minimise effects from personalisation and reactivity, only the locally archived data
was used as basis for the analysis.
110
__________________________________________________________________________
An observational protocol supported the recording procedure. Following the guide-

lines of Creswell (2009) and Flick et al. (2012) a scheme was developed to structure
the descriptive notes. This includes the following structuring elements: appearance,
number of users, general description of community, target of the community, charac-
ter of contributions and other special features. The used observation protocol can be
found in Appendix C.1.
5.3.3 Data Analysis
The subsequent content analysis should answer the research questions concerning
social interactions differences depending on different social media platforms and
specifics in relation to the mobile communications industry. Therefore, the collected
data was analysed by applying qualitative content analysis as a scientific research
technique.
The methodology of the online content analysis is based on the 'classical' content
analysis and is also gaining relevance for online generated content (Welker et al.
2014). In order to ensure a rigorous interpretation of the material, Mayring (2015) and
Krippendorff (2004) developed clear rules and systematic procedures to analyse the
data. Accordingly, content analysis is a widely accepted method in IS research and is
applied in many studies (Myers and Avison 2002). Since the research question aims
to produce novel and intriguing categories to facilitate future studies, the inductive
procedure is chosen which means that categories are identified directly from the data
(Mayring 2015). All user postings and conversations between users were determined
as a relevant material for the analysis.
To perform the qualitative content analysis, the procedure model according to May-
ring (2015) outlined in section 3.2.1.3 was applied. To facilitate the subsequent proc-
ess of analysis, all collected data was imported into MAXQDA 12 software.
Accordingly, after the first round of category development, 43 categories had

emerged. This was followed by a review of the developed category system on the
raw material, verification to the desired level of abstraction, and subsequent refine-
ment of the categories for all three social media platforms. The developed categories
were reworked where needed, whereby similar categories were grouped, and the
wording was unified and categories were arranged according to eight main catego-
ries.
111
__________________________________________________________________________
To accomplish the reliability check by means of intercoder reliability (Mayring 2015) a

second researcher familiar with the research area was introduced to the category
system (Appendix C.2) and the researcher coded a third of the randomly selected
data. The calculated Holsti coefficient of reliability was 0.82.
After the qualitative content analysis, the developed categories were quantified and
compared. First, the frequency distribution for the main categories and the subcate-
gories was calculated to gain an overview of their overall importance. Subsequently,
the importance of the categories was compared between the platforms. For this pur-
pose, it was necessary to normalise the data over the three platforms. Finally, a
cross-tabulation was calculated to gain deeper insights into whether the importance
of the subcategories differ within the main categories and what the distributions of the
three platforms looks like.
5.4 Results
The findings regarding the research questions of how social interactions differ de-
pending on different social media platforms and how they can be specified in relation
to the mobile communications industry, are presented in the following. But first, the
investigated platforms are briefly described.
The social network presence of the mobile provider aimed to use the channel to
communicate with customers and interested people to share information and place
advertisements which were 'liked' by the users. Usually, conversations took place be-
tween the provider — who took a moderator role — and the users, rather than be-
tween the users themselves. On the forum, users' conversations usually took place
between users about telecommunication topics in a prevailing good atmosphere. The
main target was to provide an independent exchange platform for interested people,
and the topics were freely selectable by the users. Multimedia content was not inte-
grated in the community and advertisements were restricted. On the review/rating
platform, the main goal was to give the possibility to rate one’s own experiences and
make the reviews public. Specifically, registered users were allowed to award stars
about telecommunication providers and their products. Multimedia content and ad-
vertisement were not integrated in the rating part of the platform. Conversations be-
tween the users were not possible.
The qualitative content analysis and the subsequent quantification of the derived
categories elaborate further differences. The descriptive results of the content analy-
112
__________________________________________________________________________
sis revealed sixteen subcategories grouped into eight main categories. The main
categories are more general with regard to people interaction, whereby the subcate-
gories are more specific to telecommunication topics. Not all categories were relevant
in every investigated platform and their importance differed. The main categories
have evolved answering the research question concerning social interactions differ-
ences depending on different social media platforms related to the mobile communi-
cations industry.
The frequency distribution of the main categories shows the comparison of the impor-
tance of the identified social interactions across the platforms for the observation pe-
riod. The comparison reveals that on the review/rating platform, only 'satisfaction' and
'dissatisfaction' were shared (Table 10), as is given for this kind of platform. The next
category of 'comment and opinion' subsumes user interactions by briefly commenting
on content or writing their opinion in detail, it seemed balanced between the social
network and the forum. The category 'reference' was dominant at the forum, and
summarises user interactions by referencing other websites, products or other con-
tent. In contrast to 'support' which seemed more important on the social network plat-
form. 'Regret' was only represented on the social network while shared 'experiences'
were dominant in the forum.
Table 10: Frequency Distribution of Social Interactions across Platforms (Main Categories)
Platforms Social net- Forum Re- Total

work view/rating
platform
Main Categories in %
Satisfaction 18.2% 8.9% 72.8% 100.0%
Dissatisfaction 30.0% 3.7% 66.3% 100.0%
Comment and Opinion 41.2% 58.8% 0.0% 100.0%
Reference 38.2% 61.8% 0.0% 100.0%
Support (asked/offered) 62.7% 37.3% 0.0% 100.0%
Recommendation (ad-
vise/disadvise) 0.0% 100.0% 0.0% 100.0%
113
__________________________________________________________________________
Platforms Social net- Forum Re- Total

work view/rating
platform
Main Categories in %
Regret 100.0% 0.0% 0.0% 100.0%
Experience 17.5% 82.5% 0.0% 100.0%
The subcategories are more specific to the mobile communications industry and de-
scribe what kind of interactions take place relating to telecommunication topics and
their overall importance. The subcategories show that the importance of the topics
strongly differed between the platforms (Table 11). While 'mobile network'- and 'cus-
tomer service'-related interactions are predominant on the social network and the re-
view/rating platform, 'tariff offer'-, 'apps'- and 'credit check'-
related topics are more relevant on the forum.
Table 11: Frequency Distribution of Social Interactions related to Mobile Communications Top-
ics across Platforms (Subcategories)
Platforms Social net- Forum Review/ Total

work rating pl.
Subcategories in %
Mobile network 55.5% 2.9% 41.7% 100.0%
Tariff offer 22.6% 62.8% 14.6% 100.0%
Customer service 18.8% 3.4% 77.8% 100.0%
Price 9.4% 27.1% 63.5% 100.0%
Provider 87.7% 12.3% 0.0% 100.0%
Apps 29.2% 70.8% 0.0% 100.0%
Credit check 0.0% 67.5% 32.5% 100.0%
Customer retention 42.5% 0.0% 57.5% 100.0%
114
__________________________________________________________________________
Platforms Social net- Forum Review/ Total

work rating pl.
Subcategories in %
Number porting 0.0% 44.4% 55.6% 100.0%
Mobile phone 0.0% 100.0% 0.0% 100.0%
Purchasing channel 0.0% 100.0% 0.0% 100.0%
Operating system 0.0% 100.0% 0.0% 100.0%
Advertisement 100.0% 0.0% 0.0% 100.0%
Contract termination 69.9% 30.1% 0.0% 100.0%
Entertainment (music, games,

gadgets) 100.0% 0.0% 0.0% 100.0%
Handling and processing 0.0% 0.0% 100.0% 100.0%
Subsequently, the importance of the categories was compared between the plat-
forms. For this purpose, it was necessary to normalise the data over the three plat-
forms. A cross-tabulation on the top five main and subcategories allowed an over-
view of the interrelatedness between them and their distribution on the three plat-
forms (Table 12).
The results illustrate that 'dissatisfaction' and 'satisfaction' stemmed predominantly

from 'mobile network'- and 'customer service'-related topics, whereby 'dissatisfaction'
about 'mobile network' seem to be an interaction only taking place on social net-
works. Whereas 'customer service'-related interactions predominantly take place on
the review/rating platform.
On the contrary, the category 'satisfaction' related to 'mobile network' was predomi-
nantly represented on the review/rating platform and not at all on Facebook. Fur-
thermore, the category 'satisfaction' in relation to 'customer service' was only repre-
sented on the review/rating platform.
'Comments and opinions' were shared most about 'offers' and 'mobile phones' and
predominantly on the forum. 'Support' was relevant for all top five subcategories but
115
__________________________________________________________________________
most with regards to 'mobile network', 'provider', and 'mobile phone'. The social net-
work platform was well represented in all combinations except 'support' regarding
'mobile phone', where the forum was the dominant platform. The category 'experi-
ence' emerged most commonly when concerning 'tariff offer', 'mobile network', and
'mobile phone' and primarily on the social network and the forum.
Table 12: Cross-tabulation: Top 5 Main Categories and Top 5 Subcategories
Main Dissatis- Satisfac- Comment Support Experi-

categories faction tion and (asked/of ence
Opinion fered)
Subcategories
Mobile network 27.8% 25.2% 1.0% 25.6% 10.9%
Social network 61.1% 0.0% 100.0% 100.0% 10.0%
Forum 2.2% 3.6% 0.0% 0.0% 0.0%
Review/rating platform 36.7% 96.4% 0.0% 0.0% 0.0%
Customer service 21.3% 14.9% 0,0% 10.2% 0.0%
Social network 10.6% 0.0% 0.0% 100.0% 0.0%
Forum 5.7% 0.0% 0.0% 0.0% 0.0%
Offer 3.7% 10.4% 14.3% 10.9% 11.6%
Social network 30.7% 55.6% 14.4% 11.7% 18.8%
Forum 0.0% 8.6% 85.6% 88.3% 81.2%
Mobile phone 0,0% 3.6% 14.4% 15.2% 7.1%
Social network 0.0% 0.0% 0.0% 0.0% 0.0%
Forum 0.0% 100.0% 100.0% 100.0% 100.0%
116
__________________________________________________________________________
Main Dissatis- Satisfac- Comment Support Experi-

categories faction tion and (asked/of ence
Opinion fered)
Subcategories
Provider 8.6% 10.9% 1.0% 16.7% 4.4%
Social network 85.8% 83.6% 100.0% 91.8% 100.0%
Forum 14.2% 16.4% 0.0% 8.2% 0.0%
This study has sought to explore how social interactions differ depending on different
social media platforms in the empirical context of one mobile communications indus-
try, referring to one mobile provider. Considering the actual and future trends in the
environment of social media and its underlying interaction data, there is a major need
for understanding the usability of social media interactions.
The main contribution of this study stemmed from the qualitative content analysis
based on UGC from three social media platforms. Eight main categories with regards
to general social interactions and sixteen underlying subcategories specific to mobile
communications topics were revealed. These categories can enrich previous re-
search on social interaction. Peters et al. (2013) consolidated four social interactions
that are dominant on social media: sharing, gaming, expressing, and networking. The
eight identified categories of 'satisfaction', 'dissatisfaction', 'comment and opinion',
'reference', 'support', 'recommendation', 'regret', and 'experience' enrich the social in-
teractions of sharing and expressing and provide a deeper understanding of them.
Furthermore, the link of social interactions to different social media types offers a dif-
ferentiated perspective on how social interactions differs depending on the social
media type. It turns out that on the social networking platform, social interactions
'support', 'comment and opinion', 'regret' and 'dissatisfaction' take precedence.
Meanwhile, the focus on the forum is more on 'recommendation', 'experience', and
'reference'. On the review/rating platform, the interactions are limited to 'satisfaction'
and 'dissatisfaction'.
117
__________________________________________________________________________
One could assume that the main categories do not strongly differ in other commodity
industries. While the subcategories are specific to the mobile communications indus-
try, where it can be assumed that the categories for other mobile operators are not
completely different. 'Dissatisfaction' interaction was most prominent on the re-
view/rating platform, and other researchers have already ascertained that people
rather share negative than positive emotions on social media platforms; therefore,
firms need to carefully handle these kinds of interactions (Kaplan and Haenlein
2010). The least represented category was 'regret'. It only occurred on the social
network platform, which could allow the conclusions that public regret seems to be
the most difficult interaction and that it can be actively used by firms to respond to
unhappy customers. The results also indicate that some interactions to mobile com-
munications topics seem to be specific to certain platforms, which might give firms
the opportunity to influence social interactions.
The presented research reveals novel categories and new perspectives on social in-
teractions differing across various social media platforms; still, this requires future
academic research. The aim of better understanding the differences between various
social media platforms, and what kind of social interactions take place on them is a
valuable supplement to the existing literature. These findings derive managerial im-
plications, especially for telecommunication firms who need to consider social inter-
actions on different social media platforms when developing their social media strat-
egy.
Nevertheless, there are some limitations of this study that should be acknowledged.
These limitations relate to both methodology and data, which deserve future research
efforts. For instance, data from three weeks and only three platforms was investi-
gated. Data from another time period or other platforms may reveal other categories.
Further, data was collected referring to one provider, the revealed categories might
differ for another provider. In addition, more data could be explored using other re-
search methods like text mining and semantic analysis techniques. Furthermore, the
findings could be enriched by further research regarding the influence of identified in-
teractions and their impact on customer's behaviour.
118
Study III: Factors affecting Herd Behaviour in the Context of Social Media
__________________________________________________________________________
6 Study III: Factors affecting Herd Behaviour in the

Context of Social Media
Chapter overview: The study presented in this chapter seeks to identify determining
factors of herd behaviour in buying decisions influenced by social media. The factors
are identified based on a qualitative content analysis of expert interviews. In section
6.1, the specific objectives and the research question are outlined and put into con-
text with existing research on the discussed topic. Section 6.2 discusses the theoreti-
cal background on herd behaviour and the buying decision process in the context of
social media. The research methodology, including the empirical field, data collec-
tion, and data analysis, are presented in section 6.3. The evaluated results are pre-
sented in section 6.4. Finally, section 6.5 concludes this study, and provides an in-
terim discussion of the results. It should also be mentioned here that major parts of
the study have already been published in the article by Pavlovic (2018) (cf. section
1.5).
PART A: Foundations
1 Introduction
3 Research Design
Study III: Factors affecting Herd Behaviour in the Context of

6 Social Media
PART C: Summary
9 Conclusion
Figure 28: Dissertation Overview - Chapter 6: Study III
119
__________________________________________________________________________
Social media have developed to become an integral part of people's lives. They are
used on a daily basis to communicate, share or rate. Considering herd behaviour —
according to which people discount own information to imitate others — the transpar-
ency of other people's opinions by online communities might influence herd behav-
iour in buying decisions. Nonetheless, despite a large body of literature exploring so-
cial media, herd behaviour and buying decisions separately, little is known about the
complex interplay of the three topics. This study seeks to identify determining factors
of herd behaviour in buying decisions influenced by social media in the mobile com-
munications industry.
Individual behaviour may be influenced by other community members, given that

people rely on the opinions of others (Zhou 2011). The literature regarding herd be-
haviour suggests that people tend to discount their own beliefs and imitate others
when taking decisions (Banerjee 1992; Sun 2013). Accordingly, it is especially impor-
tant for practitioners to pay attention to the specific behaviour of consumers in their
buying decision process.
The distinct topics of social media, herd behaviour and buying decisions are studied
in literature; a combination of topics can also be found, e.g. social media and herd
behaviour (Akehurst 2008; Sparks and Browning 2011) or social media and buying
decisions (Chen 2008). However, there is scarce research exploring what factors
may account for the phenomenon whereby individuals tend to adapt their buying de-
cisions to follow the herd and the role of social media therein.
Hence, it is important to understand the interplay of social media, herd behaviour and
buying decision. Building on these thoughts, the present paper aims to answer the
following research question in the empirical context of the mobile communications in-
dustry:
ing decisions?
In practice, the three topics are already often combined in one position; for example,
in online marketing. Accordingly, data analysis tools are used to track customers' be-
120
__________________________________________________________________________
haviour to target them with the help of online communities (Scott 2015). In order to
utilise the considerable knowledge of practitioners, the research question was ap-
proached through explorative expert interviews. In general, the target of expert inter-
views is not to confirm or reject the developed hypotheses, but rather to prompt their
evolution out of the material to provide new supporting ideas for further research; this
approach is also pursued in this study (Bogner et al. 2014). The conducted expert in-
terviews were followed by qualitative content analysis according to the approach of
Mayring (2015). Finally, the inductively developed categories are presented and dis-
cussed.
6.2.1 Herd Behaviour in the Context of Social Media
Combining economics with psychology and experimental research on decision-

making behaviour, behavioural economists have sought to better explain what influ-
ences consumers in their purchasing decisions and how this behaviour affects the
market (DellaVigna 2009). In essence, behavioural economics uses evidence of how
humans actually behave (Ariely 2008). It argues that observed human decision-
making differs from the rational welfare-maximising behaviour assumed in neoclassi-
cal economic models. Psychologist Daniel Kahneman and cognitive and mathemati-
cal psychologist Amos Tversky are certainly the most famous researchers in the field
of research into human decision-making behaviour. They have identified several
cognitive biases that show where people deviate from rational decisions or judge-
ment (Kahneman 2011; Kahneman and Tversky 1979). Herd behaviour can be clas-
sified as a phenomenon describing the notion that “everyone does what everyone
else is doing, even when their private information suggests doing something quite dif-
ferent” (Banerjee 1992), (cf section 2.1). While this phenomenon is already well ex-
plored in asset and stock markets (Hirshleifer and Hong Teoh 2003; Lux 1995),
barely any literature explores herd behaviour in the mobile communications industry.
Chen (2008) states that online herd behaviour occurs "when people use the product
evaluations of others to indicate product quality on the Internet". Furthermore, he ex-
amines herd behaviour in the realm of online book purchasing and investigates the
effectiveness of different recommendation sources. Moreover, the impact of online
reviews in the tourism industry has been explored (Akehurst 2008; Sparks and
Browning 2011). In the field of IS, Sun (2013) adds the two concepts of "discounting
121
__________________________________________________________________________
one’s own information and imitating others" in his research and identifies how herd
behaviour influences technology adoption decision-making and post-adoptive system
use.
Previous research shows that herd behaviour occurs in various industries and it is a
field that can be explored through different lenses/perspectives. Nevertheless, most
of the literature focuses on the herd behaviour phenomenon itself or reveals that
online communities might influence this phenomenon. Consequently, very little is
known about the factors that affect herd behaviour influenced by social media. Moti-
vated by this gap in the literature, the present study seeks to explore these different
factors.
6.2.2 Buying Decisions in the Context of Social Media
It is necessary to pay attention to the specific behaviour of consumers in the different

phases of the buying decision process, as well as using corresponding strategies and
instruments to address customers phase-specifically. This does not replace the iso-
lated consideration of individual psychological effects; rather, it supplements isolated
individual motivation to provide a process-oriented, holistic view of modern marketing
(Foscht and Swoboda 2011). The purchasing process is understood as the whole
process, from the emergence of a certain need to the various types of decision-
making with information acquisition and processing, product selection, shopping be-
haviour, use and eventual disposal of the product, and the increase in consumer
product experience (Kuß and Tomczak 2007). A widespread tool to gain better un-
derstanding of customers and their behaviour is a five-stage buying decision process
model, comprising need recognition, information search, evaluation of alternatives,
buying decision and post-buying behaviour (Armstrong et al. 2009). However, con-
sumers do not necessarily pass through all stages; e.g. in routine purchases, con-
sumers often skip or reverse some of these stages (Armstrong et al. 2009).
According to Foscht and Swoboda (2011), a holistic view seems to make sense if at
least three phases are distinguished (cf. section 2.2.3). They propose the phases
pre-buying phase, buying phase and post-buying phase. The focus in the pre-buying
phase is to gather information about products by using different channels. In the buy-
ing phase, the consumer ranks the product alternatives in a preference order. Never-
theless, his/her first choice does not necessarily need to be the finally-bought one.
Friends or community influence might change a consumers' preference for another
122
__________________________________________________________________________
product even if they intended to buy another one (Comegys et al. 2006). The post-
buying phase refers to post-buying action (e.g. repurchase) and customer satisfac-
tion (Comegys et al. 2006).
There is a considerable amount of research examining the influence of online com-

munities on the buying decision process (Comegys et al. 2006; Maas et al. 2014). In-
sights into how social media can be used optimally are not only valuable for science
but they also hold strong value for marketers (Kaplan and Haenlein 2010; Scott
2015). However, very little can be found regarding herd behaviour in buying deci-
sions influenced by social media. Thus, the aim of the current study is to elaborate
more on this issue.
The eminent importance of expert interviews for research practice is undisputed

(Bogner et al. 2009; Flick et al. 2012). It is a commonly used instrument for collecting
data in different research fields, including the field of IS (Myers and Avison 2002).
The benefit of obtaining good results quickly is one aspect that makes expert inter-
views an appealing option. However, the more important aspect is to use experts as
'crystallisation points' for particular knowledge that becomes effective in practice and
thus guides action for other actors, which makes this a valuable research method
(Bogner et al. 2009; Flick et al. 2012). The present research topic touches different
fields of science in terms of IS, marketing, and psychology. In order to supplement
knowledge from the literature and link the different research directions in the context
of business-related decisions regarding the research question, expert interviews
were chosen as the appropriate research method.
6.3.1 Empirical Field
According to Bogner et al. (2009), experts are defined as having the opportunity to
(at least partially) enforce their orientations. Experts have exclusive knowledge and
they are able to decisively determine from which perspective and by means of which
concepts certain problems are considered. It is exactly this practical relevance that
makes the experts interesting for the present research project. With expert inter-
views, different dimensions of expert knowledge can be accessed, namely technical,
process and interpretative knowledge. Although usually all dimensions are covered in
expert interviews, in the present study the focus is placed on process and interpreta-
tive knowledge.
123
__________________________________________________________________________
For the purpose of this study, it was important to access major business-related deci-
sion-makers who deal with the topics of social media, customer behaviour and cus-
tomer buying decisions as well as their connection. Furthermore, it was crucial to ac-
cess senior managers with a mobile communications background, as this industry
was chosen as the empirical context. The goal was to gain access to their practical
perspective on the research question and bring the different topics together, whereas
there are often investigated separately in existing literature. Therefore, it was impor-
tant to access experts who combine different perspectives such as marketing, IS or
psychology in their position. In practice, these separate fields are often combined in
one role. In order to obtain access to the interviewees, own network was used, rec-
ommendations and online research, particularly using LinkedIn and Xing to identify
potential candidates. Overall, fifteen candidates were selected and approached,
whereby eventually six in-depth interviews were conducted. The experts were all
senior management members, aside from one data scientist. The interviewees were
a CEO from an online marketing company, three senior managers from a telecom-
munication company — one responsible for online sales and marketing, one for so-
cial media, and one for business analytics — one social media consultant and one
data scientist active in the online area. An overview of the interviews conducted can
be found in Appendix D.2.
6.3.2 Data Collection
The research process followed an explorative approach. The aims of explorative ex-
pert interviews are to provide a first orientation in an unexplored field, sharpen a sci-
entific problem or generate hypotheses. The focus is on knowledge of action and ex-
perience that has been derived from practice and is reflexively accessible (Bogner
and Menz 2009; Gläser and Laudel 2010).
In order to develop a fairly elaborated topic guide for the interview, a multi-step pro-
cedure as recommended in the relevant literature was used (Gläser and Laudel
(2010). All research questions, existing studies and outcomes of a previous online
observation were compiled. Subsequently, in several steps this content was sys-
tematised and reduced to sharpen the research question. The result was a general
catalogue of questions that was then sorted and grouped into larger subject blocks.
For each block, generic terms were formulated and respective questions assigned,
before being translated into interview questions for the semi-structured interview in
the next step.
124
__________________________________________________________________________
In preparation of the expert interview, two pre-test interviews were conducted to clar-
ify the scope and feasibility of the study. The final interview guide was structured as
follows: (1) welcoming the interviewee, introducing the underlying context of the in-
terview and the procedure; (2) general experience with social media and their impact
on the mobile communications industry; (3) influence factors on herd behaviour and
buying decisions; (4) reasons for the influence of social media; (5) differences in the
influence of different influencing groups; (6) specifics regarding the influence on deci-
sions concerning mobile communications products; and (7) comments and closing.
The full interview guide is provided in Appendix D.1. The interviews lasted on aver-
age for one hour and not more than two hours. The interviews were all audio re-
corded with the permission to do so. All interviewees were interested in the research
topic and most had valuable input for the study in addition to the questions. Further-
more, all conversations were conducted in an open, constructive and good-natured
manner.
6.3.3 Data Analysis
To answer the research question concerning what factors affect the influence of
online communities on herd behaviour in buying decisions, the generated data was
analysed by applying qualitative content analysis as a scientific research technique.
In order to ensure a rigorous interpretation of the material, Mayring (2015) and Krip-
pendorff (2004) have developed clear rules and systematic procedures to analyse
the data. Accordingly, content analysis is a widely accepted method in IS research
and is applied in many studies (Myers and Avison 2002). Since the research question
aims to combine different research perspectives and consider the perspective from
practice to facilitate future studies, the inductive procedure is chosen. This approach
is data-driven, which means that categories are identified directly from the data (May-
ring 2015).
To perform the qualitative content analysis, the procedure model according to May-
ring (2015) outlined in section 3.2.1.3was applied. In order to analyse the data, the
fully-transcribed interviews and all notes from the interviews were imported to the
MAXQDA 12 software.
Accordingly, after the first round of category development 57 categories had

emerged. This was followed by the next review of the developed category system on
the raw material, where the abstraction level was further raised and the categories
125
__________________________________________________________________________
arranged according to main categories. To achieve this, theoretical considerations

from the literature were also used.
To accomplish the reliability check by means of intercoder reliability (Mayring 2015),

a second researcher familiar with the research area was introduced to the category
system (Appendix D.3) and coded three randomly selected parts of the data. The
calculated Holsti coefficient of reliability was 0.78.
6.4 Results
The analysis results of the expert interviews conducted are presented in the follow-
ing.
Main Categories
The findings regarding the research question revealed 29 subcategories grouped into
seven main categories (Figure 29) based on an inductive coding procedure of the re-
search material. The percentages indicate the frequency distribution of the catego-
ries. This means how often the various aspects that were combined into categories
were referred to from the point of view of the experts interviewed.
Drivers of social media usage 32.0%
General influence groups 17.0%
Product specifics 14.0%
Role of social media in buiyng decision process 13.0%
Customer segment specifics 10.0%
Personal vs. Mass opinion 7.0%
Social media activity 7.0%
Figure 29: Overview of the Seven Identified Influencing Factors in the Mobile Communications
Industry (Main Categories)
The analysis of the expert interviews shows that 'drivers of social media usage' is
the largest category under the research question context. This category describes
that the drivers for the use of social media have an impact on the influence of social
126
__________________________________________________________________________
media on herd behaviour in the buying decision process. This means that consumers
can exhibit different herd behaviour in a buying decision depending on what drives
their social media use. For example, consumers could use social media in their buy-
ing decision process and do so primarily because they expect an objective assess-
ment by other users via social media. These could then display more pronounced
herd behaviour than would be the case, for example, with consumers who use social
media primarily out of convenience.
The second main category 'general influence groups' shows which type of influ-
ence groups may affect herd behaviour in buying decisions. In the experts' view,
consumers are exposed to the influence of various groups in the course of a buying
decision process. This can be done by actively searching for information, e.g. on re-
view and rating sites, or more passively by exchanging opinions with friends or family
or casually observing the decisions of others. The influence on herd behaviour in the
buying decision process can differ depending on the group from which information
originates. For example, friends may have a stronger influence on herd behaviour
than, for example, social media.
The third main category 'product specifics' describes that the influence of social
media on herd behaviour in the buying decision process can differ depending on the
product. Thus, the product type would have an influence on whether herd behaviour
occurs or how strong it is. For example, it could be different for a smartphone than for
a mobile phone contract or a mobile phone cover.
The fourth main category 'role of social media in buying decision process' de-
scribes the notion that the impact of social media might differ depending on the buy-
ing phase. According to the experts buying decisions are no longer made without the
influence of social media. Consumers use them heavily, especially at the beginning
of the buying decision phase, but also in later phases. Accordingly, the influence of
social media on herd behaviour at the beginning of a buying decision can be different
from that at the end.
The fifth main category 'customer segment specifics' shows that the impact of so-
cial media on herd behaviour in buying decisions might depend on specific customer
segments or personality types. This means that, depending on the personality traits
of a consumer, he or she can be influenced to a greater or lesser extent by social
media when making a buying decision.
127
__________________________________________________________________________
The sixth main category 'personal vs. mass opinion' provides an indication con-
cerning what might have a greater influence on herd behaviour in buying decisions.
Thus, it can make a difference where the information comes from, which consumers
are exposed to when making their buying decision.
The last main category 'social media activity' states that the personal activity level
of consumers in social media can have an influence on herd behaviour in the buying
decision process. Therefore, consumers who are very active in social media may be
either more or less exposed to herd behaviour than consumers who hardly use social
media at all.
Subcategories
The frequency distribution of the subcategories is depicted in Table 13. The percent-
ages indicate the distribution of subcategories within the main category.
128
__________________________________________________________________________
Table 13: Overview of Frequency Distribution of the Subcategories
MC 1 Drivers of so- MC 2 General influ- MC 3 Product specif- MC 4 Role of so-

cial media usage ence groups ics cial media in buy-
ing decision proc-
ess
SC 1.1 Credibility SC 2.1 Review- SC 3.1 Low- SC 4.1 Information

(23%) /rating platforms involvement vs. high- provider in pre-
(42%) involvement products buying phase
SC 1.2 Wisdom of
(43%) (59%)
crowds (20%) SC 2.2 Friends and
family (22%) SC 3.2 Hardware as SC 4.2 Confirma-
SC 1.3 Convenience
high-involvement tion in buying
(14%) SC 2.3 Search en-
product (33%) phase (32%)
gine (19%)
SC 1.4 Trustworthi-
SC 3.3 Connectivity SC 4.3 Optimisa-
ness (13%) SC 2.4 Social net-
as low-involvement tion in post-buying
works (10%)
SC 1.5 Independency product (24%) phase (9%)
(7%) SC 2.5 Independent
testing foundation
SC 1.6 Transparency
(7%)
(7%)
SC 1.7 Popularity of
community (6%)
SC 1.8 Information
overload (5%)
SC 1.9 Topicality
(6%)
C 5 Customer segment spe- C 6 Personal vs. mass opin- C 7. 1 Social media activ-
cifics ion ity
SC 5.1 Technology affinity SC 6.1 Reliable personal SC 7. 1 Posters vs. lurkers

(31%) opinions (42%) (100%)
SC 5.2 Price sensitivity (27%) SC 6.2 Harmonisation of con-
flicting opinions (42%)
SC 5.3 Service affinity (16%)
SC 6.3 Valuable topicality
SC 5.4 Culture (14%)
and fast access (17%)
SC 5.5 Age (12%)
The first main category 'drivers of social media usage' is the largest category and
sums up nine subcategories. 'Credibility' refers to users authentically describing
their experiences with telecommunication products and services, which helps con-
sumers to believe in other people's opinions. 'Wisdom of crowds' describes that peo-
ple rely on opinions that seem to be shared by many people, e.g. on Amazon or other
129
__________________________________________________________________________
rating platforms, one can see how many people have rated a product or a service.
'Convenience' refers to the notion that people want to have as much transparency as
possible but they are also overwhelmed by the quantity of accessible information,
whereby this phenomenon is related to the subcategory 'information overload'. Thus,
it is more convenient for people to refer to other people's experiences and ratings
rather than trying to form their own opinion. 'Trustworthiness' comes from using
names, pictures and private data (e.g. age, gender, marital status, interests) in pro-
files of social media platforms, which seems to suggest personalisation. 'Independ-
ency' refers to objective reviews, opinions and ratings from real consumers. Whereas
'transparency' refers to information availability, the 'popularity of community' de-
scribes the notion that the greater the popularity of a platform, the stronger its poten-
tial impact on herd behaviour in buying decisions. 'Topicality' shows that social media
provide access to latest information, which also seems to be a driver for their usage.
The second main category 'general influence groups' comprises five subcatego-
ries. 'Review/rating platforms' refer to an opinion-forming function. People use infor-
mative descriptions, comments, rankings on such platforms to compare products and
prices to support their buying decision process. Nevertheless, 'friends and family' be-
long to an offline community influence group and they also seem to play a decisive
role in the buying decision process. 'Search engine' refers to the notion that by using
search engine optimisation (SEO) people's buying decisions might be indirectly influ-
enced and thus also their herd behaviour. 'Social networks' like Facebook or Twitter
are widely spread and actively used and they also seem to be an integral part of the
influencing groups. By using 'independent testing foundations', people use independ-
ent expert assessments in their buying decision process; consequently, their deci-
sions might also be influenced.
The third main category 'product specifics' includes three subcategories and re-
lates to the mobile communications products. The subcategory 'low-involvement vs.
high-involvement products' describes the notion that herd behaviour in the buying
decision might differ depending on the product. The experts did not agree on where
they would expect a stronger development of herd behaviour, namely with low- or
high-involvement products. 'Hardware as high-involvement product' states that in the
mobile communications industry mobile phones, tablets or gadgets can be assigned
to high-involvement products, whereas connectivity can be counted as a low-
involvement product.
130
__________________________________________________________________________
The fourth main category 'role of social media in buying decision process' com-
prises three subcategories. The first subcategory 'information provider in the pre-
buying phase' describes social media as being primarily used to gain a broad over-
view of the product in this phase. The second subcategory 'confirmation in buying
phase' states that in the buying phase it is simply about confirming or strengthening
one’s own decision with the help of social media. The last subcategory 'optimisation
in post-buying phase' is not about changing one’s mind but rather optimising the buy-
ing decision taken, e.g. a mobile contract renewal with the help of online communi-
ties.
The fifth main category 'customer segment specifics' has five subcategories. De-
pending on 'technology affinity', the impact of online communities on personal deci-
sions may differ. People’s personal 'price sensitivity' might have an influence on how
strongly people rely on social media information, as people depend on their 'service
affinity'. Dependent on the 'culture', people might go about their buying decisions dif-
ferently and thus are affected by different factors. 'Age' refers to differences between
age groups and their usual buying decision behaviour and handling of social media.
The sixth main category 'personal vs. mass opinion' comprises three subcatego-
ries and provides an indication concerning what might have a greater influence on
herd behaviour in buying decisions. The subcategory 'reliable personal opinions' de-
scribes the notion that personal opinions and recommendations seem to be more re-
liable than independent mass information and thus potentially more valued and hold-
ing stronger influence. 'Harmonisation of conflicting opinions' describes the phe-
nomenon whereby mass opinions and personal opinions strongly differ and people
try to average the extreme positions for themselves. 'Valuable topicality and fast ac-
cess' refers to the 'always on' trend, which allows people to access the latest informa-
tion at any time via online communities.
The last main category 'social media activity' refers to 'posters and lurkers', a phe-
nomenon that describes the different activity levels in using social media. While the
smallest part of users actively create content, a minor part at least comment on the
content and the major part are passive and only read the content. This occurrence
has also been described as the 1-9-90 rule. Thus, the activity level might affect herd
behaviour depending on the group to which one belongs.
131
__________________________________________________________________________
Furthermore, the interviewees had additional comments that are not reflected in the
category scheme but should be mentioned here. All shared the opinion that the herd
behaviour phenomenon is further reinforced by social media. Therefore, it is crucial
for firms to consider and integrate them in their marketing and sales activities. In ad-
dition, they stated their interest in further research on following topics: What is the tip
of the scale when different communities (online and offline) recommend different
products? What role do previous experiences play; for instance, with a telecommuni-
cation provider? What kind of community (online and offline) plays the strongest role
in each buying decision process for telecommunication products?
This study has sought to explore factors that influence the impact of social media on
herd behaviour in the buying decisions and gain a deeper understanding for the mo-
bile communications industry. The research shows that there is no single relevant
factor but rather multiple factors. The qualitative content analysis revealed seven
main and 29 subcategories developed through an inductive approach.
The largest main category 'drivers for social media' comprises the largest number of
subcategories, showing that there are various drivers for the usage of social media.
Some previous studies have found that factors such as perceived usefulness, com-
mitment, trust, self-efficacy and outcome expectation significantly influence the social
media usage (Zhou 2011). Some of them match the subcategories developed in this
study, while additional subcategories like 'credibility', 'wisdom of crowds ', 'independ-
ency', 'transparency' or 'popularity of community' can also be added. The second
category 'general influence groups' showed that herd behaviour in buying decisions is
not only influenced by different types of social media but also by other influencing
groups like 'family and friends' or 'independent testing foundations'. The third cate-
gory concerning 'product specifics' related to telecommunication products revealed
that herd behaviour in the buying decision might differ depending on the product
type, namely, between low- or high-involvement products; however, the experts did
not agree on where herd behaviour would be more affected. Marketing research
studies like that by Trommsdorff (2004) have investigated the different product types
regarding buying behaviour, e.g. impulse buying of consumer goods, although the
context of herd behaviour has not been explored. The main category concerning the
'role of social media in buying decision process' has worked out the different impacts
132
__________________________________________________________________________
that social media has in the each buying process step. In the beginning, social media
functions as an information provider; it also serves this role during the buying phase
to confirm the buyer’s preference. In the last phase regarding possible repurchase,
the online communities' role is to provide optimisation hints; for instance, for a mobile
contract renewal. Furthermore, the fifth main category of 'customer segment specif-
ics' shows that differences in personality related to 'technology affinity' or 'culture'
might be an influencing factor for herd behaviour in buying decisions. 'Personal vs.
mass opinion' is the third main category describing the notion that personal and mass
opinions have benefits that are valued by consumers. The finding that personal opin-
ions and recommendations coming from people whom they know and trust such as
family members and friends are more reliable and thus more likely to be considered
than mass opinion is also supported in the literature (Sinha and Swearingen 2001).
The last main category of 'social media activity' is a further factor that influences herd
behaviour in buying decisions influenced by social media. The phenomenon of 'post-
ers and lurkers' in social media is well known in the literature (Schlosser 2005) and
revealed in this study to have an influence on herd behaviour in buying decisions.
Moreover, the additional comments of the interviewees provided valuable input for
further research.
The presented research combines different perspectives from IS, marketing and psy-
chology, and reveals new perspectives that do not seem to have been mentioned in
the literature before and thus facilitate future academic research. A better under-
standing of the influencing factors that social media can have on herd behaviour in
buying decisions is a valuable supplement to the existing literature. The results of the
study indicate that there are further factors affecting herd behaviour in buying deci-
sions by social media than some single factors. These findings are also valuable for
practice as they can help telecommunication firms or other commodity markets to op-
timise their online marketing and sales activities by analysing the relevant factors re-
lated to their consumers and their buying behaviour. Nevertheless, there are some
limitations of this study that should be acknowledged. Such limitations relate to both
methodology and data, which warrant future research efforts. For instance, only six
expert interviews were conducted, and interviews with other experts could reveal
other categories. Another limitation is the interpretation of code frequency as an indi-
cator of importance, because the code frequency can be biased depending on the
focused topics of the expert.
133
Study IV: Herding along the Consumer Buying Decision Process
__________________________________________________________________________
7 Study IV: Herding along the Consumer Buying

Decision Process
Chapter overview: The study presented in chapter 7 is the most comprehensive of
this doctoral thesis. As an empirical study, it examines various factors influencing
herd behaviour along the buying decision process. In the first section of this chapter,
the specific objectives and the research question are outlined. Then, section 7.2 dis-
cusses the theoretical background on the topics: herd behaviour in buying decisions,
information sources, product differentiation, social media activity, and personality
types. The hypotheses are developed in section 7.3. The research methodology, in-
cluding procedure and sample profile, research design, and data analysis are pre-
sented in section 7.4. The descriptive results are presented in section 7.5 and the
analytical results in section 7.6 Finally, section 7.7 concludes this study, and provides
an interim discussion of the results.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
Study IV: Herding along the Consumer Buying Decision

7 Process
PART C: Summary
9 Conclusion
Figure 30: Dissertation Overview - Chapter 7: Study IV
134
__________________________________________________________________________
Already in 1992, Banerjee developed a model that "analysed a sequential decision

model in which each decision maker looks at the decisions made by previous deci-
sion makers in taking her own decision" (Banerjee 1992, p. 797). The study results
showed that human decision behaviour is characterised by doing what others are do-
ing rather than using their own information that might suggest something different
(Banerjee 1992). This phenomenon is an influential and well documented feature of
human behaviour which has been studied in a number of domains, particularly in so-
cial psychology and economics but also in IS research. A detailed overview of the re-
spective literature is presented in the state-of the-art analysis (chapter 4). With the
dissemination of web technologies, the alleged transparency of other people's opin-
ions has significantly increased, along with the possibilities to easily observe the de-
cisions and behaviour of others. With regard to the large amount of information avail-
able, people are exposed to an overwhelming uncertainty, and thus may take a
shortcut to follow a herd in technology adoption to rely on the opinions of many oth-
ers (Sun 2013). This phenomenon impacts consumers as well as firms because this
easily accessible information may greatly affect consumers' buying decisions. Never-
theless, due to the low level of informativeness, herd behaviour remains fragile, e.g.
a new information at a later stage might change someone's decision. Furthermore,
herd behaviour may be influenced by several factors such as information popularity,
valence, and product type (cf. chapter 4).
However, little IS research has been conducted to apply herd theory within the dis-
tinct buying decision phases. In the pre-buying phase, consumers actively search for
information and evaluation of alternatives (Kuß and Tomczak 2007). In the buying
phase, consumers decide on the product they want to buy (Häubl et al. 2010), and in
the post-buying phase, consumers may compare their current experience with what
they had anticipated before making a purchase (Comegys et al. 2006). So far, the ex-
isting research seems to focus exclusively on individual aspects, i.e. on individual
buying phases (usually the pre-buying phase) and/or on certain influencing factors.
For example, Chen (2008) examines herd behaviour in the realm of online books and
investigates the effectiveness of different recommendation sources only covering the
buying intention phase. Sun (2013) considers in his research the influence of herd
behaviour with focus on the post-adoptive phase. Godinho de Matos et al. (2014)
conclude that the tendency of a subscriber to adopt to the iPhone 3G increases with
135
__________________________________________________________________________
the percentage of friends who have already adopted for the buying phase. However,
the existing IS research has not advanced far enough to explain differences in herd
behaviour between the individual buying decision phases and the differing factors af-
fecting the respective phases.
A further elaboration of the individual aspects, in particular the consideration of the

individual buying decision phase and the overall holistic consideration in the present
dissertation can thus contribute to the further theory development of herd behaviour.
Thus, to investigate how herd behaviour generally differs among the consumer buy-
ing decision could contribute to ongoing theorising about herd behaviour.
Furthermore, and particularly for practitioners whose business model is based on

contractual relationships and loyal customers, as is the case in the mobile communi-
cations industry, it is critical to success to understand how herd behaviour in individ-
ual buying decision phases differ. In the pre-buying phase this is important to reduce
costs to attract customers in a most efficient manner, in the buying phase to sell the
most profitable product, and in the post-buying phase to not lose a customer to the
competition (as it is always more expensive to acquire a new customer than to keep
an existing one). Therefore, it is also crucial to understand how the instrument social
media, which is widely spread and used by firms, influences this phase and what
other factors influence this phenomenon. This is especially necessary when numer-
ous sources of information may greatly affect consumers' buying decisions. As a
typical type of experience good, the quality of a mobile communications product (i.e.
smartphone, mobile contract, entertainment service) is difficult to evaluate before
adoption. As a consequence, consumers are motivated to make extensive use of
various information sources, hoping to make well-informed decisions (Gu et al. 2012;
Luo and Zhang 2013), and are thus more reliant on the opinions of other consumers.
Building on these thoughts, the present study aims to answer the following research
questions in the context of the mobile communications industry:
phases?

136
__________________________________________________________________________
To approach these research questions, this study will present results based on an
online survey with three comparison groups. Thus, different aspects and their influ-
ence on herd behaviour in the distinct buying decision phases can be evaluated.
In line with the topics elaborated upon in the literature review (chapter 4) and the re-
vealed topics in Study III (chapter 6), the following aspects are analysed in the dis-
tinct buying decision phases based on the expert interviews:
 General herd behaviour
 Different information sources
 Product differentiation
 Social media activity
 Personality traits
7.2.1 Herd Behaviour in Buying Decisions
Herd behaviour and the buying decision process were introduced in chapter 2 and
examined in more detail in the literature review (chapter 4). Hence, in this section
only a brief outline of the main aspects is given.
The common understanding of herd behaviour is that people discount their own in-
formation and imitate others (Banerjee 1992; Bikhchandani et al. 1992; Sun 2013).
Hirshleifer and Hong Teoh (2003) developed a comprehensive herding framework

which includes various sources for conformity in human behaviour (cf. section 2.1.3.).
According to a growing body of literature, information cascades seem to be the most
promising concept to explain the underlying mechanism (Cipriani and Guarino 2009;
Sunstein 2005); they describes a state in which imitation is certain to occur. Primary
conditions that lead to the mechanism of herd behaviour are the following: first, the
presence of uncertainty about a decision, and second, the observed actions of other
individuals. An informational cascade occurs in both conditions, referring to one's
tendency to conform to the opinions of others and the acceptance of received infor-
mation as an indicator of uncertainty (Banerjee 1992; Bikhchandani et al. 1992).
Regarding the buying decision process, which is characterised by uncertainty (Kuß

and Tomczak 2007), as well as the availability of information about other users’
choices, the Internet offers an ideal environment for informational cascades. To re-
duce uncertainty in the pre-buying phase, consumers undertake a more or less in-
137
__________________________________________________________________________
tensive search for information to be able to evaluate considered alternatives. During

the buying phase, after the identification of the optimal alternatives and the determi-
nation of the satisfactory ones, a buying intention is formed and the purchase is
processed. Herd behaviour could lead to the dominance of one product or technology
over another despite the availability of close substitutes, or to the rejection of more
efficient technologies (Abrahamson 1991). Nevertheless, due to the low informative-
ness, herd behaviour is fragile; e.g. new information at a later stage might change
someone's decision, which could influence the post-buying phase.
The present study aims to shed more light into the phenomenon of herd behaviour in
the distinctive buying decision phases.
7.2.2 Different Information Sources
Consumers' decisions are directly and significantly influenced by the actions of oth-
ers. These 'actions' range from personal recommendations from a friend to passive
observation of what a stranger is wearing. Consumers are eager to obtain more rele-
vant product information and professional suggestions (e.g. ratings, comments, and
recommendations) from others’ expertise and experience, since products have be-
come increasingly complex and technical in nature (Godes et al. 2005).
In the literature review (cf. section 4.3.1.2), information sources were elaborated
upon with regards to topics addressed in the literature. The following different infor-
mation sources were identified: 'peer groups', 'experts', and 'crowd'. These are briefly
outlined here.
A 'peer group' is usually a group of people who know each other, e.g. colleagues,
family and friends, and acquaintances (Salazar et al. 2013). In this sense, the infor-
mation provided by peer groups with a higher degree of social proximity appears to
have a stronger weight in consumer's decision process (Salazar et al. 2013). WOM
and eWOM play a significant role in consumers' purchasing decision process. Never-
theless, there are different sources for eWOM and research has shown that the
source of the information plays an important role in consumers’ attitude and behav-
ioural intention formation (Purnawirawan et al. 2014).
Further, consumers often rely on the opinions of well-known experts or recommenda-

tions from well-known published publications, especially when buying complex, dura-
ble, expensive products or services (e.g. financial products, home, car and insur-
ance) (Nakayama et al. 2010). Expert sources (e.g. sources who are considered as
138
__________________________________________________________________________
knowledgeable and competent) are assumed to possess the knowledge and ability to
provide accurate information, and thus consumers tend to seek advice from experts
rather than non-experts (Pornpitakpan 2004).
Besides reviews submitted by friends or other peers, there are also those submitted
by the rest of the online community (stranger-networks), the so-called 'crowd'. In
general, the crowd consists of a great number of people with heterogeneous prefer-
ences, experiences and know-how, generating the wisdom of the crowd (Surowiecki
2004). Therefore, online reviews generated by the 'crowd' are as well an important
source for prospective buyers to observe, and learn from the mass experiences and
behaviour of other users (Lee et al. 2015; Zhang et al. 2015).
The literature addressing these aspects focuses on the question of which source of
information ultimately has the greatest influence on consumers decision behaviour.
But the picture does not seem to be so unambiguous. For example, Chen (2008)
shows in his study that recommendations of the crowd influenced subject choices
more than recommendations of an expert. In contrast, Nakayama et al. (2010) stated
that expert opinion wins out over recommendations from peers or other consumers.
Therefore, the objective of this study is to develop a more differentiated picture.
7.2.3 Product Differentiation
Consumers often have to make purchasing decisions with incomplete information

because they lack complete information on products and services. Additionally,
searching and seeking information is time consuming and can be costly, and there
are trade-offs between perceived costs and benefits of additional search (Stigler
1961). Thus, consumers are motivated to follow a buying decision process that re-
duce uncertainty. The extant research addresses the aspect that herd behaviour
might differ between different product types.
There are several ways to differentiate product types or to classify them (e.g. experi-
ence goods vs. search goods, high-involvement vs. low-involvement products). Sec-
tion 4.3.2.1 provides an overview.
A perspective to investigate the product type differences is the extent of product in-
volvement, which relates more to the individual perception of consumers. Zaich-
kowsky (1985) defines product involvement through the perceived relevance of the
product by a person based on the inherent needs, values and interests. Product in-
139
__________________________________________________________________________
volvement refers to the level of interest or attention that a consumer pays to a prod-
uct (Cohen 1983; Mitchell 1979).
There are various involvement product classifications, which are one-dimensional

and do not differentiate according to segments or situations. Different involvement
hit-lists have been developed in different studies (see, e.g., Lastovicka and Gardner
(1979), Jeck-Schlottmann (1988), and Jain and Srinivasan (1990)). Involvement lists,
which originate from different cultures, times, target groups and measurement meth-
ods, reveal the dilemma of product involvement. They cannot be generalised, but
only the involvement specific product characteristics. Zaichkowsky (1985) claims, for
example that consumer's involvement level is sensitive to personal: the intrinsic in-
terests, values or needs that motivate one to the object; physical: the characteristics
of the object that cause differentiation and increase interest; and situational: some-
thing that temporarily increases relevance or interest in the object. Laurent and
Kapferer (1985) provide a remarkable approach for the multidimensional involvement
definition. They empirically show that five independent involvement determinants are
to be considered across the most diverse product groups: interest in the product, re-
warding the nature of product or the consumption, personal expression possibilities,
perceived risk in terms of probability, and perceived risk in terms of risk costs. Gold-
smith and Emmert (1991) examined in their study additional to the scale of Zaich-
kowsky (1985) and Laurent and Kapferer (1985), the Mittal (1989) involvement scale,
to validate their operationalisation. They stated that all three are valid and research
can apparently use any of the three scales. According to Goldsmith and Emmert
(1991), the Zaichkowsky (1985) scale appears to be the most general while the scale
of Mittal (1989) is more suited toward measuring buying decision involvement. Of
these 3 scales, the Laurent and Kapferer (1985) scale appears to tap the antece-
dents of continuing product involvement. In the dissertation at hand, the developed
scale is oriented to that by Laurent and Kapferer (1985).
Existing studies have so far paid little attention to the level of product involvement in
the context of herd behaviour. Chen (2008) argues in his study that books, which he
defines as low-involvement products, and where consumers spend less time and
cognitive efforts, are more likely to elicit herd behaviour. Gu et al. (2012) examine the
relative impact of different WOM sources on retailer sales for digital cameras as high-
involvement products.
140
__________________________________________________________________________
The aim of this study is therefore to develop new insights into product involvement on
herding along the buying decision process.
7.2.4 Social Media Activity
With the increase of online communication possibilities (e.g. chats, blogs, social me-
dia), the number of participants is constantly increasing, especially within social me-
dia. Young people in particular are using social media to a high extent, with more
than 50% of them doing so daily or several times a week (mpfs 2018). This is an ad-
vantageous development for users and firms to mine information regarding customer
perception, needs or behaviour; as well as to influence customers and generate
revenue through sales of products, services, information, and advertising (Ridings et
al. 2006). But at the same time, all community platforms heavily depend on users for
generation of content (Baumer et al. 2011; Velasquez et al. 2014). Overall, driven by
social media the user's role has dramatically changed by collapsing the distinction
between media consumers and producers (Miller 2011).
Therefore, Internet users are often referred to as one homogeneous group and it is
assumed that everyone produces user-generated content and is equally chatting,
sharing, blogging, creating, etc. In fact, the reality is different. Public online participa-
tion may represent one form of activity, but there are other forms of activity levels
(Edelmann 2013). Nevertheless, the largest portion of the online environment is often
the so-called 'lurkers', a special group of website users who regularly login to online
communities but seldom post, and they can account for up to 90% of users (Non-
necke and Preece 2000). Ridings et al. (2006) categorised in their research three
groups: lurkers who never posted, infrequent posters who posted three or less times
per month, and frequent posters. They found out that posters and lurkers differ sig-
nificantly in their behaviour within online communities. Furthermore, there is a fourth
group to be mentioned, the non-users who do not participate online at all (neither ac-
tively nor passively).
Online users are given the technological possibility to access the Internet without
having to be visible or to publicly participate (Edelmann 2013). However, they may be
affected by social media content even if they do not contribute to the ongoing con-
versations (Ridings et al. 2006). There are many definitions of lurkers. Usually, 'lurk-
ing' is associated with non-participation or non-posting, although these definitions
vary in numerical terms. According to Ridings et al. (2006) posting frequency is the
141
__________________________________________________________________________
key factor in the determination of lurker status and lurkers do not post or post very in-
frequently. While other researchers define lurkers based on their contribution in the
first twelve months after subscription to a list (Batinic et al. 2002), or who have not
posted in the recent months (Nonnecke and Preece 2000) or have never or only oc-
casionally post a message (Nonnecke et al. 2006; Nonnecke and Preece 2000).
Lurkers are often described as people who observe what is going on but do not par-
ticipate or remain silent and inactive, and are thus associated with observation, inac-
tivity/passivity, silence or invisibility (Leshed 2005). Therefore, they seem to be hard
to reach or to involve e.g. in online communities (Nonnecke et al. 2006). Further, they
are seen as those members who log into a community, read blogs and discussions,
but do not contribute (Nielsen 2006; Strout 2011). Thus, they may be affected by the
virtual community content even if they do not contribute to the ongoing conversa-
tions.
On the contrary there is a more obvious group: the 'posters' who actively and publicly
participate in the community’s activities (Leshed 2005). A poster, gets actively in-
volved and contributes to the discussion with its own content, thus determining the
nature and value of the discussion. The poster invests his or her time to make a con-
tribution, expecting to receive recognition through it (as feedback from the community
and through knowing that he or she has helped another person). This is a distinct so-
cial context that the poster experiences in the community (Ridings et al. 2006).
Some people may not be that active as posters and post and create content regularly
but on the other hand they may not be that inactive as lurkers — there seems to be a
group that is more active than lurkers but less active than posters: the so-called low-
frequency posters or 'infrequent posters'. Their role is not as obvious as posters but
they are no less important. They represent some intermediate step between lurking
and posting and are seen as an additional group beyond. They are characterised by
being moderately active in online communities and posting irregularly (Ridings et al.
2006).
Despite the now high penetration of social media (DataReportal 2020a), there are
people who are not part of it: the 'non-users'. However, a distinction between non-
users and lurkers is important because lurkers are not non-users. Lurkers use the
technology and visit websites, social media etc. Non-users, conversely, are those
who for various reasons do not use communication technologies; they are an own
142
__________________________________________________________________________
group. These can be for a number of reasons, such as lack of financial resources or
skills, poor education, emotional reasons or simply because they do not like to use
them (Martin and Robinson 2007; Selwyn 2006).
At first, the literature preferred to focus on those who actively post, since lurkers were
occupied with a negative image. Many regarded lurkers as free-riders who took and
gave nothing back (Kollock and Smith 1996). Through the work of Nonnecke, Ridgins
and Edelmann, lurkers became more the subject of research and were brought into
the positive light. The experiment of Schlosser (2005) revealed that lurkers are less
affected by another's opinion than posters who adjust their attitude downward trig-
gered by another's negative opinion. The conducted literature review in the present
dissertation revealed that hardly any research can be found that deals with social
media activity in the context of herd behaviour along the purchasing process (cf.
4.3.2.3). Nevertheless, the results of Study III (cf. 6.4) revealed that this might be an
interesting aspect to be considered with regards to herd behaviour along the buying
decision process. Thus, the dissertation evaluates this aspect more deeply.
7.2.5 Personality Types
Since the work by Devaraj et al. (2008), it has also become evident in IS research
that individual differences, including personality traits, have an impact on consumers'
decisions and behaviour. They incorporated the five-factor model (FFM) of personal-
ity factors in the context of technology acceptance and found that FFM personality
dimensions can be useful predictors of users’ attitudes and beliefs.
There are many different types of personality assessments. One of the more ac-
cepted is the Myers–Briggs Type Indicator (MBTI) based on four bipolar discontinu-
ous scales: introversion–extraversion, and sensation–intuition. It classifies people
into sixteen different personality types based on the largest score obtained for each
bipolar scale. Meanwhile, the Big Five personality traits measured by the Neuroti-
cism-Extroversion-Openness Personality Inventory-Revised (NEO PI-R), which is
particularly with regard to all-important construct and predictive validity criteria supe-
rior to that of the MBTI (Furnham 1996). The Big Five framework enjoys considerable
support and has become one of the most heavily used and extensively researched
model in the academic research area on personality (John and Srivastava 1999).
Therefore, the Big Five personality model is also used in the present dissertation.
143
__________________________________________________________________________
The Big Five framework is a hierarchical model of personality traits and suggests that
most individual differences in human personality can be classified into five broad,
empirically derived domains. These 'Big Five' personality traits are broad dimensions
of individual differences in experience and behaviour: extraversion, agreeableness,
conscientiousness, emotional stability/neuroticism and openness to experience (Gos-
ling et al. 2003). Although different descriptions can be found in the research to de-
scribe these five factors, representative labels are as follows:
Extraversion is described as being sociable, gregarious, assertive, talkative, and ac-

tive (Barrick and Mount 1991). These individuals tend to be optimistic, enthusiastic,
and have a can-do attitude (Devaraj et al. 2008).
Agreeableness is associated with being courteous, flexible, trusting, good-natured,

forgiving, cooperative, soft-hearted and tolerant (Barrick and Mount 1991). It refers to
propensity of an individual to comply with others and is represented by a compas-
sionate interpersonal orientation (Devaraj et al. 2008).
Conscientiousness is a trait is represented by being careful, thorough, responsible,

organized, and planful (Barrick and Mount 1991). Conscientiousness individuals are
rule followers, and persistent, they are also associated with goal-directed behaviour
(Devaraj et al. 2008).
Emotional Stability or Neuroticism is commonly linked to feelings of anxiety, worry,

insecurity and depression, among others (Barrick and Mount 1991). People who
show neuroticism traits are less likely to try new experiences, and are more likely to
have self-efficacy and self-esteem issues and try to avoid new situations (Rosen and
Kluemper 2008).
Openness to Experience is associated with traits such as being imaginative, cul-

tured, curious, original, broad-minded, intelligent, and artistically sensitive (Barrick
and Mount 1991). Highly open people are creative and attracted by novelty and new
ideas (Devaraj et al. 2008).
Several rating tools have been developed to measure the Big Five dimensions. The
most comprehensive one is the 240-item NEO-PI-R by Costa and McCrae (1992)
which allows specific facets within each dimension and takes about 45 minutes to
complete. Due to the length of this version, several researchers have developed
shorter versions, for example: the well-established 44-item Big Five Inventory (BFI)
(John and Srivastava 1999), the 60-item NEO Five-Factor Inventory (NEO-FFI)
144
__________________________________________________________________________
(Costa and McCrae 1992) or the 100-item instrument (TDA) by Goldberg (1992).
Nevertheless, researchers recognised the need for an even briefer measure of the
Big Five (John and Srivastava 1999). To meet the need for a very brief measure,
Gosling et al. (2003) have developed and evaluated a 5- and 10-item inventory. Of
the two instruments, the Ten-Item Personality Inventory (TIPI) is psychometrically
superior, and it takes only about one minute to complete. It is recommended when a
short instrument is needed, e.g. if the personality traits represent only one dimension
in research (Gosling et al. 2003).
The Big Five model is also used in IS research as a standard tool for personality
measurement. In recent years, researchers have increasingly focused on the aspects
of individual personality differences in relation to different topics, e.g. the participation
in social networks and social media use, video games, virtual reality, and online
shopping (Correa et al. 2010; Hughes et al. 2012; Kober and Neuper 2013; Ryan and
Xenos 2011; Wilson et al. 2010; Witt et al. 2011). Svendsen et al (2013) investigated
the impact of the Big Five personality traits on behavioural intentions to use a tech-
nology, finding that perceived usefulness and ease of use fully mediated the effects
of conscientiousness and extraversion on intentions, while emotional stability directly
affected intentions. Other researchers have also considered the personality traits in
the context of herd behaviour, especially in the financial industry these aspects were
investigated; see, e.g. Bashir et al. (2013), and Chitra and Ramya Sreedevi (2011).
However, there is still little IS research that brings together the different aspects of
personality traits, herd behaviour and consumer buying decisions, which also be-
comes apparent in the results of the conducted literature review (cf. 4.3.2.3). The re-
sults of Pavlovic (2018) (Study III in chapter cf. 6.4) revealed that this might be an
aspect affecting herd behaviour along the buying decision process.
145
__________________________________________________________________________
7.3 Hypotheses
The following Table 14 gives an overview on the formulated hypotheses relating to

herd behaviour along the buying decision process.
Table 14: Overview of developed Hypotheses
Related Topic Hypotheses
Herd Behav- H 1a Observation of others in the pre-buying phase is positively asso-

iour in along ciated with imitating others.
the Buying De-
H 1b Observation of others in the buying phase is positively associ-
cision Process
ated with imitating others.
H 1c Observation of others in the post-buying phase is negatively as-

sociated with the intention to discontinue previous decisions.
Herd Behav- H 2a Consumer ratings exert a greater influence on herd behaviour

iour by Infor- than expert ratings in the pre-buying phase.
mation
H 2b Personal recommendations from trusted friends exert a greater
Sources
influence on herd behaviour than impersonal mass opinions on
social media in the buying phase.
H 2c Personal recommendations from trusted friends exert a greater

influence on herd behaviour than impersonal mass opinions on
social media in the post-buying phase.
Herd Behav- H 3a The degree of product involvement has an influence on herd be-
iour by Product haviour in the pre-buying phase.
Type
H 3b The degree of product involvement has an influence on herd be-
haviour in the buying phase.
H 3c The degree of product involvement has an influence on herd be-

haviour in the post-buying phase
Herd Behav- H 4a The level of social media activity has an influence on herd be-
iour by Social haviour in the pre-buying phase.
Media Activity
H 4b The level of social media activity has an influence on herd be-
haviour in the buying phase.
H 4c The level of social media activity has an influence on herd be-
146
__________________________________________________________________________
Related Topic Hypotheses
haviour in the post-buying phase.
H5 The social media activity level affects the influence of social

media on buying decisions.
Herd Behav- H 6a Personality has an influence on herd behaviour in the pre-buying

iour by Per- phase.
sonality Type
H 6b Personality has an influence on herd behaviour in the buying
phase.
H 6c Personality has an influence on herd behaviour in the post-

buying phase.
7.3.1 Herd Behaviour in the Buying Decision Phases
As outlined above, herd literature has suggested two conditions for herd behaviour to
occur: observation of others and uncertainty regarding the decision to be made. Fur-
thermore, herd behaviour is related to imitating others and discounting one's own in-
formation (Sun 2013). Both are given with regard to consumers' decisions concerning
a mobile communications product and therefore make herd behaviour commonplace.
With the dissemination of social media, people increasingly interact with each other
to discuss topics, exchange ideas and have easy access to other people's views or
actions. For example, one may see a satisfaction report regarding a mobile network
provider or that the new Samsung was rated as the best smartphone by other con-
sumers. Consumers might come across these kinds of information before or after
their purchase, which plays a not inconsiderable role in herd behaviour. Given that
people rely on the opinions of others, individual behaviour may be influenced by oth-
ers along the whole buying decision process (Zhou 2011). Imitating others denotes
that individuals follow others’ decisions or behaviour when herding.
Consumers are eager to reduce their uncertainty about a product along the whole
buying decision process (Kuß and Tomczak 2007). Uncertainty is defined as "an in-
dividual's perceived inability to predict something accurately" (Milliken 1987, p. 136).
Milliken (1987) identified three types of uncertainty: state uncertainty (lack of knowl-
edge how components of the environment might change), effect uncertainty (inability
147
__________________________________________________________________________
to predict what the impact of environmental events or changes will be), and response
uncertainty (inability to predict the likely consequences of a choice).
Based on the definition of uncertainty, it can be assumed that effect uncertainty and
response uncertainty occur in all three phases of the buying decision process; how-
ever, it might be higher in the pre-buying phase as the consumer has no or less own
information.
Therefore, the aim of the pre-buying phase is to gather information about products by
using different channels. Consumers spend their time searching for information and
evaluating alternatives (Kuß and Tomczak 2007). Making a buying decision is the
main mission in the buying phase — product alternatives are ranked and consumers
choose which product to buy (Häubl et al. 2010). To reduce uncertainty, consumers
tend to simply follow the decisions or recommendations made by predecessors
(Banerjee 1992; Bikhchandani et al. 1992). Accordingly, the following two hypotheses
are proposed to articulate that herd behaviour will affect the pre-buying and the buy-
ing decision phases.
H 1a Observation of others in the pre-buying phase is positively associated with

imitating others.
H 1b Observation of others in the buying phase is positively associated with

imitating others.
In the post-buying phase, consumers may compare their current consumption ex-
perience with what they had anticipated before making a purchase, and at the same
time, they may share their consumption experience through different communication
channels, e.g. review and rating platforms or social media in terms of tweets, blog
posts, or likes (Hennig-Thurau et al. 2004) to discuss the details of the product they
have purchased and recommend a product they feel satisfied with. Nevertheless, re-
sponse uncertainty, which are defined as inability to predict the likely consequences
of a choice might still occur, especially when observation of others may carry more
weight than their own information, leading them to accept others' adoption and thus
disregarding their own information when making decisions (Shen et al. 2014). How-
ever, in the post-buying phase another phenomenon may also occur, the so-called
'status quo bias'. A series of decision-making experiments show that individuals dis-
proportionately stick with the status quo, which means that there is a preference for
the current state or endowments (Kahneman et al. 1991; Samuelson and Zeckhauser
148
__________________________________________________________________________
1988). Therefore, when users actively discount their own preference on the foregone
technology in the herding condition, they are likely to think they are responsible for
the decision of giving up what they originally preferred (Sun 2013; Zou et al. 2015).
Thus, the following hypothesis is proposed:
H 1c Observation of others in the post-buying phase is negatively associated with

the intention to discontinue previous decisions.
7.3.2 Herd Behaviour by Information Sources
In contrast to the above three hypotheses, which relate to differences in herd behav-
iour between the particular buying decision phases, this section formulates three hy-
potheses related to the effectiveness of different information sources in consumers'
buying decisions.
Research on herd behaviour points out that multiple information sources have a dif-
ferent influence herding behaviour in consumers' buying decisions (cf. 4.3.1.2). The
cues examined in the study at hand for eliciting herd behaviour include consumer
and expert ratings, social networks and trusted friends. Whereas consumer and ex-
pert ratings are typical information sources in the pre-buying phase (Maas et al.
2014), the influence of consumer (Check24) versus expert ratings (Stiftung Warent-
est) is compared in the study at hand. According to the buying decision process
model of Foscht and Swoboda (2011), there is no active search for information or al-
ternatives in the post-buying phase (cf. 2.2.3). Therefore, the comparison with one's
own buying decision usually takes place rather randomly, e.g. in exchange with
friends. Hence, the influence of information from Facebook versus trusted friends' in-
put is evaluated in the buying and post-buying phases.
Bikhchandani et al. (1992) indicated that the source of information may also matter.
Generally, people follow previous adopters, believing in the 'wisdom of the crowd'.
However, they might follow a specific group of people more, such as IT experts or
fashion leaders. Then, they believe that they are more likely to have more accurate
information than the general public (Bandura 1986).
Expertise can be viewed as 'authoritativeness', 'competence' and 'expertness'

(Applbaum and Anatol 1972). Previous research shows that perceived level of exper-
tise positively impacts subject compliance with source recommendations (Crisci and
Kassinove 1973). Thus, an expert should be perceived as possessing more expertise
than other consumers. On the other hand, consumer reviews and ratings help con-
149
__________________________________________________________________________
sumers to evaluate offerings through the lens of previous customers. Consumers find
online user reviews helpful mainly from two aspects: a large volume of user reviews
and the valence of user reviews that provides information about the products and
services (Zhou and Duan 2016). Consumers are influenced more by 'collective intel-
ligence' than by a small group of experts (Surowiecki 2004). For example, Nielsen
(2015) state in their 'Global Trust in Advertising Survey' that WOM recommendations
remain the most credible. Hence, firms actively encourage users to provide feedback
and generate content. Following this line of research, the following hypothesis is pro-
posed:
H 2a Consumer ratings exert a greater influence on herd behaviour than expert

ratings in the pre-buying phase.
Some previous researchers consider anonymity as an advantage for eWOM (Chat-

terjee 2001; Goldsmith and Horowitz 2006), and others anticipate the eWOM on so-
cial media to be more influential on consumers’ buying decisions as it occurs be-
tween people who know each other (Moran and Muzellec 2017; Park et al. 2007).
Factors like information quantity, information readiness, detailed information and
dedicated information are reasons why consumers prefer anonymous reviews or rat-
ings more than friends' recommendations (Erkan and Evans 2018).
Users can better interpret recommendations from friends based on their knowledge
of these friends because friends share many experiences which might lead them to
develop similar views. Yet, recommendations from personal contacts, such as
friends, family, teammates, and co-workers, have a larger influence than impersonal
sources (Katz and Lazarsfeld 1955). Regardless of the number of reviews by the
crowd, a user may choose a rating value similar to her/his friends (Lee et al. 2015).
Dewan et al. (2017) add another perspective: popularity and proximity influence. As
stated above, popularity describes consumers' decision-making based on aggregate
information on the preferences and actions of others. Proximity influence describes
the role of social ties on individual consumption. In their study applied to the music
industry, they ascertained that the two types of influence are substitutes for one an-
other, and proximity influence, when present, will dominate the effect of popularity in-
fluence.
Indeed, this debate remains unresolved. Whether interaction among anonymous

people or interaction among people's friends and acquaintances is more influential on
150
__________________________________________________________________________
consumers’ buying decisions still remains undetermined; however, the influence of

both types was separately tested and found influential. Based on the discussions
above, the following two hypotheses are proposed:
H 2b Personal recommendations from trusted friends exert a greater influence on

herd behaviour than impersonal mass opinions on social media in the buying
phase.
H 2c Personal recommendations from trusted friends exert a greater influence on

herd behaviour than impersonal mass opinions on social media in the post-
buying phase.
7.3.3 Herd Behaviour by Product Type
For further investigation of the moderating effect of product types with regard to herd
behaviour, three hypotheses are here proposed to articulate that the level of product
involvement will affect herd behaviour along the buying decision process.
Previous works have shown that consumers are influenced by others in decision-
making. Herd behaviour is common in situations in which consumers gather product
information from other consumer choices and incorporate that into their own deci-
sion-making. Furthermore, the type of product can also influence consumer's deci-
sion-making process. Thus, existing studies have paid little attention to comparing
herd behaviour between the different level of product involvement. Since there
seems to be little research to be found on the different influence of herd behaviour on
high- and low-involvement products along the purchasing process, the dissertation at
hand sheds light on these aspects.
Consumers often have to make purchasing decisions with incomplete information

because they lack complete information on products and services. Additionally,
searching and seeking information is time consuming and can be costly, and there
are trade-offs between perceived costs and benefits of additional search (Stigler
1961). Thus, consumers are motivated to follow a buying decision process that re-
duce uncertainty. For example, the quality of a mobile service is difficult to evaluate
before adoption and requires one's experience. As a consequence, consumers are
motivated to make extensive use of various information sources, hoping to make
well-informed decisions (Gu et al. 2012; Luo and Zhang 2013), and are thus more re-
liant on the opinions of other consumers. Numerous studies, state that consumers of-
ten conduct pre-buying searches for high-involvement products (e.g., digital cam-
151
__________________________________________________________________________
eras) and visit eWOM during the search process (Gu et al. 2012; Luo and Zhang
2013). Lin et al. (2013) found in their study that product involvement has an moderat-
ing effect on buying intention. Chen (2008) argues that books as low-involvement
products that are relatively low priced, and where consumers spend less time and
cognitive efforts, elicit herd behaviour. Furthermore, Yi-Fen and Ya-Ju (2010) exam-
ined the herd effect and the moderating role of product involvement on bidder
choices in online auctions. They found that herding behaviour is stronger in high-
involvement than low-involvement participants. Thus, the following hypotheses are
proposed:
H 3a The degree of product involvement has an moderating effect on herd behavi-

our in the pre-buying phase.
H 3b The degree of product involvement has an influence on herd behaviour in the

buying phase.
Additionally, the level of product involvement seems to play an important role in the
post-buying phase. Richins and Bloch (1991) found that consumers with high product
involvement show higher disconfirmation in satisfaction changes than consumers
with low-involvement. Further, Graff et al. (2012) investigated the post-buying of mo-
bile phone buying, among other aspects, as well as the influence of other people's
opinions and product involvement on the cognitive dissonance of consumers. They
found that consumers follow the opinions of others who are close to them and prod-
uct involvement impacts consumers' cognitive state. Following this line of research,
the following hypothesis is proposed:
H 3c The degree of product involvement has an influence on herd behaviour in the

post-buying phase.
7.3.4 Herd Behaviour by Social Media Activity
Considering the aspects of social media activity in the context of herd behaviour
along the purchasing process, in the following three hypotheses are proposed. In ad-
dition, it is examined whether the different social media types differ in their impact on
the different activity groups.
The increasing spread of social media has also raised the level of interaction be-
tween people. Hence, the transparency of other people's opinions has significantly
152
__________________________________________________________________________
increased, and likewise the opportunities to interact with others or to take a shortcut
to follow other people's opinions increased.
However, not all Internet users are the same and different groups must be distin-
guished. Lurkers are known as the silent participants in virtual communities; never-
theless, they are part of virtual communities, even if passive, and are still consumers
of its knowledge and content. Therefore, they can react to advertising or selling ac-
tivities. Thus, they may be affected by the virtual community content even if they do
not contribute to the ongoing conversations (Ridings et al. 2006). Nonnecke et al.
(2004) thus argue that both lurkers and infrequent posters should be considered in
research alongside the group of posters.
Thus, the question arises of whether herd behaviour in the purchasing process is the
same for all groups, or whether there are differences here. Ho et al. (2017) observed
in their research systematic differences in rating behaviours between posters and in-
frequent posters. Further, the experiment of Schlosser (2005) revealed that lurkers
are less affected by another's opinion than posters who adjust their attitude down-
ward triggered by another's negative opinion.
The research of Bartikowski and Walsh (2014) investigates why and how product re-
views from consumer opinion platforms affect individual users’ buying behaviour and
whether these relationships depend on user type: posters or lurkers. They found that
lurkers are much less influenced by the opinions/product reviews of others than post-
ers. The reason for this is that posters are more likely to conform to community
norms because they identify more with the community. Lurkers, on the other hand,
are less involved in the community and their sense of 'we' is not very strong. Rather,
they use the opinion platform primarily to fulfil their functional goals in terms of gath-
ering information needs (Bartikowski and Walsh 2014). According to the findings of
Huang et al. (2012) it can be said that lurkers use the opinions of others primarily as
a source of information. Infrequent posters rely more on the opinion of others and
posters, as the most active group; meanwhile, the primary posters and creators of
content refer to others' opinions only as one source of information and seldom accept
the opinion of others.
Following this line of research, the following hypotheses are proposed considering
the aspects of the activity level and herd behaviour along the buying decision proc-
ess:
153
__________________________________________________________________________
H 4a The level of social media activity has an influence on herd behaviour in the
pre-buying phase.
H 4b The level of social media activity has an influence on herd behaviour in the
buying phase.
H 4c The level of social media activity has an influence on herd behaviour in the
post-buying phase.
The existing literature suggests that posters and lurkers join online communities
equally because they share common interests and want to improve their understand-
ing of a topic (Preece et al. 2004). Lurkers do not engage in posting because they
mainly want to satisfy their informational needs (Preece et al. 2004). In contrast,
posters try to satisfy social-emotional needs and present themselves, improve their
image or promote their status and convince others of their opinion (Balasubramanian
and Mahajan 2001; Jang et al. 2008). As each social media type offer different op-
tions for self-communication and self-marketing, it can be assumed that the type of
social media then can also affect the different user groups in the buying decision
process differently. Therefore, the following hypothesis is proposed:
H5 Different social media types have a differing impact on the individual activity
groups.
7.3.5 Herd Behaviour by Personality Type
Considering the aspects of personality traits in the context of herd behaviour along
the buying decision process, in the following three more hypotheses are proposed.
For this purpose, the established Big Five framework is used. It includes the person-
ality traits of extraversion, agreeableness, conscientiousness, emotional stability or
neuroticism, and openness (Devaraj et al. 2008).
The consideration of personality traits in relation to herd behaviour in the buying de-
cision process does not seem to have been extensively investigated in IS research
so far (cf. 4.3.2.3), Although some researchers point out that a closer examination of
this aspect would be relevant; e.g. Chen (2008) emphasises at the end of his study
that it would be worth examining the influence of personality traits on herd behaviour.
Researchers who have addressed the aspect of personality in the context of con-
sumers' decisions mostly stem from other research streams, such as economics or
154
__________________________________________________________________________
finance. Some of the studies that support the discussion in relation to herd behaviour
in the buying decision process are presented below.
The study of Madhavi and Rao (2018) indicates the influence of Big Five factors of
personality on buying behaviour including post-buying in the automobile sector. Do-
bre and Milovan-Ciuta (2015) highlighted in their study the influences of personality
on important decision-making variables linked to the customer’s online visiting, buy-
ing and post-buying process. They discovered that personality influences the pre-
buying and buying process in terms of priority and importance given to selection cri-
teria of online stores. Further, the assessment of satisfaction, of post-buying and post
visiting processes differs depending on the personality traits of the consumers (Dobre
and Milovan-Ciuta 2015). The study of Chitra and Ramya Sreedevi (2011) shows that
investors' personality traits have an impact on their decisions and also have a strong
impact on determining the method of investment.
The research conducted by Svendsen et al. (2013) investigates the degree to which
users’ assessments of the core constructs of Technology Acceptance Model are in-
fluenced by their personality and indicates that personality influence behavioural in-
tention both directly and mediated through the Technology Acceptance Model beliefs.
Overall, it can therefore be said that some researchers have looked at the influence
of personality differences on purchasing behaviour, taking various aspects into ac-
count. Although there is less literature on herd behaviour in the buying decision proc-
ess to be found in this context. Following the line of existing research, the following
hypotheses are proposed:
H 6a Personality has an influence on herd behaviour in the pre-buying phase.
H 6b Personality has an influence on herd behaviour in the buying phase.
H 6c Personality has an influence on herd behaviour in the post-buying phase.
7.4.1 Procedure and Sample Profile
To test the hypotheses, an online survey was conducted. The data was collected us-
ing an online self-administered and voluntary survey. A split-sample approach was
used, where participants were randomly assigned to one of three groups: a control
group, and treatment groups 1 and 2. Data was collected anonymously. The ques-
tionnaire was programmed in Unipark Software, and a pre-test with seven partici-
155
__________________________________________________________________________
pants was carried out to test clarity, language usage and comprehensiveness before
the survey was finally launched. Unipark offers the possibility to testers to comment
on the questionnaire to provide their feedback. Accordingly, parts of the question-
naire were rephrased to improve clarity, and the socio-demographic questions were
put at the end in order to reduce a possible termination at the beginning. The survey
was disseminated using university/business school mailing lists and through social
networking sites such as Facebook.
A total of 867 participants (57.2% response rate) from different universities and busi-
ness schools in Germany responded to the survey. The use of students was consid-
ered appropriate for this study since they actively engage in eWOM communications
about products and services (Bailey 2005), and use social media the most compared
to other age groups (Social 2018). The majority of participants were between the
ages of 18 and 29 (90.8%); 66.7% were female and 33.3% were male.
7.4.2 Research Design
In the following, the research design is outlined in detail. First, the structure of the
survey is presented. This is followed by the description of the comparison groups,
and finally, a depiction of the used items and scales.
7.4.2.1 Structure of the Survey
The survey is structured in six sections (cf. Table 15). At the beginning of the survey,
the respondent is informed about the context of the study, i.e. that it is a survey within
the framework of a promotion and that purchasing decisions are involved. The re-
spondent is given the name of the contact person for the study and is informed that
at the end of the survey he can take part in a lottery for an Amazon voucher. Fur-
thermore, it is pointed out that participation is anonymous and that the survey takes
5-10 minutes.
In the second section, the general description of the personality is dealt with first.
Then, the average activity in social media and the type of activity in it are discussed.
The third section deals with the product involvement for the respective product
(smartphone, mobile phone contract, or entertainment service), according to the
automatically assigned product category.
In the following section, the respondent is guided through a buying decision process.
The respondent is first asked about his or her most important sources of information
156
__________________________________________________________________________
when buying the allocated product. In this section, the respondent is automatically
assigned to a comparison group in the background, which receives different informa-
tion during the buying decision process. The process is divided into three phases:
pre-buying, buying and post-buying phase.
Then, the demographic data of the subjects are collected.
The survey ends with the final page, where the respondents are thanked for their par-
ticipation and have the opportunity to make comments on the survey. Finally, they
can take part in the lottery. The complete questionnaire programmed in Unipark can
be found in the Appendix E.1.
Table 15: Structure of the Survey
Section Content
Introduction  Welcome
 Context of the survey
 Contact person
 Assurance of anonymity
 Expected time needed to reply to the questionnaire
Segmentation  Personality type

 Social media activity
Product  Product type (automatic, random assignment to

one product type):
o Smartphone
o Mobile contract
o Entertainment service
 Product involvement
Buying Decision Process per  Information source

product  Buying decision process per comparison group
(automatic, allocation to one of the three compari-
son groups):
o Pre-buying phase
o Buying phase
o Post-buying phase
157
__________________________________________________________________________
Section Content
Demographic Data  Age

 Gender
 Household type
 Profession
 Educational attainment
End  Thanks for participation and possibility for feed-

back
 Possibility to participate in a lottery
Since the study is a survey with three comparison groups, some conditions had to be
solved in the programming of the questionnaire in order to generate different com-
parison groups. The survey was programmed in such a manner that the respondent
is automatically assigned to a product category after the introduction and segmenta-
tion section. This ensured that a balanced number per product category is available
for further analysis. Furthermore, the test persons were assigned to different com-
parison groups, which received different information during the buying decision proc-
ess. The illustration (see Figure 31) gives an overview of the distribution process to
the comparison groups.
Control group Buying decision phases
Even Treatment group 1

Smartphone spread
Buying decision phases
Treatment group 2 Buying decision phases
Random Mobile Even Treatment group 1 Buying decision phases

Respondent assignment contract spread
Entertainment Even Treatment group 1 Buying decision phases

service spread
Figure 31: Overview of the Distribution Process to the Comparison Groups
158
__________________________________________________________________________
7.4.2.2 Comparison Groups
The study uses three comparison groups: one control group and two treatment
groups. The participants were randomly assigned to one of the groups in order to
achieve an even distribution of the groups. Each respondent was guided through a
hypothetical purchasing decision process (pre-buying, buying, post-buying) for a mo-
bile communications product (smartphone, mobile contract, or mobile entertainment
service). Before each decision step, subjects were presented with a choice of two
products, A or B, each with ratings from different information resources (consumer
reviews, expert reviews, trusted friends or Facebook); the control group received no
external information during the whole buying decision process.
As review and rating platforms and expert reviews are typical information sources in
the pre-buying phase to gather information (Maas et al. 2014; Mudambi and Schuff
2010), they are chosen as information sources in this decision step. The treatment
groups each received star ratings from 'Check24' (consumer review and rating plat-
form) and 'Stiftung Warentest' (expert review and rating platform/foundation for prod-
uct testing), with opposite preferences for A or B. Stars varied markedly (2 out of 5
stars or 4 out of 5 stars) among the products to reflect the preferences of the respec-
tive source. The stars reflected possible real situations related to online consumer
evaluation or expert evaluation of mobile communications products. In order to ana-
lyse whether the influence of consumer reviews or expert opinions is stronger, two
pieces of information were given in the pre-buying phase. First, treatment group 1 re-
ceived information from 'Check24' and treatment group 2 received information from
'Stiftung Warentest'. In the second step, the respondents who had previously re-
ceived information from 'Check24' then received information from 'Stiftung Warentest'
with the opposite ratings, and the other treatment group received information in the
opposite order (Table 1). The two information rounds are chosen to, first, identify the
general difference in the information source, and second, in the case of opposing
statements, identify which source has the stronger influence on consumers' decision-
making. The overall preference choices of two products were operationalised by ask-
ing: "After you read the information, what is your preference?" Responses were given
using the scale: A, Neutral, or B.
159
__________________________________________________________________________
Table 16: Provided information in the Pre-Buying Phase
Control Group Treatment Group 1 Treatment Group 2
• No input from information • Information from Check24 • Information from Stiftung

resources (review and rating platform) Warentest (foundation for
A: 4 from 5 stars product testing)
 request for preference
B: 2 from 5 stars A: 2 from 5 stars
B: 4 from 5 stars
• Information from Stiftung
Warentest (foundation for • Information from Check24
product testing) (review and rating platform)
A: 2 from 5 stars A: 4 from 5 stars
B: 4 from 5 stars B: 2 from 5 stars
 request for preference  request for preference
In the next decision step, the buying phase, the respondents were asked how they
would decide if they had to make a buying decision now. Before this final hypothetical
buying decision, Treatment group 1 received the information that product B had re-
ceived twice as many likes on Facebook with corresponding positive comments.
Treatment group 2 received the information that the majority of their trusted friends
had chosen product B and that they were very enthusiastic about it. Trusted friends
were chosen as an information source because they are a typical source for consul-
tation before making a choice (Lee et al. 2015). By contrast, Facebook is a rather im-
personal mass opinion source, where opinions are reflected in the number of likes.
To examine the different influence of the two information sources, the final buying
decision with regard to the two mobile communications products were operational-
ised by asking: "After reading the information and if you were to take a buying deci-
sion now, what would you buy?" Responses were given using the scale: A, Neutral,
or B.
160
__________________________________________________________________________
Table 17: Provided information in the Buying Phase
• No input from information • Information from Facebook • Information from trusted

resources A: 20.000 likes friends
B: 40.000 likes A: No recommendation
 request for buying deci-
B: Clear recommendation
sion  request for buying decision
 request for buying decision
In the last step, the post-buying phase, respondents were informed that there was an
offer they could make use of. It was called "Love it, or exchange it", which means that
the respondent had the opportunity to return the product within four weeks. In addi-
tion, the respondents were informed that they should imagine that they have been
using the product they had previously chosen already for three weeks, and that they
were very satisfied with it. Before they have to make a decision if they want to make
use of the offer, treatment group 1 (who got the information from Facebook during
the buying phase) got the information that most of their trusted friends had decided
for the other product than themselves and are very enthusiastic about it. Treatment
group 2 (who had received the information from their trusted friends during the buy-
ing phase) now received the information that the other product on Facebook had re-
ceived twice as many likes and numerous positive comments. The control group also
received the "Love it, or exchange it" offer, but without external information concern-
ing the products. The questionnaire was programmed in such a manner that the re-
spondent was always advertised the product for which she/he had not previously de-
cided, which was done in order to see whether she/he would revise the previous buy-
ing decision. The likelihood to revise the previously made buying decision was opera-
tionalised by the question: "How likely would you make use of the return option?" Re-
sponses were made using a seven-point Likert-type scale (1 very unlikely, 7 very
likely).
161
__________________________________________________________________________
Table 18: Provided information in the Post-Buying Phase
• No input from information • Information from trusted • Information from Facebook

resources friends 40.000 likes for the oppo-
site of the chosen option
 request for likelihood re- Clear recommendation for
garding the return option the opposite of the cho- 20.000 likes for the op-
sen option tion that was chosen
request for likelihood re-  request for likelihood

garding the return option regarding the return op-
tion
7.4.2.3 Items and Scales
Here, the used items and scales are presented.
Information Sources
To identify the relevant information sources within the buying decision process the
respondents were asked for their most important and second most important informa-
tion sources. The respondents were asked to indicate which source of information is
the most important for them before buying a smartphone or entering into a mobile
phone contract or entertainment service (depending on the group they were assigned
to). To avoid bias by the order of the answers, the answer options were rotated and
only a single choice answer was possible. The provided list of information sources is
mainly inspired by Maas et al. (2014), MindTake (2017), and Pavlovic (2018), and in-
cludes the following sources:
 Social networks (e.g. Facebook, Twitter, LinkedIn)

 Expert (e.g. Stiftung Warentest)
 Online Forums
 Comparison platforms/review and rating platform (e.g. Check24)
 Blogs
 Family
 Video portals (e.g. YouTube)
 Friends
 Acquaintances/work colleagues
162
__________________________________________________________________________
 Others
In order to identify reasons behind the choice of information sources, the following
question was posed: "For what reasons would this source be most important to you?"
The respondent could give two reasons and the answer choices were rotated. The
provided list is mainly inspired by Maas et al. (2014) and Pavlovic (2018), and in-
cludes the following items:
 Transparency
 Credibility
 Objectivity
 Trustworthiness
 Personal relationship
 Speed
 Actuality
 Mass opinion
 Convenience
 Others
After the respondents were asked about the most important sources of information in
the buying decision process and the reasons for this, they were asked to add their
second most important source with the corresponding reasons. Thereby, the same
answer options were provided as most important information source. Since consum-
ers usually use several sources of information in the buying decision process, the two
most important ones should be identified here.
Product Involvement
Different product types from the mobile communications industry are chosen for fur-
ther evaluation in this dissertation at hand. Smartphones and mobile phone contracts
are among the standard products of mobile phone providers in Germany, but enter-
tainment services are now also among the standard services offered by providers
(Bundesnetzagentur 2019a). However, the various mobile phone products may well
differ in their product involvement, as involvement depends on individual perception
of consumers based on their inherent needs, values and interests (Zaichkowsky
1985). It refers to the level of interest or attention that a consumer pays to a product
(Cohen 1983; Mitchell 1979).
163
__________________________________________________________________________
The developed scale of Laurent and Kapferer (1985) was used to determine the in-
volvement level per product. They empirically showed that five independent involve-
ment determinants are to be considered across the most diverse product groups. The
following five items were offered to the participants to describe the product for them-
selves:
 Interest in the product,

 Fun to use
 Personal expression possibilities; identification with the product
 Perceived degree of risk; 'probability of falling into it'
 Perceived risk in terms of risk costs, e.g. in the event of loss or damage
The following question was asked to identify the degree of product involvement of
each subject, depending on the product group. Thereby, the respondent was ran-
domly allocated to a product, e.g. for smartphone the question was asked: "If you
were to buy a new smartphone for yourself — how important would the following as-
pects be to you?" Accordingly, the question was raised for the other products. Re-
sponses were made using a seven-point Likert-type scale (1 very unimportant, 7 very
important).
Social Media Activity
The level of activity in social media can vary greatly from person to person. For ex-
ample, there are very active people ('posters'), very passive people ('lurkers') and
those in between who participate from time to time ('infrequent posters'). Previous re-
searchers have found that the behaviour of the different groups differs (e.g. (Ridings
et al. 2006)). In order to be able to analyse the effect of the activity level in social
media on herd behaviour in the buying decision process in further analyses, the test
persons were asked about their activity level. Thereby, existing scales from literature
were reviewed and items were carefully adapted or developed. According to Ridings
et al. (2006), posting frequency is the key factor in the determination of lurker status
and lurkers do not post or post very infrequently. In previous research, posting be-
haviour was measured by the self-reported number of new posts per month, the
number of responses to other posts per month, and reading rates or creation of con-
tent (Leshed 2005; Ridings et al. 2006). As this research was done some years ago
and the penetration and usage of social media has increased by over 80% from 2015
164
__________________________________________________________________________
to 2020 (DataReportal 2020a), the questions about the frequency of usage per month
were adapted to questions about usage per week and per day.
The questions were structured in three blocks relating to active content creation, par-
ticipation in content and pure consumption of content. Considering the entire spec-
trum of social media, for example social networks (e.g. Facebook, Xing), comparison
portals (e.g. Check24, Amazon), video portals (e.g. YouTube), blogs (e.g. Twitter) or
forums the respondents were asked to specify the frequency of use, either per day
or, in the case of less than once per day, the frequency per week.
 First, participants were asked about how often they created their own content
(e.g. texts, videos, photos, own reviews).
 Then, they were asked to specify their participation in social media through
sharing, commenting (e.g. 'like', 'dislike') or rating (e.g. start ratings) content.
 The third question related to the passive consumption (e.g. reading, watching,
listening) of social media content.
Personality Types
The Big Five framework, well established in IS research, was used in this dissertation
to consider personality differences. The five personality factors: extraversion, agree-
ableness, conscientiousness, emotional stability/neuroticism and openness to ex-
perience are measured by using the 10-item scale (TIPI) as the recommended short
instrument by Gosling et al. (2003). This scale was developed as a short measure of
the five major dimensions of personality. The TIPI has an adequate level of validity,
reliability and external correlations. It can therefore be used as a proxy for the longer
Big Five version (Gosling et al. 2003). Furthermore, the German version of the ques-
tionnaire was used, which was developed by Rammstedt and John (2007) to facilitate
their cross-cultural research.
The respondents were asked to answer the following question: "How well do the fol-
lowing statements describe your personality?" using a seven-point Likert-type scale
(1 disagree strongly, 7 agree strongly). To avoid bias by the order of the answers, the
answer options were rotated.
 I see myself as someone who:

o is reserved
o is generally trusting
165
__________________________________________________________________________
o tends to be lazy
o is relaxed, handles stress well
o has few artistic interests
o is outgoing, sociable
o tends to find fault with others
o does a thorough job
o gets nervous easily
o has an active imagination
7.4.3 Data Preparation and Analytical Methods
Before testing the hypotheses, the data were first cleaned up to ensure a stable data
set. First, only respondents were included who had answered all decision questions;
674 participants remained, which corresponds to 77.7% of the total number of par-
ticipants. Secondly, all participants (n=33) who had a processing time of less than
five minutes (5% quantile) were not included. Since the average processing time
(median) was 8.29 minutes, it was assumed that a processing time that was over
40% less than the average processing time would not produce meaningful results.
After both cleansing rounds, a dataset of 641 remained for further analysis (Table
19).
Table 19: Data Cleansing
Respondents participated in the survey 867
Respondents answered all questions 77.7%
Data set after first data cleansing 674
Respondents with a processing time lower than five

5.1%
minutes
Data set after second data cleansing 641
H1 has the aim of examining whether herd behaviour occurs or does not occur along
the buying decision phases. Due to the research design of two treatment groups (1
and 2) each receiving different information, it was necessary to aggregate treatment
group 1 and 2 to one comparison group to avoid misinterpretation. This was neces-
sary before testing hypothesis H1. As the control group did not receive any external
166
__________________________________________________________________________
information, therefore, there was no difference between pre-buying and buying deci-
sion phase, thus the control group is the same for the analysis of both steps. Fur-
thermore, with regard to H1, H3, H4, H5 and H6 data from all groups was dichoto-
mised into "Neutral" and "A or B". In addition, a group indicator was required to define
the affiliation to the control or comparison group. To test H2 dichotomisation was not
necessary as the data is based on the control group and the two treatment groups,
which is relevant for the analysis concerning different information sources and their
influence on herd behaviour in the buying decision process.
To test if external information has an impact on the pre-buying and buying decision
process, the Pearson´s chi-squared test of independence test was used to determine
whether there is a significant difference between the expected frequencies and the
observed frequencies (Pearson 1900). In the case of independence, the same deci-
sion ratios would be expected in both groups. In the case of low sample sizes
Fisher´s exact test result is reported (Fisher 1922). The effect size of a chi-square
test is reported as phi coefficient in case of 2x2 contingency tables, and as Cramer´s
V in case of comparison of variables with more than two values. In the post-buying
phase, the respondents were not asked to choose a product A or B but rather to indi-
cate their willingness to exchange their product. Since the expressions ('very unlikely'
to 'very likely') are ordinal, the decision to exchange can also be understood as a
metric value and a mean value comparison was carried out. The t-test could not be
used, as the Shapiro–Wilk test showed a significant deviation from the normal distri-
bution assumption for all groups (p < 0.001) (Shapiro and Wilk 1965). Therefore, the
nonparametric Mann–Whitney U test was used to verify the significance of the corre-
spondence of two distributions (Wilcoxon 1945). As effect size measurement eta-
squared is reported. For the comparison of more than two groups/distributions the
Kruskal–Wallis test was applied (Kruskal and Wallis 1952). All analyses in this study
were performed using SPSS 25.
7.5 Descriptive Results
7.5.1 Demographics
Since the target group were students, 87.2% of the test persons stated to be stu-
dents. In terms of educational attainment, 63.4% said they had a high school degree
and 34.5% a university degree. The majority of participants were between the ages
167
__________________________________________________________________________
of 18 and 29 (90.8%); 66.7% were female and 33.3% were male. Most of them lived
either in a flat share (30.6%), with their partner (25.6%) or with their family (24.6%).
7.5.2 Information Sources
Figure 32 lists the frequency distribution of the most important and second most im-
portant information sources during the buying decision process. The expert evalua-
tion thus proves to be the most important source of information for a good 50% of the
participants, followed by review/rating platforms with 47%. Friends are also consid-
ered to be an important source of information by more than 44% of the participants. A
total of 15 people cited both times the same source, i.e. both as the most important
and the second most important source of information.
26.7%
Expert (e.g. Stiftung Warentest)
23.4%
Comparison platforms / 25.6%

review-/ rating platform (e.g. check24) 21.4%
22.8%
Friends
21.2%
7.8%
Family
10.5%
5.5%
Online Forums
8.3%
4.5%
Acquaintances / Work colleagues
7.0%
3.0%
Video portals (e.g. youtube)
3.4%
3.0%
Others
1.4%
Social networks 0.9%

(e.g. FB, Twitter, Linked-In) 2.2%
0.3%
Blogs
1.2%
n = 641 Most important Second-most important
Figure 32: Information Sources within the Buying Decision Process
Table 20 gives an overview of the received responses under 'Others' for the most im-
portant and second most important information sources within the buying decision
process.
168
__________________________________________________________________________
Table 20: Other Information Sources within the Buying Decision Process
Most important Second most important
 IT expert  Store of the provider

 Technical information on products, own  Own experiences
opinion  Homepage of the provider
 various sales consultations  Own research
 Advice programmes, e.g. WISO  Intuition
 Provider portals, e.g. Telekom.de
 Consulting
 Own research
 Technical advice
 Financial tip
 Mobile Competence Center GmbH
 Store of the provider
 Website of the provider
 I make my own comparisons and cal-
culate what is most intelligent
 Own research from most of the above
mentioned
 Google
 Mix of video portals, comparison por-
tals and friends/acquaintances
In Table 21, the reasons for the choice of information sources within the buying deci-
sion process are presented. Participants were allowed to give multiple responses, so
numbers do not sum to 100%.
According to this, respondents seem to be consulted social networks mainly for

speed, 50% indicated this reason as important. Experts are mainly concerned with
credibility (51.5%), objectivity (57.9%) and trustworthiness (46.8%). In online forums,
the reasons are distributed relatively evenly, with objectivity and speed being the
most likely reasons. Review/rating platforms seem to be used as a source of informa-
tion mainly because of mass opinion (31.1%), objectivity (28.7%), and transparency
(25.6%). Only 3 persons mentioned blogs as the most important source of informa-
tion, the reasons are trustworthiness (50%), personal relationship (50%) and speed
169
__________________________________________________________________________
(50%). The reason to state the family as the most important source of information is
mainly trustworthiness (68%) and credibility (46%), but also personal relationship
(42%). Video portals are listed as actuality (47.4%) and can be reached quickly.
Friends are also mentioned mainly for the reasons of trustworthiness (57.5%) and
credibility (46.6%), but also the personal relationship (38.4%) plays a role here again.
A comparable picture emerges among the acquaintances, where trustworthiness and
credibility are also considered to be decisive reasons. The reasons for the second
most important information sources do not really differ and are therefore not listed
separately.
Table 21: Reasons for choosing the Information Sources (in%)
Most important
Friends
Social networks
Experts
Family
Acquaintance
Video portals
Others
Online forum
Review/rating
Blogs
Information
Source
Reasons for In-

formation
Source
Transparency 0 15.2 11.4 25.6 0 4.0 15.8 6.8 3.4 26.3
Credibility 33.3 51.5 22.9 13.4 0 46.0 15.8 46.6 44.8 16.3
Objectivity 0 57.9 28.6 28.7 0 6.0 21.1 8.9 17.2 31.6
Trustworthi- 16.7 46.8 14.3 15.2 50.0 68.0 26.3 57.5 58.6 15.8
ness
Personal rela- 0 0.6 2.9 1.8 50.0 42.0 5.3 38.4 17.2 15.8
tionship
Speed 50.0 0 28.6 13.4 50.0 0 36.8 5.5 3.4 15.8
Actuality 33.3 5.3 20.0 25.6 0 0 47.4 6.8 10.3 26.3
Mass opinion 33.3 0 22.9 31.1 0 0 0 0 3.4 5.3
Convenience 16.7 0.6 20.0 17.7 0 4.0 15.8 6.8 10.3 5.3
Others 0 0.6 0 1.2 0 8.0 5.3 0 0 15.8
Total (n) 11 305 60 285 3 89 36 259 188 35
170
__________________________________________________________________________
7.5.3 Social Media Activity
To be able to display the results it was necessary to generate three binary variables
on social media usage behaviour.
If the three items "Creation of content per day", "Participation in social media per
day", and "Consumption of social media per day" were answered with "at least once
a day", the persons were classified as "Poster", "Participant" and "Lurker", respec-
tively.
It also occurred that respondents had overlaps in their answers, which means that
the three individual binary variables are not disjunct, i.e. many have described them-
selves as both a poster and participant and/or a lurker. Therefore, the 'highest' alloca-
tion was chosen to trump the others. Thus, if someone called himself a poster, he is
categorised as a poster, regardless of whether he also called himself a participant
and/or a lurker. Participant is someone who does not call himself a poster, but a par-
ticipant. A lurker is someone who has stated "less than 1x per day" for both poster
and participant, according to the following assignment:
user = 3, if Poster = 1
user = 2, if Poster = 0 & Participant = 1
user = 1, if Poster = 0 & Participant = 0 und Lurker = 1
user = 0, if Poster = 0 & Participant = 0 und Lurker = 0
If this procedure is applied, the following results are obtained:
The majority of interviewed persons are participants (50.1%), followed by consumers

('Lurkers') (34.5%). Only very few participants (9%) are high-activity users, called
'Posters'. In fact, 6.4% stated to be non-users, i.e. they stated “less than 1x per day”
for creation, consumption, or participation in social media.
171
__________________________________________________________________________
Participants 50.1%
Lurker 34.5%
Poster 9.0%
Non-User 6.4%
n = 641
Figure 33: Overview of Social Media Activity
7.5.4 Product Involvement per Product Type
To evaluate the product related results, first, new variables were calculated for the
five product involvement queries ('interest in the product', 'fun to use', 'personal ex-
pression possibilities/identification with the product', perceived degree of risk; "prob-
ability of "falling into it", 'perceived risk in terms of risk costs. e.g. in the event of loss
or damage). In the following way, the items on product involvement were recorded on
the 7-point Likert scale:
 1 or 2 was selected => 'Low Involvement'

 3, 4 or 5 was selected => 'Neutral'
 6 or 7 was selected => 'High Involvement'
The responses for the product category smartphone are presented in Table 22.
Table 22: Product Involvement of a Smartphone
Smartphone Low Neutral High Involve-

Involvement ment
(n= 210)
Interest in the product 1.9% 41.4% 56.7%
Fun to use 3.8% 32.9% 63.3%
Personal expression possibili- 25.7% 50.5% 23.8%

ties/identification with product
172
__________________________________________________________________________
Smartphone Low Neutral High Involve-

Involvement ment
(n= 210)
Perceived degree of risk; "prob-

3.8% 53.3% 42.9%
ability of "falling into it"
Perceived risk in terms of risk

costs. e.g. in the event of loss or 14.4% 62.7% 23.0%
damage
The responses for the product category 'Mobile Contract' are presented in Table 23.
Table 23: Product Involvement of a Mobile Contract
Mobile Contract Low Neutral High Involve-

(n= 209) Involvement ment
Fun to use 9.1% 47.6% 43.3%
Personal expression possibili-

ties//identification with the 28.7% 56.9% 14.4%
product

5.3% 36.8% 57.9%

damage
173
__________________________________________________________________________
The responses for the product category 'Entertainment Service' are presented in Ta-
ble 24.
Table 24: Product Involvement of an Entertainment Service
Entertainment Service Low Neutral High Involve-

(n= 221) Involvement ment
Fun to use 0.5% 10.9% 88.7%
Personal expression possibili-

ties//identification with the 13.1% 68.8% 18.1%
product

3.2% 47.3% 49.5%

damage
Subsequently, an overall variable for product involvement was created for each
product type according to the following procedure:
1. Sum of the 5 individual questions (with 3 categories)

2. Division of the total values again into the 3 categories 'Low-Involvement', 'Neu-
tral' and 'High involvement' according to the following scheme:
 Sum ≤ 7: 'Low-Involvement', i.e. a maximum of 2 of the 5 questions were
marked 'Neutral', all others with 'Low-Involvement'
 Sum between 8 and 12: 'Neutral', i.e. between 2 answers 'Low involve-
ment', 3 answers 'Neutral' and 3 answers 'Neutral', 2 answers 'High-
Involvement'
 Sum > 12: 'High-Involvement'
The product involvement per product type is depicted in Figure 34. The 'High-
Involvement' rate is therefore highest for the product type 'Entertainment Service' at
174
__________________________________________________________________________
53.2% and lowest for 'Mobile Contract' at 28.7%. The 'Neutral' rate is highest for
'Smartphone'.
1.4%
Smartphone 68.6% 30.0%
3.3%
Mobile Contract 67.9% 28.7%
1.4%
Entertainment Service 45.5% 53.2%
100%
n = 641 Low-Involvement Neutral High-Involvement
Figure 34: Product Involvement per Product Type
7.5.5 Personality Types
As described above the 10-item questionnaire was used to assess the personality
type of the participants. Each item is requested using a 7-point Likert scale with val-
ues 1 (does not apply at all) to 7 (does apply completely). Calculating the mean of
two of the items result in five numeric variables defining the five personality types. To
allocate the participants to one of the personality types, a dichotomisation is done in
the following kind: only participants with values of at least 6 are allocated to this spe-
cific personality type. As the majority of participants are allocated to more than one
personality type (i.e. the assignment was not perfectly selective), only these 232 par-
ticipants, who are allocated to exactly one personality type, are included in the follow-
ing analysis. The distribution of these 232 participants is depicted in Figure 35.
175
__________________________________________________________________________
Conscientousness 32.3%
Openness to Experiences 32.3
Extraversion 22.0%
Emotional Stability 8.2%
Agreeableness 5.2%
n = 232
Figure 35: Distribution of Personality Types in the Sample
7.6 Statistical Analysis
7.6.1 Herd Behaviour along the Buying Decision Process
To test H 1a and H 1b, a chi-squared test was performed to determine the existence
of significant differences regarding consumers' pre-buying and buying decisions in-
fluenced by external information. Additionally, the phi coefficient depict the strength
of association.
Table 25 lists the results, which indicate a strong correlation between belonging to
one of the two groups and the pre-buying decision (2 = 325.933, p < 0.001). Re-
spondents who received a tip from other consumers or experts opted for A or B more
frequently to a significant extent (97.3%). Meanwhile, those from the control group
(without external information) remain predominantly neutral (67.5%). The phi coeffi-
cient (ɸ = 0.713, p < 0.001) has a relatively high value here, which indicates a strong
correlation between the pre-buying decision and group membership. Thus, H 1a is
supported.
176
__________________________________________________________________________
Table 25: Herd Behaviour in the Pre-Buying Phase
Control Group (n=203) Comparison Group (n=438) Total

No external information Information from consumer re- (n=641)
views & experts
Neutral 67.5% 2.7% 23.2%
A or B 32.5% 97.3% 76.8%
Total 100% 100% 100%
Chi- squared test: 2 = 325.933, df = 1, p < 0.001
ɸ = 0.713, p < 0.001
Table 26 indicates a strong correlation between belonging to one of the two groups
and the buying decision (2 = 189.166, p < 0.001). Individuals who received a hint
from Facebook or trusted friends opted for A or B more frequently to a significant ex-
tent (86.5%). Meanwhile, respondents from the control group (without external infor-
mation) remain predominantly neutral (67.5%). The correlation between group mem-
bership and buying decision is not quite as strong as in the pre-buying phase, but still
significant (ɸ = 0.546, p < 0.001). Therefore, H 1b is supported.
Table 26: Herd Behaviour in the Buying Phase
Control Group (n=203) Comparison Group (n=438) Total

No external information Information from Facebook & (n=641)
trusted friends
Neutral 67.5% 13.5% 30.8%
A or B 32.5% 86.5% 69.2%
Total 100% 100% 100%
Chi- squared test: 2 = 189.166, df = 1, p < 0.001
ɸ = 0.546, p < 0.001
177
__________________________________________________________________________
To test H 1c, a Mann–Whitney U test was performed to determine the existence of

significant differences concerning consumers' post-buying decision influenced by ex-
ternal information. Furthermore, eta-squared and Cohen's d depict the strength of
association.
Table 27 lists the results. The Mann–Whitney U test indicates a significant correla-
tion, close to the significance level between belonging to one of the two groups and
the post-buying decision to exchange the chosen product from the buying decision
(Mann–Whitney U test: = 38304, p = 0.044). Yet, the two effect size measures (Eta-
squared (η2) = 0.006 and Cohen's d = 0.157) show only a very small dependence.
The control group has a slightly higher mean value than the comparison group, i.e.
the control group is somewhat more change-friendly. However, the medians that
should be better used for skewed distributions are the same and show that half the
people do not want to change their choices, which again speaks against a difference.
It can be assumed that the difference, which is only just significant, is due to the
number of cases. Hence, H 1c is supported.
Table 27: Herd Behaviour in the Post-Buying Phase
Control Group (n=203) Comparison Group (n=438)

No external information Information from:
Facebook & Trusted Friends
Mean1 2.85 2.56
Standard Deviation 1.76 1.71
25% Quantile 1.00 1.00
Median1 2.00 2.00
75% Quantile 4.00 4.00
1
Mean and median on a 7-point scale, where 1 indicated 'very unlikely' and 7 indicated
'very likely'
Mann–Whitney U test = 38304, p = 0.044
Eta-squared (η2) = 0.006, Cohen´s d = 0.157
178
__________________________________________________________________________
7.6.2 Herd Behaviour by Information Sources
of significant differences regarding the influence of different information sources in
consumers' pre-buying and buying decisions. Additionally, Cramer´s V depicts the
strength of association.
Table 28 lists the results of the first information round in pre-buying decision phase.
The chi-square test comparing treatment group 1 with the control group results in a
test statistic of 244.26 (df = 2, p < 0.001); meanwhile, Cramer´s V equals 0.764 with
p < 0.001. These values denote a large effect size. Comparing treatment group 2
with the control group results in a chi-square test statistic of 255.42 (df = 2, p <
0.001) and Cramer´s V = 0.775 with p < 0.001, which also denotes a large effect
size. The results of both treatment groups indicate a strong correlation between re-
ceiving external information and the pre-buying decision. But almost exactly the
same percentage of participants followed the input of Check24 (consumers' view) as
the input of Stiftung Warentest (experts' view). Thus, H 2a cannot yet be supported.
Table 28: Herd Behaviour in the Pre-Buying Phase with Different information Sources
(1st information round)
Control Group Treatment Group 1 Treatment Group 2 Total

(n=203) (n=216) (n=222)
(n=641)
No external informa- Consumers favour A Experts favour B
tion
A 16.7% 92.6% 1.8% 37.1%
Neutral 67.5% 6.5% 5.4% 25.4%
B 15.8% 0.9% 92.8% 37.4%
Total 100% 100% 100% 100%
Chi-squared test: Chi- squared test:

 = 244.26, df = 2,
2
2 = 255.42, df = 2,
p < 0.001 p < 0.001
Cramer´s V = 0.764, Cramer´s V = 0.775,

p < 0.001 p < 0.001
179
__________________________________________________________________________
The results of the direct comparison of opposing statements (1st and 2nd information
round) between experts and consumers are shown in Table 9. Comparing treatment
group 1 with the control group yields a chi-square test statistic of 152.23 (df = 2, p
<0.001) and Cramer´s V = 0.605 with p < 0.001. The comparison of treatment group
2 and the control group yields a chi-square test statistic of 131.93 (df = 2, p < 0.001)
and Cramer´s V = 0.558 with p < 0.001. Again, both show a significant correlation be-
tween belonging to one of the two groups and the buying decision. However, ap-
proximately 5% more participants decide to follow the input of consumers (75.6%)
than to follow the input of experts (71.0%). Therefore, H 2a is supported.
Table 29: Herd Behaviour in the Pre-Buying Phase with Different information Sources
(2nd information round)

(n=203) (n=216) (n=222)
(n=641)
No external informa- Experts favour B Consumers favour A
tion
A 16.7% 10.4% 75.6% 32.7%
Neutral 67.5% 18.6% 16.4% 33.4%
B 15.8% 71.0% 8.0% 33.9%
Total 100% 100% 100% 100%
Chi- squared test: Chi- squared test:

2 = 152.23 , df = 2, 2 = 131.93, df = 2,
p < 0.001 p < 0.001
Cramer´s V = 0.605, Cramer´s V = 0. 558,

p < 0.001 p < 0.001
Table 30 illustrates a significant correlation between belonging to one of the treat-

ment groups and the buying decision. The comparison of treatment group 1 and the
control group results in a chi-square test statistic of 174.71 (df = 2, p < 0.001) and
Cramer´s V = 0.649 with p < 0.001. Meanwhile, comparison of treatment group 2 with
the control group yields a chi-square test statistic of 272.59 (df = 2, p < 0.001) and
180
__________________________________________________________________________
Cramer´s V = 0.734 with p < 0.001. Thus, more participants (>8%) follow the input of
trusted friends than the input of Facebook. H 2b is therefore supported.
Table 30: Herd Behaviour in the Buying Phase with Different information Sources

(n=203) (n=216) (n=222)
(n=641)
No external informa- Facebook favour B Trusted friends fa-
tion vour B
A 16.7% 1.4% 2.3% 6.6%
Neutral 67.5% 18.4% 8.7% 30.8%
B 15.8% 80.2% 89.0% 62.6%
Total 100% 100% 100% 100%
Chi- squared test: Chi- squared test:

2 = 174.71, df = 2, 2 = 272.59, df = 2,
p < 0.001 p < 0.001
Cramer´s V = 0.649, Cramer´s V = 0.734,

p < 0.001 p < 0.001
To test H 2c, a Mann–Whitney U test was performed to determine the existence of

significant differences concerning the influence of different information sources and
consumers' post-buying decisions. Furthermore, eta-squared and Cohen's d depict
the strength of association.
Table 31 lists the results. The Mann–Whitney U test indicates a significant correla-
tion, close to the significance level between belonging to one of the two treatment
groups and the post-buying decision to exchange the chosen product from the buying
decision (Mann–Whitney U test: 19453, p = 0.059 for treatment group 1 and 18851, p
= 0.109 for treatment group 2). The two effect size measures show only a very small
dependence in both treatment groups: eta-squared (η2) = 0.008 and 0.006, Cohen's
d = 0.180 and 0.155. The control group has a slightly higher mean value than the two
treatment groups, but through comparing the different inputs it seems that friends
seem to have a slightly higher influence than Facebook (Mean = 2.6 versus 2.53),
though on a very low level. However, the medians that should be better used for
181
__________________________________________________________________________
skewed distributions are the same in all comparison groups and show that half the
people do not want to change their choices, which again speaks against a difference.
Thus, H 2c is not supported.
Table 31: Herd Behaviour in the Post-Buying Phase with Different information Sources

(n=203) (n=216) (n=222)
No external information Information from trusted Information from Face-
Friends who favour the book where the oppo-
opposite of the chosen site of the chosen option
option is favoured
Mean1 2.85 2.60 2.53
Standard
1.76 1.84 1.57
Deviation
25%
1.00 1.00 1.00
Quantile
Median1 2.00 2.00 2.00
75%
4.00 4.00 4.00
Quantile
1
Mean and median on a 7-point scale, where 1 indicates 'very unlikely' and 7 indicates
'very likely'
Mann–Whitney U test = Mann–Whitney U test =

19453, p = 0.059 18851, p = 0.109
Eta-squared (η2) = Eta-squared (η2) =

0.008, Cohen´s d = 0.006, Cohen´s d =
0.180 0.155
7.6.3 Herd Behaviour by Product Involvement
The target of the further analysis is to identify, whether the degree of product in-
volvement has an influence on herd behaviour along the buying decision process.
Since three product types were offered in the survey, which differ in their product in-
182
__________________________________________________________________________
volvement (see Figure 34), they were combined for further analysis to obtain a total
product involvement variable across all product types and all individual queries.
Only the comparison groups (treatment groups 1 and 2) were used for the further
analyses (n=438), since only these groups were influenced in their buying decisions
by external information.
The following Figure 36 provides an overview of the product involvement distribution

across all product groups. This elaboration of the product involvement across all
product groups is necessary to be able to analyse the influence of product involve-
ment on herd behaviour in the buying decision process more closely.
Neutral 61.2%
High Involvement 36.1%
Low Involvement 2.7%
n = 438
Figure 36: Product Involvement per Product Type
of significant differences regarding the influence of different level of product involve-
ment in consumers' pre-buying and buying decisions. Additionally, the phi coefficient
depicts the strength of association.
Table 32 lists the results for the pre-buying phase. These results do not indicate a
significant correlation between the pre-buying decision and product involvement level
as the chi-square statistic is 1.89 (df = 1, p = 0.269). Consequently, the decision in
the pre-buying phase does not depend on product involvement in the comparison
group. Furthermore, Cramer´s V (V = 0.066, p = 0.335) has a relatively low value
here, which indicates a low association. H 3a is thus not supported.
183
__________________________________________________________________________
Table 32: Herd Behaviour by Product Involvement in the Pre-Buying Phase
Low Neutral High Involvement Total

Involvement (n=438)
Neutral 8.3% 3.0% 1.9% 2.7%
A or B 91.7% 97.0% 98.1% 97.3%
Total 100% 100% 100% 100%
Chi- squared test: 1.89, df = 1, p = 0.269 (exact test according to Fisher)
Cramer´s V = 0.066, p = 0.335
Table 33 lists the results for the buying phase, indicating a significant correlation be-
tween the buying decision and product involvement. Thus, the higher the product in-
volvement, the higher the probability of deciding for A or B. The chi-square test re-
sults in a test statistic of 19.98 (df = 1, p < 0.001). However, Cramer´s V is 0.215 and
thus relatively low (despite significance), which indicates a low strength of the corre-
lation. (It can be assumed that the significant result is mainly due to the large number
of cases.) Despite the low strength of the association, there is a significant correla-
tion. Thus, H 3b is supported.
Table 33: Herd Behaviour by Product Involvement in the Buying Phase
Low Neutral High Involvement Total

Involvement (n=438)
Neutral 54.5% 14.8% 8.3% 13.5%
A or B 45.5% 85.2% 91.7% 86.5%
Total 100% 100% 100% 100%
Chi- squared test: 19.98, df = 1, p < 0.001
Cramer´s V = 0.215, p < 0.001
Table 34 lists the results of herd behaviour by product involvement in the post-buying
phase. The comparison of the three product involvement groups (Figure 36) with re-
184
__________________________________________________________________________
gard to their decision in the post-buying phase is performed with the Kruskal–Wallis
test. The Kruskal–Wallis test indicates that the three product involvement groups do
not differ significantly from one another in terms of their willingness to switch. H =
3.893, df = 2, p = 0.143. The medians are identical in all groups and the means differ
only slightly. Interestingly, the maximum in the low involvement group is just 5,
whereas the maximum reached the highest possible value 7 in the other groups.
Hence, H 3c is not supported.
Table 34: Herd Behaviour by Product Involvement in the Post-Buying Phase
Low Neutral High Involvement

Involvement (n = 258) (n = 155)
(n=11)
Mean1 2.64 2.45 2.75
Standard
1.63 1.67 1.79
Deviation
Minimum 1 1 1
Median1 2.00 2.00 2.00
Maximum 5 7 7
1
'very likely'
H = 3.893, df = 2, p = 0.143
7.6.4 Herd Behaviour by Social Media Activity
In this section the influence of the social media activity on herd behaviour along the
buying decision process is analysed.
As in the analysis of the product involvement, only the comparison groups (treatment
groups 1 and 2) were used for the analysis (n=438) because only these groups were
influenced in their buying decisions by external information.
Table 35 presents the results, which do not indicate a significant correlation between
the social media activity level and the decision in the pre-buying phase. The chi-
185
__________________________________________________________________________
square test results in a test statistic of 6.929 (df = 3, p = 0.074). Consequently, the
decision in the pre-buying phase does not depend on the social media activity level in
the comparison group. Furthermore, Cramer´s V (V = 0.126, p = 0.074) has a rela-
tively low value here, which confirms the low association.
H 4a is thus not supported. Nevertheless, 'participants' seem to be the most decisive.
Table 35: Herd Behaviour by Social Media Activity in the Pre-Buying Phase
Non-User Lurker Participants Poster Total

(n=438)
Neutral 10.3% 2.5% 1.9% 2.7% 2.7%
A or B 89.7% 97.5% 98.1% 97.3% 97.3%
Total 100% 100% 100% 100% 100%
Chi- squared test = 6.929, df = 3, p = 0.074
Cramer´s V = 0.126, p = 0.074
Table 36 depicts a significant correlation between the buying decision and social me-
dia activity. Therefore, the higher the product involvement, the higher the probability
of deciding for A or B. The chi-square test results in a test statistic of 11.784 (df = 3, p
= 0.008). However, Cramer´s V is 0.165 and thus relatively low (despite significance),
which indicates a low strength of the correlation.
Hence, H 4b is supported. Thereby, participants are the most decisive while posters
are the most reserved and most often remain neutral.
186
__________________________________________________________________________
Table 36: Herd Behaviour by Social Media Activity in the Buying Phase
Non-User Lurker Participants Poster Total

(n=438)
Neutral 25.0% 17.4% 8.0% 19.4% 13.5%
A or B 75.0% 82.6% 92.0% 80.6% 86.5%
Total 100.0% 100.0% 100.0% 100.0% 100.0%
Chi- squared test = 11.784, df = 3, p = 0.008
Cramer´s V = 0.165, p = 0.008
Table 37 assesses herd behaviour by social media activity in the post-buying phase.
The comparison of the social media activity with regard to their decision in the post-
buying phase is performed with the Kruskal–Wallis test. The Kruskal–Wallis test indi-
cates that the social media activity level do not differ significantly from one another in
terms of their willingness to change an already chosen product. Thus, H 4c is not
supported.
Table 37: Herd Behaviour by Social Media Activity in the Post-Buying Phase
Non-User Lurker Participants Poster

(n=28) (n=128) (n=208) (n=36)
Mean1 2.71 2.53 2.52 2.83
Standard
1.88 1.74 1.64 1.92
Deviation
Minimum 1 1 1 1
Median1 2 2 2 2
Maximum 7 7 7 7
1
'very likely';
H = 0.861, df = 3, p = 0.835
187
__________________________________________________________________________
To ascertain whether the level of activity in social media has an influence on how dif-
ferent information sources affect the buying decision process, we turn to Table 38.
For example, the influence of Facebook might differ depending on the social media
activity.
The results presented in Table 38 indicate a significant correlation between the level
of social media activity and the influence on the consumers' decision by a review-
and rating platform (2 = 15.147, p = 0.002). Even though there is a rather weak cor-
relation between activity level and buying decision (V = 0.215, p < 0.001), it can be
clearly seen that 'participants' most often choose A or B and are thus most strongly
influenced in their decision by reviews and ratings. In contrast, 'non-users' most often
remain neutral.
Table 38: Review-/Rating Platform Influence depending on Social Media Activity
Review Non-User Lurker Participants Poster Total

/rating (n=216)
platform
Neutral 30.8% 7.2% 2.8% 7.7% 6.5%
A or B 69.2% 92.8% 97.2% 92.3% 93.5%
Total 100% 100% 100% 100% 100%
Chi- squared test: 15.147, df = 3, p = 0.002
Cramer´s V = 0.265, p = 0.002
Table 39 lists the results for the expert ratings, which show a barely significant (2 =
7.815, p = 0.050) but weak correlation (V = 0.192, p = 0.050) between the level of
social media activity and the influence on the consumers' decision by expert ratings.
'Participants' most often decide for A or B influenced by expert ratings, while 'non-
users' most often stay neutral, closely followed by 'posters'.
188
__________________________________________________________________________
Table 39: Expert Rating influence depending on Social Media Activity
Expert Non-User Lurker Participants Poster Total

rating (n=213)
Neutral 38.5% 15.9% 12.4% 30.8% 16.4%
A or B 61.5% 84.1% 87.6% 69.2% 83.6%
Total 100% 100% 100% 100% 100%
Cramer´s V = 0.192, p = 0.050
Table 40 lists the results for Facebook, highlighting a significant correlation between
the level of social media activity and the influence on the consumers' decision by
Facebook input. The chi-square test results in a test statistic of 10.071 (df = 3, p =
0.018). However, Cramer´s V is with 0.218 is relatively low, which indicates a rather
low strength of the correlation. 'Participants' most often decide for A or B influenced
by Facebook, while 'non-users' most often remain neutral.
Table 40: Facebook influence depending on Social Media Activity
Facebook Non-User Lurker Participants Poster Total

(n=212)
Neutral 46.2% 22.2% 12.4% 15.4% 18.4%
A or B 53.8% 77.8% 87.6% 84.6% 81.6%
Total 100% 100% 100% 100% 100%
Cramer´s V = 0.218, p = 0.018
Overall, for all social media types there seems to be a correlation between social
media activity and buying decisions. 'Participants' are most strongly influenced by so-
cial media in their buying decision and 'non-user' of social media the least. The influ-
189
__________________________________________________________________________
ence of review and ratings seems to be strongest and that of experts the least. Thus,
H 5 is supported.
7.6.5 Herd Behaviour by Personality Type
The following analysis determines whether the personality type influences the deci-
sion in the pre-buying, buying or post-buying phase.
The results listed in Table 41 do not indicate a significant correlation between the
personality type and the decision in the pre-buying phase. The chi-square test results
in a test statistic of 2.801 (df = 4, p = 0.592). Consequently, the decision in the pre-
buying phase does not depend on the personality type in the comparison group. Fur-
thermore, Cramer´s V (V = 0.131, p = 0.592) has a relatively low value here, which
indicates a low association.
Thus, H 6a is not supported. However, the 'emotional stability' type most often re-
mains neutral.
Table 41: Herd Behaviour by Personality Type in the Pre-Buying Phase
Extraver- Agree- Conscien- Emotional Openness Total

sion ableness tiousness Stability to Experi- (n=163)
ences
Neutral 0.0% 0.0% 5.6% 8.3% 3.5% 3.7%
A or B 100.0% 100.0% 94.4% 91.7% 96.5% 96.3%
Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Quadrat = 2.801, df = 4, p = 0.592
Cramer´s V = 0.131, p = 0.592
Table 41 lists the next set of results, which does not indicate a significant correlation
between the personality type and the decision in the pre-buying phase. The chi-
square test results in a test statistic of 3.418 (df = 4, p = 0.490). Consequently, the
decision in the pre-buying phase does not depend on the personality type in the
comparison group. Furthermore, Cramer´s V (V = 0.145, p = 0.490) has a relatively
low value here, which indicates a low association.
Therefore, H 6b is not supported. However, the 'agreeableness' type most often re-
mains neutral.
190
__________________________________________________________________________
Table 42: Herd Behaviour by Personality Type in the Buying Phase
Extraver- Agree- Conscien- Emotional Openness Total

sion ableness tiousness Stability to Experi- (n=162)
ences
Neutral 15.6% 28.6% 9.3% 8.3% 8.8% 11.1%
A or B 84.4% 71.4% 90.7% 91.7% 91.2% 88.9%
Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%
Chi-Quadrat = 3.418, df = 4, p = 0.490
Cramer´s V = 0.145, p = 0.490
Table 43 lists the results of herd behaviour by personality type in the post-buying
phase. The comparison of the personality type with regard to their decision in the
post- buying phase is performed with the Kruskal–Wallis test. The Kruskal–Wallis test
indicates that the social media activity level do not differ significantly from one an-
other in terms of their willingness to change an already chosen product. H = 6.382, df
= 4, p = 0.172
H 6c is thus not supported. Nevertheless, the 'emotional stability' type shows the
lowest willingness to switch, the median is one value below the other types, the
maximum even 4. Nevertheless, the difference is not enough since the sample size is
relatively small.
Table 43: Herd Behaviour by Personality Type in the Post-Buying Phase
Extraversion Agreeable- Conscien- Emotional Openness to

ness tiousness Stability Experiences
Mean1 2.68 2.57 2.69 1.5 2.52
Standard
1.80 2.07 2.07 0.9 1.76
Deviation
Minimum 1 1 1 1 1
191
__________________________________________________________________________
Extraversion Agreeable- Conscien- Emotional Openness to

ness tiousness Stability Experiences
Median1 2 2 2 1 2
Maximum 7 6 6 4 6
1
Mean and median on a 7-point scale, where 1 indicates ‘‘very unlikely’’ and 7 indicates
‘‘very likely’’
H = 6.382, df = 4, p = 0.172
7.6.6 Strongest Influence on Herd Behaviour
Comparing the individual hypothesis test results with each other, i.e. mainly p-value
and effect size measures, it can be seen that hypotheses (H1 to H2) deliver strong
results.
Hence, the pre-buying and buying decision is only influenced by whether the con-
sumer has received any information at all or from which source it came (H2a and
H2b). According to this, in the pre-buying phase information from consumers is fol-
lowed more than that from experts, and in the buying phase friends are followed
more than Facebook.
The further influence of product involvement, social media activity and personality
type can be regarded as rather small, since tests are only partially and then only
weakly significant with low effect strength.
It is most likely to be seen that product involvement shows a certain influence in the
buying phase (p < 0.001, Φ = 0.215).
With regard to social media activity, people who consider themselves to be partici-
pants are probably most likely to be influenced by rating platforms (p < 0.002, V =
0.265) and Facebook (p < 0.018, V = 0.216); the expert opinion has no effect at all.
The personality type seems to have the weakest influence, the tests show no signifi-
cant results here. One reason for this may be that the personality types are not selec-
tive and most people combine a mixed form in their personality, i.e. have more or
less of each type.
192
__________________________________________________________________________
This study aimed to answer two research questions. First, to investigate how herd
behaviour generally differs in the individual consumer buying decision phases (pre-
buying, buying, post-buying). Second, to learn more about various factors influencing
herd behaviour in the individual buying decision phases. The addressed factors were
different information sources (experts, review/rating platforms, friends, FB). Further,
the impact of the level of product involvement (low/high involvement), the degree of
social media activity (poster, lurker, participant), and personality traits (Big Five) were
closely investigated.
To answer these questions, an online survey was conducted with 867 respondents.
Students were selected as the target group, since they actively engage in eWOM
communications about products and services (Bailey 2005), and because they use
social media the most compared to other age groups (Social 2018).
The results of the evaluation with the help of descriptive statistics is summarised and
discussed in the following.
With regards to the information sources in the pre-buying phase, the results show
that expert evaluation proves to be the most important source (50.1%), followed by
review/ rating platforms (47%), and friends (44%), which is largely consistent with ex-
isting studies (Maas et al. 2014; MindTake 2017). Thereby, experts are mainly con-
cerned with credibility, objectivity and trustworthiness. Existing research describes
expert sources as knowledgeable and competent with the competence to provide ac-
curate information, and thus consumers tend to seek advice from experts as they are
credible sources (Pornpitakpan 2004) and since information stemming from experts
is perceived as objective (Tseng et al. 2017).
Review/rating platforms are chosen as a source of information mainly for the follow-
ing reasons: mass opinion, objectivity, and transparency. Moreover, these results
also correspond to previous statements in the literature. Consumers value the 'collec-
tive intelligence' (Surowiecki 2004) which is reflected in the high ranking of 'mass
opinion'. Consumer reviews and ratings help consumers to evaluate offerings through
the lens of previous customers, what is considered objective (Zhou and Duan 2016).
Factors like information quantity, information readiness, detailed information and
193
__________________________________________________________________________
dedicated information are why consumers value anonymous reviews (Erkan and Ev-
ans 2018).
The results show that friends are valued mainly for the reasons of trustworthiness
and credibility, but also the personal relationship plays a role. Existing research ar-
gues that due to personal relationships, users can better interpret recommendations
from friends based on their knowledge, as friends share many experiences which
might lead to develop similar views (Katz and Lazarsfeld 1955). Another aspect that
is outlined in literature is the proximity influence, which describes the role of social
ties and can be associated with trustworthiness (Dewan et al. 2017).
The analysis of the product involvement revealed that entertainment services have
the highest involvement rates among the participants, driven by personal assessment
of the attributes: 'interest in the product' and 'fun to use'. On the other hand, 'mobile
contract' is rated as low-involvement, and 'smartphones' are classified as neutral.
The assessment conducted by Graff et al. (2012, p. 8) goes in the same direction: "a
mobile phone does not fit the high involvement products category for most individu-
als, but still it is not a low-involvement product". Nevertheless, the picture for the
other two products might differ for other target groups than students, since students
hold strong interests in entertainment services (BVDW 2019) and perceive a mobile
contract as a given.
The results of the social media activity evaluation show that the majority of inter-
viewed persons are 'participants' (50.1%), followed by 'lurkers' (34.5%). Only very
few respondents (9%) are high-activity users, called 'posters', and 6.4% stated to be
non-users. Initially something like a 90-9-1 rule was established in the literature. This
rule states that in large online communities 90% of the members can access the in-
formation provided without to contribute something themselves, 9% of the members
occasionally contribute and 1% of the members are responsible for the contents
(Mason 1999; Nielsen 2006). In the meantime, these figures are considered outdated
and various studies show significantly lower numbers of 'lurkers', e.g. Nonnecke and
Preece (2000) present in their study lurker rates of 45.5%. In the present study, 'lurk-
ers' are significantly lower and 'posters' significantly higher than represented in the
original rule. However, this is not surprising since social media activity has increased
over the last years (DataReportal 2020a) and the test participants are students, who
194
__________________________________________________________________________
probably have a comparatively higher activity level anyway. However, the basic rank-
ing of the group still follows the original rule.
The analysis of the personality traits, showed that most participants in the study be-
long to personality types 'openness to experiences' and 'conscientiousness', and the
least belong to 'agreeableness'.
Within the scope of the present study, hypotheses were developed and tested with
statistical methods on the various factors influencing herd behaviour in the individual
buying decision phases. An overview of the test results is shown in Table 44.
195
__________________________________________________________________________
Table 44: Overview on Results of the tested Hypotheses
Related Topic Hypotheses Test results
Herd Behav- H 1a Observation of others in the pre-buying phase is supported

iour in along positively associated with imitating others.
the Buying
H 1b Observation of others in the buying phase is supported
Decision
positively associated with imitating others.
Process
H 1c Observation of others in the post-buying phase is supported
negatively associated with the intention to dis-
continue previous decisions.
Herd Behav- H 2a Consumer ratings exert a greater influence on supported

iour by Infor- herd behaviour than expert ratings in the pre-
mation buying phase.
Sources
H 2b Personal recommendations from trusted friends supported
exert a greater influence on herd behaviour than
impersonal mass opinions on social media in the
buying phase.
H 2c Personal recommendations from trusted friends not sup-

exert a greater influence on herd behaviour than ported
impersonal mass opinions on social media in the
post-buying phase.
Herd Behav- H 3a The degree of product involvement has an influ- not sup-
iour by Prod- ence on herd behaviour in the pre-buying phase. ported
uct Type
H 3b The degree of product involvement has an influ- supported
ence on herd behaviour in the buying phase.
H 3c The degree of product involvement has an influ- not sup-

ence on herd behaviour in the post-buying phase ported
Herd Behav- H 4a The level of social media activity has an influence not sup-
iour by Social on herd behaviour in the pre-buying phase. ported
Media Activity
H 4b The level of social media activity has an influence supported
on herd behaviour in the buying phase.
H 4c The level of social media activity has an influence not sup-
196
__________________________________________________________________________
Related Topic Hypotheses Test results
on herd behaviour in the post-buying phase. ported
H5 The social media activity level affects the influ- supported

ence of social media on buying decisions.
Herd Behav- H 6a Personality has an influence on herd behaviour in not sup-

iour by Per- the pre-buying phase. ported
sonality Type
H 6b Personality has an influence on herd behaviour in not sup-
the buying phase. ported
H 6c Personality has an influence on herd behaviour in not sup-

the post-buying phase. ported
Examining these hypotheses with the survey, it was ascertained that consumers
mostly follow others in the pre-buying phase. This phenomenon decreases in the
buying phase and is no longer evident in the post-purchase phase. This supports
claims that people tend to stick to their status quo rather than follow the herd (Kah-
neman et al. 1991; Samuelson and Zeckhauser 1988).
Comparing recommendations from other consumers versus recommendations from

experts shows that consumers' choices are more effectively influenced by recom-
mendations from other consumers than from experts, which demonstrates the in-
creased importance of UGC for retailers. Further, recommendations from trusted
friends compared with social media input has a stronger influence on herding in the
buying decision process, which shows the dominant effect of proximity influence over
the popularity influence (Dewan et al. 2017). However, in the post-buying phase, only
a marginally stronger and overall very minor influence is present.
Generally, product involvement is related to the consumers' amount of interest in the

product. Considering the influence of product involvement on herd behaviour in the
individual buying decision phases, it becomes evident that this influence is only pre-
sent in the buying phase. In this phase, the higher the product involvement, the
higher the probability of following others. This follows the line of argumentation of
other studies that consumers want to take well-informed decisions for high-
involvement products are therefore rely on the information of others, and that herd
197
__________________________________________________________________________
behaviour is stronger in high-involvement than low-involvement participants (Gu et al.

2012; Luo and Zhang 2013; Yi-Fen and Ya-Ju 2010). This contradicts the statement
of Chen who argues that low-involvement products where consumers spend less
time and cognitive efforts elicit herd behaviour.
With regard to the increase use of social media by consumers, the question was ex-
amined to what extent their own level of activity has an effect on herd behaviour. The
results show that 'participants' are the most decisive group, whereby 'posters' are the
most reserved and most often remain neutral. Since most research (Bartikowski and
Walsh 2014; Schlosser 2005) has only differentiated between 'posters' and 'lurkers'
(missing infrequent posters: participants), the results are not comparable. Different
from the results of Huang et al. (2012), who differentiated between the three groups
and state that 'lurkers' use the opinions of others primarily as a source of information,
infrequent posters rely more on the opinion of others, and posters, as the most active
group, refer to others' opinions only as one source of information and seldom accept
the opinion of others. Nevertheless, a significant correlation between social media
activity and herd behaviour can only be observed in the buying phase.
Considering the influence of the social media type depending on the activity level, the
results confirmed a correlation. Whereby, 'participants' are most strongly influenced
by social media in their buying decision and 'non-user' of social media the least. The
influence of review and ratings seems to be strongest and that of experts the least.
Lurkers do not engage in posting, because they mainly want to satisfy their informa-
tional needs (Preece et al. 2004). In contrast, posters try to satisfy social-emotional
needs and present themselves, improve their image or promote their status and con-
vince others of their opinion (Balasubramanian and Mahajan 2001; Jang et al. 2008).
As each social media type offer different options for self-communication and self-
marketing, it can be assumed that the type of social media then can also affect the
different user groups in the buying decision process differently.
The analysis of the influence of personality types on herd behaviour in the buying de-
cisions show that the personality traits seem not to significantly affect consumers'
behaviour.
In summary, it can be said that there are also some potential research and practical
implications from the identification of differences in herd behaviour between different
198
__________________________________________________________________________
buying phases. The study at hand adds a detailed consideration of herd behaviour in
the individual buying decision phases and supplements existing literature, which refer
primarily on single buying decision phases, like the buying intention (Chen 2008) or
Sun (2013) and Zou et al. (2015) who look deeper in the post-adoptive phase. In ad-
dition, it complements research focusing on information sources. Less can be found,
however, on how different sources of information affect the herd behaviour in the in-
dividual buying phases in direct comparison to each other. However, this much more
reflects the real buying process, where customers get information from different
sources at the same time. In addition, further aspects that have not yet been re-
flected upon in the literature are shown with regard to herd behaviour in the buying
decision process. These are product involvement, social media activity, and person-
ality traits.
With regard to practical implications, mobile communications providers can bear in

mind that consumers follow the opinions of others to different extents, depending on
the buying phase they are in, and how they best use different information sources for
themselves. By understanding the mechanism of herding, they may strategically
manage cues to generate a desirable herding effect among consumers to improve
purchases and create value.
199
Overall Discussion
__________________________________________________________________________
Chapter overview: This chapter draws on the summaries and interim discussions of
the previous chapters (chapters 4 to 7) and provides an overall synthesis of the study
results. Based on the findings of the conducted studies the contribution to existing IS
research on the phenomenon of herd behaviour in buying decisions is outlined in
section 8.1. Furthermore, indications for practice, in particular for the mobile commu-
nications industry, are provided in section 8.2.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 37: Dissertation Overview - Chapter 8: Overall Discussion
200
Overall Discussion
__________________________________________________________________________
8.1 Contribution to IS Research
The overarching objective of the dissertation at hand was to enhance existing knowl-
edge on herd behaviour in the individual buying decision phases in the age of social
media in the context of the mobile communications industry — in particular to shed
light on different aspects influencing this phenomenon.
To approach this objective, four studies were conducted. The first study (chapter 4)
— a comprehensive literature review — forms the foundation for the empirical inves-
tigations and clarifies the state-of-the-art research on herd behaviour in buying deci-
sions in general, without an industry focus. The second study (chapter 5) — the first
empirical one — is based on non-reactive data collected from three different social
media platforms with regard to one mobile communications provider and a subse-
quent qualitative content analysis. The explorative approach had the target to explore
social interactions taking place on social media platforms with regards to mobile
communications products, and to provide novel and intriguing categories on this is-
sue. Expert interviews were used in the third study (chapter 6) to utilise the combined
and overarching knowledge of practitioners from the mobile communications industry
with regards to the herd behaviour, social media, and consumer buying decisions.
Thus, new supporting ideas for further research have been generated. The last and
most comprehensive empirical study of this dissertation (chapter 7) was based on an
online survey with three comparison groups in the context of the mobile communica-
tions industry. Thus, different aspects and their influence on herd behaviour in the
distinct buying decision phases were evaluated using descriptive and analytical sta-
tistics.
To the author's knowledge, this is the first project to specifically address the topic of
herd behaviour in the individual buying decision phases in the context of social me-
dia, and to do so in the empirical context of the mobile communications market. In
the following, the key findings are summarised and it is explained how the results can
contribute to recent IS research. The present research contributes to IS research in
multiple ways, and the results hopefully allow other researchers gain helpful insights.
201
Overall Discussion
__________________________________________________________________________
The dissertation structured the research field, which combined the three research
fields: herd behaviour, buying decisions, and social media, by means of a systematic
literature review in Study I. The analysis revealed that out of the 48 articles only one
article deals with the mobile communications industry, which is why no industry focus
is taken but a general understanding is focused. It was found that the individual top-
ics in the research have already been examined, but only a few contributions deal
with the combination of the respective topics in a targeted approach. Likewise, most
of the publications are only based on individual factors influencing herd behaviour.
These mostly focus on the topic of information type; for example, Li and Wu (2018)
and Jabr and Zheng (2014) studied the impact of different 'information types' on herd-
ing. When considering social media aspects, the emphasis is on the popularity of in-
formation and the quality of information. The individual phases in the buying decision
process of customers are rarely taken into account, or rather most research work re-
fers to the pre-buying phase, while a few also deal with the post-buying phase (e.g.
Sun (2013)). However, no work has been found that combines all three buying deci-
sion phases in one paper. Besides the identification of central research gaps, catego-
ries were developed through a qualitative summary of the material, which give a ho-
listic overview of the combined research area, and thus supplements existing re-
search in these fields.
The qualitative investigation in Study II made three central theoretical contributions:
 First, a qualitative survey of social interactions based on UGC from three so-
cial media platforms in the empirical context of the mobile communications in-
dustry. Comparable qualitative and empirical surveys have not been available
in the literature to date. This was achieved by means of a content analysis de-
rived category system includes all relevant social interactions and thus pro-
vides a clear overview of topics addressed by consumers on different plat-
forms on telecommunication topics. Such a holistic view of relevant parame-
ters has not yet been considered.
 Secondly, the identified social interactions were logically grouped into so-
called main categories. Thereby, eight main categories with regards to gen-
eral social interactions and sixteen underlying subcategories specific to mo-
bile communications topics were revealed. These categories can enrich pre-
202
Overall Discussion
__________________________________________________________________________
vious research on social interaction. The eight identified categories enrich the
social interactions of sharing and expressing and provide a deeper under-
standing of them.
 Third, the link of social interactions to different social media types offers a dif-
ferentiated perspective on how social interactions differs depending on the
social media type.
Study III has sought to explore factors that influence the impact of social media on
bile communications industry. These results supplement existing research in multiple
ways:
 First, experts with a mobile communications background were used as 'crys-

tallisation points' for linked knowledge from different research directions (IS,
marketing, psychology) in the context of business-related decisions. The pre-
sented research combines different perspectives from IS, marketing and psy-
chology, and reveals new perspectives that do not seem to be mentioned in
the existing literature.
 Second, using qualitative content analysis, inductive categories were devel-
oped from the results of the expert interviews. Seven main categories and 29
subcategories were developed through an inductive approach, and the sub-
categories were logically grouped together. Thus, a comprehensive overview
of influencing factors on herd behaviour in the context of social media is pro-
vided, along with a better understanding of the phenomenon, which is a valu-
able supplement to the existing literature. To date, some of the influencing fac-
tors identified have not been described or related to influencing factors of herd
behaviour. The results showed that there is no single relevant factor but rather
multiple factors.
 Third, the identified categories were described and explored by placing them
in the context of existing literature, which contributed to the creation of a
common understanding. Some of the factors (e.g. social media activity, per-
sonality type) identified have also provided important input for the quantitative
investigation in Study IV.
203
Overall Discussion
__________________________________________________________________________
The comprehensive quantitative investigation in Study IV essentially contributes to

IS research in the following way:
 Firstly, the herd behaviour in the three individual buying decision phases was
examined in the context of the mobile communications industry. This seems
not to have been investigated in the literature before and thus provides a valu-
able contribution to the IS research community.
 Furthermore, various factors influencing herd behaviour in the respective buy-
ing decision phases were examined and the strength of their impact was in-
vestigated.
 The combination of influencing factors (information source, product involve-
ment, social media activity, personality type) have mostly only been examined
separately by researchers; investigating them here in a holistic study provided
the opportunity to compare the influencing factors in their effect on herd be-
haviour in the individual buying decision phases.
8.2 Contribution from a Management Perspective
The results of the studies conducted have several important practical implications,
and thus can provide managers with the guidance on how they should deal with herd
behaviour among consumers, and how to benefit from social media in this context, all
while considering other relevant factors.
First, the results show that social interaction is platform specific for the investigated
mobile communications provider. For example, on the social networking platform, so-
cial interactions 'support', 'comment and opinion', 'regret' and 'dissatisfaction' take
precedence. Meanwhile, the focus on the forum is more on 'recommendation', 'ex-
perience', and 'reference'. On the review/rating platform, the interactions are limited
to 'satisfaction' and 'dissatisfaction'. In the development of a company's social media
strategy, it is important to consider which social media should be actively controlled
by the company, and which platform content should be analysed in order to gain in-
sights into critical issues (e.g. satisfaction, complaints, etc.). Especially in the mobile
communications industry, which in the past was often characterised by negative cus-
tomer experiences (Connect 2019), which is also reflected in the present study re-
sults (e.g. 'dissatisfaction' interaction was most prominent on the review/rating plat-
form), active control, analysis and understanding of social interactions of customers
is required. To contribute to customer satisfaction and thus be successful, firms need
204
Overall Discussion
__________________________________________________________________________
to carefully handle these kinds of interactions. It is necessary to build up the corre-

sponding analytical and strategic competencies in the company in order to develop a
target-oriented social media strategy, considering social interaction aspects. The as-
pects described above can also provide valuable input for other mobile operators.
The awareness of herd behaviour in the buying decision process also seems to have
arrived in the economy. It is not without reason that there are programs such as
'friends recruit friends' or targeted information about which products are primarily
bought by customers and how these are evaluated, as is the case with Amazon.com.
"A number of services have emerged over the last few years that attempt to integrate
with online social network services (SNS) so that users can access reviews submit-
ted by their friends in addition to those submitted by the rest of the online community"
(Lee et al. 2015, p. 2241). For instance, TripAdvisor, Yelp, and some other popular
review platforms have added features to allow users to identify reviews by friends. In
addition, a better understanding of what other factors influence the herd behaviour of
consumers is relevant to be able to optimise online marketing and sales activities by
analysing the relevant factors related to their consumers and their buying behaviour.
In order to be able to assess the effect and interaction of the individual factors in
practice, it is necessary to have the corresponding data (e.g. customer data, transac-
tion data, behavioural data), and to use the corresponding analysis and tracking
tools.
Further practical implications arise from the fourth study. Firstly, information is pro-
vided on where customers can obtain information before buying mobile communica-
tions products. The most popular sources are expert evaluations, review rating plat-
forms, and friends. These insights help mobile providers to adapt their marketing and
sales communication to the respective information sources. Furthermore, the reasons
for choosing certain channels guides marketers in which wording and keywords to
use on the channel in order to meet customers' expectations that are specific to that
channel (e.g. credibility, objectivity).
With regard to product involvement, practitioners from the mobile communications in-
dustry get to know that entertainment services have the highest involvement rates
among the participants, 'mobile contract' is rated as low-involvement, and 'smart-
205
Overall Discussion
__________________________________________________________________________
phones' are classified as neutral. From this, practitioners can deduce which focus
they should put in marketing.
Understanding the existence of different types of activity in the social media world is
of great relevance to industry. It is important to understand that the group of passive
consumers is significantly higher than the group of active content creators. Therefore,
as a company, you have to think about how to reach and activate the passive group.
Furthermore, it is particularly relevant for the mobile communications industry, whose

customer relationship is primarily based on contractual relationships, to take the indi-
vidual buying decision phases into account. The knowledge about herd behaviour in
the individual phases can help to actively control them.
Further, this paper also depicts that consumer ratings exert a greater influence on
herd behaviour than expert ratings in the pre-buying phase, and personal recom-
mendation from trusted friends exert a greater influence on herd behaviour than im-
personal mass opinions (Facebook) in the buying and post-buying phase. in the post-
buying phase, external information seems to be of secondary importance or not influ-
ence customer decisions as much as in the previous phases. With this knowledge,
practitioners can align the corresponding channels with the appropriate information to
the individual buying decision phases.
In terms of social media activity, it becomes clear that firms can influence 'partici-
pants' in particular in their buying decisions, although they represent the largest
group, but the other two groups (lurker, poster) must not be neglected. Firms have to
be aware that their decisions may be influenced differently than 'participants', and this
should be reflected in their marketing strategy.
Even if personality types and customer segments are an appropriate tool, especially
in marketing, particularly in terms of communication, willingness to pay, product pref-
erences, customer value etc. (Hwang et al. 2004). They seem to play a subordinate
role as an influencing factor on herd behaviour in the decision-making process, and
therefore there is no need to focus on this in practice.
206
Conclusion
__________________________________________________________________________
9 Conclusion
Chapter overview: This concluding chapter provides an overall summary of the re-
search project of this dissertation. Therefore, the results of the dissertation that have
contributed to answering the research questions are summarised in section 9.1.
Then, the limitations of the studies conducted are outlined in section 9.2. Lastly, an
outlook and suggestions for future research opportunities are provided in section 9.3.
PART A: Foundations
1 Introduction
3 Research Design

6 Media
PART C: Summary
9 Conclusion
Figure 38: Dissertation Overview - Chapter 9: Conclusion
207
Conclusion
__________________________________________________________________________
9.1 Research Questions addressed
The research gaps addressed in the present dissertation were closed with chapters 4
to 7. The objective of this research work was determined by seven research ques-
tions. In the following, the results for the individual research questions are summa-
rised concisely.
Before entering the empirical field of the conducted empirical studies in the mobile
communications industry, an analysis of existing scientific research was carried out to
take inventory of the existing knowledge about herd behaviour in consumers' buying
decisions in the context of social media. Therefore, the literature review (chapter 4)
targeted answering the following research questions:


The statistical evaluation of the 48 articles identified in the timeframe from 2008 to
2018 suggests an increasing interest in the studied phenomenon, particularly from
2014 onwards, while the most frequently research streams are IS, economics, and
marketing. Multiple methods (e.g. survey, experiments, and mathematical modelling)
and variety of theories (e.g. social influence theory, information cascade theory, herd-
ing theory) are applied in literature to study the phenomenon.
The qualitative evaluation of the literature set identified revealed three main topics
with regards to the first RQ 1a: 'information type', 'information source', 'technology
adoption'. Thereby, the treatment of the topic 'information type' in literature has more
than doubled in the period between 2014 and 2018. The results with regards to the
second research question (RQ 1b), summarises a total of six topic clusters: 'popular-
ity', 'valence', 'information quality', 'product type', 'homophily', and 'others', with being
'popularity' and 'information quality' the most frequently discussed topics, especially
in the second time period.
Overall, the results of the literature review show general important aspects with re-
gards to herd behaviour in the purchasing decision and the role of social media
208
Conclusion
__________________________________________________________________________
therein. But it can be concluded that single aspects concerning herd behaviour in the
buying decision process have been researched. However, there is a missing com-
prehensive view that brings together the different aspects along the entire buying de-
cision process.
Second, by means of a non-reactive study collected data from three social media
platforms to explore social interactions with regard to mobile communications topics
related. The conducted study (chapter 5) aimed to answer the following research
questions:
RQ 2a How do social interactions differ depending on different social media plat

forms?

This study has sought to explore how social interactions differ depending on different
social media platforms in the empirical context of the mobile communications indus-
try. The main contribution of this study stemmed from the qualitative content analysis
based on UGC from three social media platforms. Thereby, eight main categories
with regards to general social interactions and sixteen underlying subcategories spe-
cific to mobile communications topics were revealed. The eight identified categories
of 'satisfaction', 'dissatisfaction', 'comment and opinion', 'reference', 'support', 'rec-
ommendation', 'regret', and 'experience' enrich the social interactions of sharing and
expressing and provide a deeper understanding of them. Furthermore, the link of so-
cial interactions to different social media types offers a differentiated perspective on
how social interactions differs depending on the social media type. The results also
indicate that some interactions to mobile communications topics seem to be specific
to certain platforms, which might give firms the opportunity to influence social interac-
tions. The presented research reveals novel categories and new perspectives on so-
cial interactions differing across various social media platforms.
Third, to identify different factors affecting herd behaviour in the context of social me-
dia, expert interviews were conducted (chapter 6) with the target to answer the fol-
lowing research question:
209
Conclusion
__________________________________________________________________________
ing decisions?
This study has sought to explore factors that influence the impact of social media on
bile communications industry. The research shows that there is no single relevant
factor but rather multiple factors. The qualitative content analysis revealed seven
main categories: 'drivers for social media', 'general influence groups', 'product specif-
ics', 'role of social media in buying decision process', 'customer segment specifics',
'personal vs. mass opinion', 'social media activity', and 29 subcategories developed
through an inductive approach. The presented research combines different perspec-
tives from IS, marketing and psychology and reveals new perspectives that do not
seem to be mentioned in the extant literature and thus facilitate future academic re-
search.
Fourth, an online survey was conducted to shed reveal the impact of different factors
on herd behaviour in the individual buying decision phases with regard to mobile
communication products. Thus, the objective of the quantitative study presented in
chapter 7 is to answer the following research questions:
phases?

This study aimed to investigate how herd behaviour generally differs in the individual
consumer buying decision phases (pre-buying, buying, post-buying), and to find out
more about various factors influencing herd behaviour in the individual buying deci-
sion phases. The addressed factors were different information sources (experts, re-
view-/rating platforms, friends, FB). Additionally, the impact of the level of product in-
volvement (low/high involvement), the degree of social media activity (poster, lurker,
participant), and personality traits (Big Five) were more closely investigated.
210
Conclusion
__________________________________________________________________________
The descriptive results revealed in the pre-buying phase expert evaluations, re-
view/rating platforms, and friends are the most important sources, mainly due to
credibility, objectivity and trustworthiness, mass opinion, transparency. The analysis
of the product involvement revealed that entertainment services have the highest in-
volvement rates among the participants. The results of the social media activity
evaluation show that the majority of interviewed persons are 'participants', followed
by 'lurkers'.
The statistical analysis led to following findings: the observation of others in the pre-
buying and the buying phase is positively associated with imitating others, and the
observation of others in the post-buying phase is negatively associated with the in-
tention to discontinue previous decisions. That is, the more advanced the purchasing
process, the less pronounced the herd behaviour. Further, consumer ratings exert a
greater influence on herd behaviour than expert ratings in the pre-buying phase, and
personal recommendation from trusted friends exert a greater influence on herd be-
haviour than impersonal mass opinions on social media in the buying and post-
buying phase. Product Involvement shows a certain influence in the buying phase,
and with regards to social media activity, people who consider themselves to be par-
ticipants are probably most likely to be influenced by rating platforms. Finally, the
personality type seems to have the least influence on herd behaviour in buying deci-
sions.
9.2 Limitations
The methods used in this research work are subject to some limitations, which are
listed below for the conducted studies in this work.
Study I
The review presented of existing scholarly literature on herd behaviour in buying de-
cisions in study I has limitations. Although a sufficient number of literature sources
have been examined to gain a deeper understanding of the current state of research
on herd behaviour in the buying decision process in the context of social media, it
cannot be guaranteed that relevant articles have been overlooked, despite the care-
ful selection. This is mainly due to the fact that the literature search combined three
fields of research and that there are also some synonyms (e.g. social media, social
network, online community, etc.) for these topics in different fields of research (eco-
nomics, psychology, IS, etc.). Although an attempt has been made to take the vari-
211
Conclusion
__________________________________________________________________________
ous synonyms into account, it cannot be ruled out that individual formulations have
not been used and thus relevant articles have not been included in the analysis.
Study II
With regard to the qualitative Study II, there are some methodological limitations as-
sociated with the data collection procedure used, and hence the results are not fully
valid. In particular, the methodological limitations of the study must be addressed.
The data collection was carried out by means of non-reactive data collection from
three different social media platforms to explore social interactions with regards to
mobile communications topics. The collected data from three weeks can be consid-
ered as snapshot. This non-random sample also imposes methodological limitations.
In particular, the chosen form of data collection cannot achieve a high external valid-
ity in the sense of replicability or generalisation of the results. This means that data
that would be collected at a different time, or over a longer period, or with regard to
another mobile provider could lead to different results. Nevertheless, to avoid differ-
ences with regards to social interactions based on different mobile providers or time
differences between platforms, the elements like related organisation (o2) and the
time period of the retrieved data were kept constant. In the context of limitations, it
should also be mentioned that only three platforms have been investigated, each rep-
resentative of one social media type. With regard to the research question, this was
considered acceptable, but it cannot be denied that content from other social media
platforms could have led to different results.
Non-reactive data allow rich insights into individuals’ perspectives, interpretations,

and constructions of meaning, and thus is also less affected by issues relating to self-
reported biases like social desirability biases (Hewson 2017). This typical advantage
of a qualitative research design also brings some limitations, as the collected data re-
treat from objective comparison (Creswell 2009).
In order to overcome these limitations, the data collection tool and the data evalua-
tion instrument are adequately described. First, an observational protocol supported
the recording procedure and a content-analytical process model was used. More-
over, a coding guideline with central definitions and anchor examples was used, and
the involvement of a second scientists contributed to an objectification of the evalua-
tion process.
212
Conclusion
__________________________________________________________________________
Study III
Considering Study III, there are a few methodological limitations inherent in the data
collection procedure, and therefore the results are not fully valid. Against this back-
ground, in particular the methodological limitations of the study should be disclosed.
The data collection was carried out by means of semi-structured expert interviews to
explore various factors influencing herd behaviour in the context of social media. A
total of six experts from a mobile provider, an online marketing company, and free-
lance social media experts were interviewed. Since predominantly senior managers
were interviewed, they may have made some inaccurate or biased statements. Man-
agers have a recognised tendency to present an image to the outside world that is in
line with socially and organisationally desirable standards, particularly if they believe
that the disclosure of information could have an impact on their career or if they are
highly committed to their organisations (Huber and Power 1985). Since the expert in-
terviews were anonymised, it can be assumed that an expected impact on the career
of the interviewees is not a reason for them to adjust their statements accordingly,
but this cannot be ruled out entirely.
Qualitative research designs stand out in particular for the openness of the investiga-
tion; however, this central feature is also a central limitation (Gläser and Laudel 2010;
Kuckartz 2014). Although a major advantage of qualitative data is that a detailed de-
scription of the object of investigation is provided, the data collected often elude ob-
jective comparison (Bortz and Döring 2007; Flick et al. 2012).
The researcher’s own social network account, external recommendations, and online
research, particularly using LinkedIn and Xing, were employed to identify potential
candidates. This non-random sample presents methodological constraints. In particu-
lar, with the selected form of data collection, no high external validity in terms of rep-
licability or generalisation of the results can be achieved, which means that a further
sample may lead to different results. Therefore, the selection of the sample was
made with due consideration that the experts come from different areas of expertise
and partly from different firms; the motivation behind this was to obtain a more holistic
view. Furthermore, all interviews were conducted by the same interviewer, and there-
fore an interviewer bias cannot be fully excluded.
To ensure the highest possible reliability, the data collection tool and the data evalua-
tion instrument are adequately described. This includes the documentation of the in-
213
Conclusion
__________________________________________________________________________
terview guidelines, the description of the situation in which the interviews were con-
ducted and the content-analytical process model. Furthermore, a coding guideline
with central definitions and anchor examples created. Especially the coding of the
material by a second scientists contributed to an objectification of the process of
knowledge.
Study IV
In Study IV, the established hypothesis system was examined in a quantitative study.
In the following, the methodological limitations of the data collection and data evalua-
tion procedures will be disclosed.
The data collection was carried out by means of a standardised online survey primar-
ily of students. A total of 867 people took part in the survey, and after data cleansing,
641 were taken into consideration for further analysis. The use of a student sample,
may result in a limitation in generalisability, as other target groups might have a dif-
ferent decision-making behaviour and might also have a different attitude towards
social media. Furthermore, the research design represents limitations, considering
the restrictions imposed on the subjects by the hypothetical buying decision process.
Especially in the post-buying phase, the test persons are asked to assume that they
have decided on a product and are then asked to reassess the situation. Neverthe-
less, the advantages were considered to be predominant with regard to the research
questions examined, and the experimental element combine the advantages of sur-
veys such as, generalisability and external validity with advantages of experiments
such as valid causal inference and internal validity (Nock and Guterbock 2010).
Quantitative methods are distinguished by a clearly structured procedure, and clearly

defined and described analysis methods. The collection and evaluation of data does
not require interpretation by the researcher. The results obtained are therefore inter-
subjectively comprehensible, and the investigations carried out can be replicated.
The quality of the research results and the quality of the conclusions drawn can
therefore be assessed by outsiders and checked in comparative studies (Schwaiger
and Meyer 2011).
The use of suitable scales is of paramount importance in the research process. Their
quality has a decisive influence on the quality of the data collected and the research
results achieved. The criteria of objectivity, reliability and validity are usually used to
assess the quality of a scale (Herrmann and Homburg 2000). Only where these crite-
214
Conclusion
__________________________________________________________________________
ria are met can quantitative research methods exploit their advantages such as com-
parability, validity and replicability of the results. Therefore, wherever possible, al-
ready validated scales are used. Further, to ensure a rigorous data collection proc-
ess, detailed information about the survey instrument used in the study is provided
and the complete questionnaire provided.
Moreover, it should be mentioned that the data evaluation procedures are focused on
statistical tests, and additional statistical analysis methods could provide further in-
sight into the results from other perspectives.
All empirical studies (Study II – Study IV) were conducted in the German market in
the context of the mobile communications industry. Thus, it cannot be results in other
countries, due to cultural differences and hence different consumer behaviour, could
produce a different picture. In addition, different market conditions in mobile commu-
nications may also result in different consumer attitudes and behaviour.
9.3 Further Research Opportunities
The results and limitations of the studies provide a number of indications for further
research. For example, the samples of the empirical studies were limited to Ger-
many. In order to rule out country-induced bias, various countries could be included
in the sample of a future survey.
Another interesting perspective for future research would be to examine herd behav-
iour in the buying decision process in other industry contexts. This would be interest-
ing in a twofold context. First, one could look at other industries that also offer com-
modity products (e.g. insurance, energy) or other industries that are more emotional,
such as the fashion industry.
Future research could also enrich the current results by utilising other methodology
and other data. In terms of exploring social interactions, automated data collection of
non-reactive data would provide a much larger amount of data as a basis for analy-
sis. This would allow more social media platforms to be considered and compared
with each other. Using text analysis tools, the large amounts of generated data could
then be analysed more automatically. Furthermore, the findings could be enriched by
further research regarding the influence of identified interactions and their impact on
customer's behaviour.
215
Conclusion
__________________________________________________________________________
With regards to the research of influencing factors on herd behaviour in buying deci-
sions further research ideas are proposed below.
First of all, other influencing factors can be identified and their effect on herd behav-
iour analysed. Further interesting aspects in the context of herd behaviour in buying
decisions could also be investigated by predictive analytics through Big Data in social
media.
Secondly, further samples can be considered in order to either achieve a representa-

tive cross-section of the entire population. Or, different target groups (e.g. pensioners
vs. students or academics vs. non-academics) can be compared in terms of herd be-
haviour in the buying decision process.
Additionally, research could be expanded by conducting laboratory experiments un-

der controlled conditions. In this case, the test persons could be confronted with
more tangible decision scenarios, for example by programming websites that suggest
a real social media world or by including products with real product features in the
experiment.
Furthermore, with the help of a longitudinal study, the purchasing process could be
depicted differently and, above all, the post-buying phase could be better reflected. In
this case, customers would be confronted with an actual product decision and would
be allowed to use real products. After a certain period of time, a new survey would be
conducted, taking into account influencing factors. In this case, the buying decision
process would not be purely hypothetical.
Since the research field around social media is an extremely dynamic one, the es-
sential literature and thus the theory and empirical research always lags behind the
times, an update of the theory could make another valuable contribution to research.
216
217
References
__________________________________________________________________________
References
Abdalla Mikhaeil, C., and George, A. 2016. "Group Decision Making in Oss: A
Dialectic Perspective on Herding".
Abrahamson, E. 1991. "Managerial Fads and Fashions: The Diffusion and Rejection
of Innovations," Academy of Management Review (16:3), pp. 586-612.
Ackert, L. F., Church, B. K., and Ely, K. 2008. "Biases in Individual Forecasts:
Experimental Evidence," Journal of Behavioral Finance (9:2), pp. 53-61.
Aggarwal, C. C., and Zhai, C. X. 2012. Mining Text Data. New York: Springer-Verlag.
Ajzen, I., and Fishbein, M. 1973. "Attitudinal and Normative Variables as Predictors
of Specific Behavior," Journal of personality and Social Psychology (27:1), p.
41.
Ajzen, I., and Kuhl, J. 1985. "Action-Control: From Cognition to Behavior," From
intentions to actions: A theory of planned behavior), pp. 11-39.
Akehurst, G. 2008. "User Generated Content: The Use of Blogs for Tourism
Organisations and Tourism Consumers," Service Business (3:1), p. 51.
Ali, S., and Kartik, N. 2012. "Herding with Collective Preferences," Economic Theory
(51:3), pp. 601-626.
Allport, F. H. 1924. "Response to Social Stimulation in the Group," Social psychology
(11), p. 260.
Allport, G. W., and Odbert, H. S. 1936. "Trait-Names: A Psycho-Lexical Study,"
Psychological monographs (47:1), p. i.
Alves, H., Fernandes, C., and Raposo, M. 2016. "Social Media Marketing: A
Literature Review and Implications," Psychology & Marketing (33:12), pp.
1029-1038.
Amy Wenxuan, D., and Li, S. 2018. "Herding in the Consumption and Purchase of
Digital Goods and Moderators of the Herding Bias," Journal of the Academy of
Marketing Science), pp. 1-19.
Anderson, C. 2006. The Long Tail: Why the Future of Business Is Selling Less of
More. Hachette Books.
Applbaum, R. F., and Anatol, K. W. E. 1972. "The Factor Structure of Source
Credibility as a Function of the Speaking Situation," Speech Monographs
(39:3), pp. 216-222.
218
References
__________________________________________________________________________
Ariely, D. 2008. Predictably Irrational : The Hidden Forces That Shape Our
Decisions. New York, NY: Harper.
Armstrong, G., Kotler, P., Harker, M., and Brennan, R. 2009. Marketing an
Introduction, (9 ed.). Harlow: Pearson Education Limited.
Arndt, J. 1967. "Role of Product-Related Conversations in the Diffusion of a New
Product," Journal of Marketing Research (4:3), pp. 291-295.
Asch, S. E. 1956. "Studies of Independence and Conformity: I. A Minority of One
against a Unanimous Majority," Psychological monographs: General and
applied (70:9), p. 1.
Asendorpf, J. B., and Neyer, F. J. 2012. Psychologie Der Persönlichkeit. Springer-
Verlag.
Atkinson, R. C., and Shiffrin, R. M. 1968. "Human Memory: A Proposed System and
Its Control Processes".
Ayanso, A., Lertwachara, K., and Vachon, F. 2007. "Diversity or Identity Crisis? An
Examination of Leading Is Journals," Communications of the Association for
Information Systems (20:42).
Babić Rosario, A., Sotgiu, F., De Valck, K., and Bijmolt, T. H. 2016. "The Effect of
Electronic Word of Mouth on Sales: A Meta-Analytic Review of Platform,
Product, and Metric Factors," Journal of Marketing Research (53:3), pp. 297-
318.
Bailey, A. A. 2005. "Consumer Awareness and Use of Product Review Websites,"
Journal of Interactive Advertising (6:1), pp. 68-81.
Baker, R. K., and White, K. M. 2010. "Predicting Adolescents’ Use of Social
Networking Sites from an Extended Theory of Planned Behaviour
Perspective," Computers in Human Behavior (26:6), pp. 1591-1597.
Balasubramanian, S., and Mahajan, V. 2001. "The Economic Leverage of the Virtual
Community," International journal of electronic commerce (5:3), pp. 103-138.
Baltes, P. B., and Smelser, N. J. 2004. International Encyclopedia of the Social &
Behavioral Sciences. Elsevier.
Bandura, A. 1969. "Social-Learning Theory of Identificatory Processes," Handbook of
socialization theory and research (213), p. 262.
Bandura, A. 1986. Social Foundations of Thought and Action: A Social Cognitive
Theory. Englewood Cliffs, NJ, US: Prentice-Hall, Inc.
219
References
__________________________________________________________________________
Banerjee, A. V. 1992. "A Simple Model of Herd Behavior," The Quarterly Journal of
Economics (107:3), pp. 797-817.
Bankhofer, H. v. U., Nissen, V., Stelzer, D., Straßburger, S., Herzwurm, G., and
Stelzer, D. 2008. "Wirtschaftsinformatik Versus Information Systems-Eine
Gegenüberstellung,").
Bänsch, A. 2017. Käuferverhalten. Walter de Gruyter GmbH & Co KG.
Barrick, M. R., and Mount, M. K. 1991. "The Big Five Personality Dimensions and
Job Performance: A Meta‐Analysis," Personnel psychology (44:1), pp. 1-26.
Bartikowski, B., and Walsh, G. 2014. "Attitude Contagion in Consumer Opinion
Platforms: Posters and Lurkers," Electronic Markets (24:3), pp. 207-217.
Bashir, T., Azam, N., Butt, A. A., Javed, A., and Tanvir, A. 2013. "Are Behavioral
Biases Influenced by Demographic Characteristics & Personality Traits?
Evidence from Pakistan," European Scientific Journal (9:29).
Batinic, B., Reips, U.-D., and Bosnjak, M. 2002. Online Social Sciences. Hogrefe &
Huber Seattle, WA.
Baumer, E. P., Sueyoshi, M., and Tomlinson, B. 2011. "Bloggers and Readers
Blogging Together: Collaborative Co-Creation of Political Blogs," Computer
Supported Cooperative Work (CSCW) (20:1-2), pp. 1-36.
Bayer, J. B., Triệu, P., and Ellison, N. B. 2020. "Social Media Elements, Ecologies,
and Effects," Annual review of psychology (71), pp. 471-497.
Belch, G. E., and Belch, M. A. 2003. Advertising and Promotion: An Integrated
Marketing Communications Perspective. The McGraw− Hill.
Bettman, J. R. 1979. Information Processing Theory of Consumer Choice. Addison-
Wesley Pub. Co.
Bettman, J. R., Luce, M. F., and Payne, J. W. 1998. "Constructive Consumer Choice
Processes," Journal of consumer research (25:3), pp. 187-217.
Bharati, P., Zhang, C., and Chaudhury, A. 2014. "Social Media Assimilation in Firms:
Investigating the Roles of Absorptive Capacity and Institutional Pressures,"
Information Systems Frontiers (16:2), pp. 257-272.
Bickart, B., and Schindler, R. M. 2001. "Internet Forums as Influential Sources of
Consumer Information," Journal of Interactive Marketing (15:3), pp. 31-40.
Bikhchandani, S., Hirshleifer, D., and Welch, I. 1992. "A Theory of Fads, Fashion,
Custom, and Cultural Change as Informational Cascades," Journal of political
Economy (100:5), pp. 992-1026.
220
References
__________________________________________________________________________
Bikhchandani, S., Hirshleifer, D., and Welch, I. 1998. "Learning from the Behavior of
Others: Conformity, Fads, and Informational Cascades," Journal of Economic
Perspectives (12:3), pp. 151-170.
Bikhchandani, S., and Sharma, S. 2000. "Herd Behavior in Financial Markets," IMF
Staff papers (47:3), pp. 279-310.
Bitkom. 2018. "Online Gleich Nach Dem Aufwachen Und Kurz Vor Dem Einschlafen:
Jeder Zweite Nutzt Social Media Im Bett." Retrieved 26.3.2020, from
https://www.bitkom.org/Presse/Presseinformation/Online-gleich-nach-dem-
Aufwachen-und-kurz-vor-dem-Einschlafen-Jeder-Zweite-nutzt-Social-Media-
im-Bett.html
Bogner, A., Littig, B., and Menz, W. 2009. "Introduction: Expert Interviews — an
Introduction to a New Methodological Debate," in Interviewing Experts, A.
Bogner, B. Littig and W. Menz (eds.). London: Palgrave Macmillan UK, pp. 1-
13.
Bogner, A., Littig, B., and Menz, W. 2014. Interviews Mit Experten. Wiesbaden:
Springer VS.
Bogner, A., and Menz, W. 2009. "The Theory-Generating Expert Interview:
Epistemological Interest, Forms of Knowledge, Interaction," in Interviewing
Experts, A. Bogner, B. Littig and W. Menz (eds.). London: Palgrave Macmillan
UK, pp. 43-80.
Bortz, J., and Döring, N. 2007. Forschungsmethoden Und Evaluation Für Human-
Und Sozialwissenschaftler: Limitierte Sonderausgabe. Springer-Verlag.
Bosman, D. J., Boshoff, C., and Van Rooyen, G.-J. 2013. "The Review Credibility of
Electronic Word-of-Mouth Communication on E-Commerce Platforms,"
Management Dynamics: Journal of the Southern African Institute for
Management Scientists (22:3), pp. 29-44.
Brooks Jr, F. P. 1995. The Mythical Man-Month: Essays on Software Engineering,
Anniversary Edition, 2/E. Pearson Education India.
Brucks, M. 1985. "The Effects of Product Class Knowledge on Information Search
Behavior," Journal of consumer research (12:1), pp. 1-16.
Bundesnetzagentur. 2019a. "Anzahl Der Mobilfunkanschlüsse Pro 100 Einwohner in
Deutschland Von 1990 Bis 2018." Retrieved 26.3.2020, from
https://de.statista.com/statistik/daten/studie/13264/umfrage/penetrationsrate-
der-deutschen-mobilfunknetze-seit-1990/
221
References
__________________________________________________________________________
Bundesnetzagentur. 2019b. "Tätigkeitsbericht Telekommunikation 2018/2019."

Retrieved 7.32020, from https://www.bundesnetzagentur.de/SharedDocs
/Mediathek/Taetigkeitsberichte/2019/TK_20182019.pdf?__blob=publicationFil
e&v=9
BVDW. 2019. "Welche Der Folgenden Streaming-Dienste Nutzen Sie? ." Retrieved
10.2.2020, from https://de.statista.com/statistik/daten/studie/1074604/umfrage/
nutzung-von-streaming-diensten-nach-altersgruppen-in-deutschland/
Cai, H., Chen, Y., and Fang, H. 2009. "Observational Learning: Evidence from a
Randomized Natural Field Experiment," American Economic Review (99:3),
pp. 864-882.
Casaló, L. V., Flavián, C., and Guinalíu, M. 2011. "Understanding the Intention to
Follow the Advice Obtained in an Online Travel Community," Computers in
Human Behavior (27:2), pp. 622-633.
Chaiken, S., and Trope, Y. 1999. Dual-Process Theories in Social Psychology.
Guilford Press.
Chatterjee, P. 2001. "Online Reviews: Do Consumers Use Them?," Advances in
Consumer Research (28:1), pp. 129-133.
Chen, G. M. 2011. "Tweet This: A Uses and Gratifications Perspective on How Active
Twitter Use Gratifies a Need to Connect with Others," Computers in Human
Behavior (27:2), pp. 755-762.
Chen, Y.-F. 2008. "Herd Behavior in Purchasing Books Online," Computers in
Chen, Y., Wang, Q., and Xie, J. 2011. "Online Social Interactions: A Natural
Experiment on Word of Mouth Versus Observational Learning," Journal of
Marketing Research (48:2), pp. 238-254.
Cheng, Y.-H., and Ho, H.-Y. 2015. "Social Influence's Impact on Reader Perceptions
of Online Reviews," Journal of Business Research (68:4), pp. 883-887.
Cheung, C. M. K., Xiao, B., and Liu, I. L. B. 2012. "The Impact of Observational
Learning and Electronic Word of Mouth on Consumer Purchase Decisions:
The Moderating Role of Consumer Expertise and Consumer Involvement,"
2012 45th Hawaii International Conference on System Sciences, pp. 3228-
3237.
Cheung, M. Y., Luo, C., Sia, C. L., and Chen, H. 2009. "Credibility of Electronic
Word-of-Mouth: Informational and Normative Determinants of on-Line
222
References
__________________________________________________________________________
Consumer Recommendations," International journal of electronic commerce

(13:4), pp. 9-38.
Chiou, J.-S., and Cheng, C. 2003. "Should a Company Have Message Boards on Its
Web Sites?," Journal of Interactive Marketing (17:3), pp. 50-61.
Chitra, K., and Ramya Sreedevi, V. 2011. "Does Personality Traits Influence the
Choice of Investment?," IUP Journal of Behavioral Finance (8:2), pp. 47-57.
Cipriani, M., and Guarino, A. 2009. "Herd Behavior in Financial Markets: An
Experiment with Financial Market Professionals," Journal of the European
Economic Association (7:1), pp. 206-233.
Cohen, J. B. 1983. "Involvement and You: 1000 Great Ideas," ACR North American
Advances).
Colliander, J., and Dahlén, M. 2011. "Following the Fashionable Friend: The Power
of Social Media: Weighing Publicity Effectiveness of Blogs Versus Online
Magazines," Journal of advertising research (51:1), pp. 313-320.
Comegys, C., Hannula, M., and Väisänen, J. 2006. "Longitudinal Comparison of
Finnish and Us Online Shopping Behaviour among University Students: The
Five-Stage Buying Decision Process," Journal of Targeting, Measurement and
Analysis for Marketing (14:4), pp. 336-356.
Connect. 2019. "Kundenbarometer Mobilfunk 2019." Retrieved 12.1.2020, from
https://www.connect.de/vergleich/connect-kundenbarometer-mobilfunk-2019-
zufriedenheit-anbieter-3199757-8459.html
Correa, T., Hinsley, A. W., and de Zúñiga, H. G. 2010. "Who Interacts on the Web?:
The Intersection of Users’ Personality and Social Media Use," Computers in
Costa, P. T., and McCrae, R. R. 1992. "Manual for the Revised Neo Personality
Inventory (Neo-Pi-R) and Neo Five-Factor Inventory (Neo-Ffi)," Odessa, FL:
Psychological Assessment Resources).
Creswell, J. W. 2009. "Research Design: Qualitative and Mixed Methods
Approaches," London and Thousand Oaks: Sage Publications).
Creswell, J. W., and Miller, D. L. 2000. "Determining Validity in Qualitative Inquiry,"
Theory into practice (39:3), pp. 124-130.
Creswell, J. W., and Poth, C. N. 2016. Qualitative Inquiry and Research Design:
Choosing among Five Approaches. Sage publications.
223
References
__________________________________________________________________________
Crisci, R., and Kassinove, H. 1973. "Effect of Perceived Expertise, Strength of

Advice, and Environmental Setting on Parental Compliance," The Journal of
Social Psychology (89:2), pp. 245-250.
Daft, R. L., and Lengel, R. H. 1986. "Organizational Information Requirements, Media
Richness and Structural Design," Management science (32:5), pp. 554-571.
DataReportal. 2020a. "Anzahl Der Aktiven Social-Media-Nutzer Weltweit in Den
Jahren 2015 Bis 2020" Retrieved 24.3.2020, from https://de.statista.com/
statistik/daten/studie/739881/umfrage/monatlich-aktive-social-media-nutzer-
weltweit/
DataReportal. 2020b. "Ranking Der Größten Social Networks Und Messenger Nach
Der Anzahl Der Nutzer Im Januar 2020 (in Millionen)." Retrieved 23.3.2020,
from https://de.statista.com/statistik/daten/studie/181086/umfrage/die-weltweit-
groessten-social-networks-nach-anzahl-der-user/
Davis, F. D. 1989. "Perceived Usefulness, Perceived Ease of Use, and User
Acceptance of Information Technology," MIS quarterly), pp. 319-340.
DellaVigna, S. 2009. "Psychology and Economics: Evidence from the Field," Journal
of Economic Literature (47:2), pp. 315-372.
Deutsch, M., and Gerard, H. B. 1955. "A Study of Normative and Informational Social
Influences Upon Individual Judgment," Journal of Abnormal and Social
Psychology (51:3), pp. 629-636.
Devaraj, S., Easley, R. F., and Crant, J. M. 2008. "Research Note—How Does
Personality Matter? Relating the Five-Factor Model to Technology Acceptance
and Use," Information systems research (19:1), pp. 93-105.
Dewan, S., Ho, Y.-J., and Ramaprasad, J. 2017. "Popularity or Proximity:
Characterizing the Nature of Social Influence in an Online Music Community,"
Information Systems Research (28:1), pp. 117-136.
Dobre, C., and Milovan-Ciuta, A.-M. 2015. "Personality Influences on Online Stores
Customers Behavior," Ecoforum Journal (4:1), p. 9.
Doh, S.-J., and Hwang, J.-S. 2009. "How Consumers Evaluate Ewom (Electronic
Word-of-Mouth) Messages," CyberPsychology & Behavior (12:2), pp. 193-197.
Duan, W., Gu, B., and Whinston, A. B. 2009. "Informational Cascades and Software
Adoption on the Internet: An Empirical Investigation," MIS Quarterly (33:1), pp.
23-48.
224
References
__________________________________________________________________________
Dulleck, U., Kerschbamer, R., and Sutter, M. 2011. "The Economics of Credence
Goods: An Experiment on the Role of Liability, Verifiability, Reputation, and
Competition," American Economic Review (101:2), pp. 526-555.
Edelmann, N. 2013. "Reviewing the Definitions of 'Lurkers' and Some Implications for
Online Research," CyberPsychology, Behavior & Social Networking (16:9), pp.
645-649.
Eighmey, J., and McCord, L. 1998. "Adding Value in the Information Age: Uses and
Gratifications of Sites on the World Wide Web," Journal of business research
(41:3), pp. 187-194.
Engel, J., Blackwell, R., and Miniard, P. 1995. Consumer Behavior, (8th, Forth Worth
ed.). Dryden Press, Texas.
Erkan, I., and Evans, C. 2018. "Social Media or Shopping Websites? The Influence of
Ewom on Consumers’ Online Purchase Intentions," Journal of Marketing
Communications (24:6), pp. 617-632.
Evans, C., Hackney, R., Rauniar, R., Rawski, G., Yang, J., and Johnson, B. 2014.
"Technology Acceptance Model (Tam) and Social Media Usage: An Empirical
Study on Facebook," Journal of Enterprise Information Management).
Eynon, R., Fry, J., and Schroeder, R. 2017. "The Ethics of Online Research," The
SAGE handbook of online research methods (2), pp. 19-37.
Fielding, N. G., Lee, R. M., and Blank, G. 2017. "The Sage Handbook of Online
Research Methods." 55 City Road, London.
Fink, A. 2002. "The Survey Handbook 2nd Edition. The Survey Kit." Sage
Publications, Inc.
Fishbein, M., and Ajzen, I. 1977. "Belief, Attitude, Intention, and Behavior: An
Introduction to Theory and Research".
Fisher, R. A. 1922. "On the Interpretation of Χ 2 from Contingency Tables, and the
Calculation of P," Journal of the Royal Statistical Society (85:1), pp. 87-94.
Fiske, S. T. 1980. "Attention and Weight in Person Perception: The Impact of
Negative and Extreme Behavior," Journal of personality and Social
Psychology (38:6), p. 889.
Fleder, D. M., Hosanagar, K., and Buja, A. 2010. "Recommender Systems and Their
Effects on Consumers: The Fragmentation Debate," EC (229), p. 230.
Flick, U., Kardorff, E., and Steinke, I. 2012. Qualitative Forschung: Ein Handbuch.
Reinbek bei Hamburg: Rowohlt.
225
References
__________________________________________________________________________
Foscht, T., and Swoboda, B. 2011. Käuferverhalten, (1 ed.). Wiesbaden: Gabler

Verlag.
Fowler, F. J. 2013. Survey Research Methods. SAGE Publications.
Fowler, J. H., Christakis, N. A., Steptoe, and Roux, D. 2009. "Dynamic Spread of
Happiness in a Large Social Network: Longitudinal Analysis of the
Framingham Heart Study Social Network," BMJ: British Medical Journal), pp.
23-27.
French, J. R., Raven, B., and Cartwright, D. 1959. "The Bases of Social Power,"
Classics of organization theory (7), pp. 311-320.
Freud, S. 1922. "Group Psychology and the Analysis of the Ego," Infection (64), p.
65.
Fritsche, I., and Linneweber, V. 2006. "Nonreactive Methods in Psychological
Research," in Handbook of Multimethod Measurement in Psychology, E.D.
Michael Eid (ed.). Washington, DC: American Psychological Association, pp.
189-203.
Früh, W. 2011. Inhaltsanalyse. Theorie Und Praxis. Konstanz: UVK.
Furnham, A. 1996. "The Big Five Versus the Big Four: The Relationship between the
Myers–Briggs Type Indicator (Mbti) and Neo-Pi Five Factor Model of
Personality," Personality and Individual Differences (21:2), pp. 303-307.
Gee-Woo, B., Ahuja, M. K., Ayoung, S., and Lee Xian, Y. 2015. "Sustainability of a
Virtual Community: Integrating Individual and Structural Dynamics," Journal of
the Association for Information Systems (16:6), pp. 418-447.
Gibbs, G. R. 2007. "Analytic Quality and Ethics," Analyzing qualitative data. London,
UK: Sage Publications).
Gläser, J., and Laudel, G. 2010. Experteninterviews Und Qualitative Inhaltsanalyse.
Wiesbaden: VS Verlag
Godes, D., Mayzlin, D., Chen, Y., Das, S., Dellarocas, C., Pfeiffer, B., Libai, B., Sen,
S., Shi, M., and Verlegh, P. 2005. "The Firm's Management of Social
Interactions," Marketing Letters (16:3), pp. 415-428.
Godinho de Matos, M., Ferreira, P., and Krackhardt, D. 2014. "Peer Influence in the
Diffusion of Iphone 3g over a Large Social Network," MIS Quarterly (38:4), pp.
1103-A1115.
Goffman, E. 1959. "The Presentation of Self in Everyday Life, New York, Ny:
Doubleday Anchor Books".
226
References
__________________________________________________________________________
Goldberg, L. R. 1992. "The Development of Markers for the Big-Five Factor

Structure," Psychological assessment (4:1), p. 26.
Goldsmith, R. E., and Emmert, J. 1991. "Measuring Product Category Involvement: A
Multitrait-Multimethod Study," Journal of Business Research (23:4), pp. 363-
371.
Goldsmith, R. E., and Horowitz, D. 2006. "Measuring Motivations for Online Opinion
Seeking," Journal of Interactive Advertising (6:2), pp. 2-14.
Gosling, S. D., Rentfrow, P. J., and Swann, W. B. 2003. "A Very Brief Measure of the
Big-Five Personality Domains," Journal of Research in Personality (37:6), pp.
504-528.
Graff, J., Sophonthummapharn, K., and Parida, V. 2012. "Post-Purchase Cognitive
Dissonance : Evidence from the Mobile Phone Market," International Journal
of Technology Marketing (7:1), pp. 32-46.
Granovetter, M. S. 1973. "The Strength of Weak Ties'," American Journal of
Sociology (78:6), pp. 1360-1380.
Grayson, K. 2007. "Friendship Versus Business in Marketing Relationships," Journal
of Marketing (71:4), pp. 121-139.
Greene, J. C., and Caracelli, V. J. 1997. Advances in Mixed-Method Evaluation: The
Challenges and Benefits of Integrating Diverse Paradigms. Jossey-Bass
Publishers San Francisco, CA.
Greenwood, B. N., and Gopal, A. 2015. "Research Note—Tigerblood: Newspapers,
Blogs, and the Founding of Information Technology Firms," Information
Systems Research (26:4), pp. 812-828.
Greenwood, B. N., and Gopal, A. 2017. "Ending the Mending Wall: Herding, Media
Coverage, and Colocation in It Entrepreneurship," MIS Quarterly (41:3), pp.
989-A914.
Gregor, S., and Hevner, A. R. 2013. "Positioning and Presenting Design Science
Research for Maximum Impact," MIS quarterly), pp. 337-355.
Gu, B., Park, J., and Konana, P. 2012. "Research Note—the Impact of External
Word-of-Mouth Sources on Retailer Sales of High-Involvement Products,"
Information Systems Research (23:1), pp. 182-196.
Hanna, R., Rohm, A., and Crittenden, V. L. 2011. "We’re All Connected: The Power
of the Social Media Ecosystem," Business Horizons (54:3), pp. 265-273.
227
References
__________________________________________________________________________
Harwood, T. G., and Garry, T. 2003. "An Overview of Content Analysis," The
marketing review (3:4), pp. 479-498.
Häubl, G., Dellaert, B. G. C., and Donkers, B. 2010. "Tunnel Vision: Local Behavioral
Influences on Consumer Decisions in Product Search," Marketing Science
(29:3), pp. 438-455.
Hennig-Thurau, T., Gwinner, K. P., Walsh, G., and Gremler, D. D. 2004. "Electronic
Word-of-Mouth Via Consumer-Opinion Platforms: What Motivates Consumers
to Articulate Themselves on the Internet?," Journal of Interactive Marketing
(18:1), pp. 38-52.
Hennig-Thurau, T., Malthouse, E. C., Friege, C., Gensler, S., Lobschat, L.,
Rangaswamy, A., and Skiera, B. 2010. "The Impact of New Media on
Customer Relationships," Journal of service research (13:3), pp. 311-330.
Herrmann, A., and Homburg, C. 2000. "Marktforschung: Ziele, Vorgehensweisen Und
Methoden".
Hewson, C. 2017. "The Sage Handbook of Online Research Methods." 55 City Road
55 City Road, London: SAGE Publications Ltd.

Hirschmeier, S., Tilly, R., and Schoder, D. 2016. "Information Quality Needs
Throughout the Purchase Process".
Hirshleifer, D., and Hong Teoh, S. 2003. "Herd Behaviour and Cascading in Capital
Markets: A Review and Synthesis," European Financial Management (9:1),
pp. 25-66.
Ho, Y.-C., Wu, J., and Tan, Y. 2017. "Disconfirmation Effect on Online Rating
Behavior: A Structural Model," Information Systems Research (28:3), pp. 626-
642.
Holsti, O. R. 1969. "Content Analysis for the Social Sciences and Humanities,"
Reading. MA: Addison-Wesley (content analysis).
Hong, Y., and Pavlou, P. A. 2014. "Product Fit Uncertainty in Online Markets: Nature,
Effects, and Antecedents," Information Systems Research (25:2), pp. 328-344.
Horton, J. J. 2014. "Misdirected Search Effort in a Matching Market: Causes,
Consequences and a Partial Solution," Proceedings of the fifteenth ACM
conference on Economics and computation, pp. 357-357.
Howard, J. A., and Sheth, J. N. 1969. "The Theory of Buyer Behavior," New York.
228
References
__________________________________________________________________________
Hsu, C.-L., and Lin, J. C.-C. 2008. "Acceptance of Blog Usage: The Roles of
Technology Acceptance, Social Influence and Knowledge Sharing Motivation,"
Information & management (45:1), pp. 65-74.
Huang, A. H., Chen, K., Yen, D. C., and Tran, T. P. 2015. "A Study of Factors That
Contribute to Online Review Helpfulness," Computers in Human Behavior
(48), pp. 17-27.
Huang, J.-H., and Chen, Y.-F. 2008. "Herding in Online Product Choice," Psychology
& Marketing (23:5), pp. 413-428.
Huang, L.-T., and Farn, C.-K. 2009. "Effects of Virtual Communities on Purchasing
Decision-Making: The Moderating Role of Information Activities," PACIS 2009
Proceedings), p. 55.
Huang, L.-T., Farn, C.-K., and Jeng, H.-T. 2012. "Motivations for Using Information
for Decision Making in Virtual Communities: The Moderating Effects of Usage
Behavior," Pacific Asia Journal of the Association for Information Systems
(4:1).
Huang, P., Lurie, N. H., and Mitra, S. 2009. "Searching for Experience on the Web:
An Empirical Examination of Consumer Behavior for Search and Experience
Goods," Journal of Marketing (73:2), pp. 55-69.
Huber, G. P., and Power, D. J. 1985. "Retrospective Reports of Strategic‐Level
Managers: Guidelines for Increasing Their Accuracy," Strategic management
journal (6:2), pp. 171-180.
Hughes, D. J., Rowe, M., Batey, M., and Lee, A. 2012. "A Tale of Two Sites: Twitter
Vs. Facebook and the Personality Predictors of Social Media Usage,"
Computers in Human Behavior (28:2), pp. 561-569.
Hwang, H., Jung, T., and Suh, E. 2004. "An Ltv Model and Customer Segmentation
Based on Customer Value: A Case Study on the Wireless Telecommunication
Industry," Expert Systems with Applications (26:2), pp. 181-188.
Iriberri, A., and Leroy, G. 2009. "A Life-Cycle Perspective on Online Community
Success," ACM Computing Surveys (41:2).
Jabr, W., and Zheng, Z. 2014. "Know Yourself and Know Your Enemy: An Analysis of
Firm Recommendations and Consumer Reviews in a Competitive
Environment," MIS Quarterly (38:3), pp. 635-A610.
Jain, K., and Srinivasan, N. 1990. "An Empirical Assessment of Multiple
Operationalizations of Involvement," ACR North American Advances).
229
References
__________________________________________________________________________
Janetzko, D. 2017. "The Sage Handbook of Online Research Methods." 55 City

Road, London: SAGE Publications Ltd.
Jang, H., Olfman, L., Ko, I., Koh, J., and Kim, K. 2008. "The Influence of on-Line
Brand Community Characteristics on Community Commitment and Brand
Loyalty," International Journal of Electronic Commerce (12:3), pp. 57-80.
Jeck-Schlottmann, G. 1988. "Anzeigenbetrachtung Bei Geringem Involvement,"
Marketing: Zeitschrift für Forschung und Praxis), pp. 33-43.
John, O. P., and Srivastava, S. 1999. "The Big Five Trait Taxonomy: History,
Measurement, and Theoretical Perspectives," Handbook of personality:
Theory and research (2:1999), pp. 102-138.
Joinson, A. N. 2008. "Looking at, Looking up or Keeping up with People? Motives
and Use of Facebook," Proceedings of the SIGCHI conference on Human
Factors in Computing Systems, pp. 1027-1036.
Kahneman, D. 2011. Thinking, Fast and Slow. London: Penguin Group.
Kahneman, D., Knetsch, J. L., and Thaler, R. H. 1991. "Anomalies: The Endowment
Effect, Loss Aversion, and Status Quo Bias," Journal of Economic
Perspectives (5:1), pp. 193-206.
Kahneman, D., and Tversky, A. 1979. "Prospect Theory: An Analysis of Decision
under Risk," Econometrica (47:2), pp. 263-291.
Kaplan, A. M., and Haenlein, M. 2010. "Users of the World, Unite! The Challenges
and Opportunities of Social Media," Business Horizons (53:1), pp. 59-68.
Kaplan, B., and Maxwell, J. A. 2005. "Qualitative Research Methods for Evaluating
Computer Information Systems," in Evaluating the Organizational Impact of
Healthcare Information Systems. Springer, pp. 30-55.
Kapoor, K. K., Tamilmani, K., Rana, N. P., Patil, P., Dwivedi, Y. K., and Nerur, S.
2018. "Advances in Social Media Research: Past, Present and Future,"
Information Systems Frontiers (20:3), pp. 531-558.
Kassarjian, H. H. 1982. "The Development of Consumer Behavior Theory," ACR
North American Advances).
Katz, E., and Lazarsfeld, P. F. 1955. Personal Influence. NewYork: Free Press.
Katz, M. L., and Shapiro, C. 1985. "Network Externalities, Competition, and
Compatibility," The American economic review (75:3), pp. 424-440.
Kavanaugh, A. L., Reese, D. D., Carroll, J. M., and Rosson, M. B. 2005. "Weak Ties
in Networked Communities," The Information Society (21:2), pp. 119-131.
230
References
__________________________________________________________________________
Kelley, H. H., and Michela, J. L. 1980. "Attribution Theory and Research," Annual
review of psychology (31:1), pp. 457-501.
Kelly, M., and O Grada, C. 2000. "Market Contagion: Evidence from the Panics of
1854 and 1857," American Economic Review (90:5), pp. 1110-1124.
Kelman, H. C. 1958. "Compliance, Identification, and Internalization Three Processes
of Attitude Change," Journal of conflict resolution (2:1), pp. 51-60.
Keutel, M., Michalik, B., and Richter, J. 2014. "Towards Mindful Case Study
Research in Is: A Critical Analysis of the Past Ten Years," European Journal
of Information Systems (23:3), pp. 256-272.
Keynes, J. 1936. "The General Theory of Employment, Interest and Money
Macmillan Cambridge University Press, for Royal Economic Society in 1936."
Khazanchi, D., and Munkvold, B. E. 2003. "On the Rhetoric and Relevance of Is
Research Paradigms: A Conceptual Framework and Some Propositions," 36th
Annual Hawaii International Conference on System Sciences, 2003.
Proceedings of the: IEEE, p. 10 pp.
Kiecker, P., and Cowles, D. 2002. "Interpersonal Communication and Personal
Influence on the Internet: A Framework for Examining Online Word-of-Mouth,"
Journal of Euromarketing (11:2), pp. 71-88.
Kim, H.-S., Brubaker, P., and Seo, K. 2015a. "Examining Psychological Effects of
Source Cues and Social Plugins on a Product Review Website," Computers in
Human Behavior (49), pp. 74-85.
Kim, S., Lee, J., and Yoon, D. 2015b. "Norms in Social Media: The Application of
Theory of Reasoned Action and Personal Norms in Predicting Interactions with
Facebook Page Like Ads," Communication Research Reports (32:4), pp. 322-
331.
Kim, S. J., Maslowska, E., and Malthouse, E. C. 2018. "Understanding the Effects of
Different Review Features on Purchase Probability," International Journal of
Advertising (37:1), pp. 29-53.
King, R. A., Racherla, P., and Bush, V. D. 2014. "What We Know and Don't Know
About Online Word-of-Mouth: A Review and Synthesis of the Literature,"
Journal of interactive marketing (28:3), pp. 167-183.
Kirman, A. 1993. "Ants, Rationality, and Recruitment," The Quarterly Journal of
Economics (108:1), pp. 137-156.
231
References
__________________________________________________________________________
Klein, H. K., and Myers, M. D. 1999. "A Set of Principles for Conducting and
Evaluating Interpretive Field Studies in Information Systems," MIS quarterly),
pp. 67-93.
Klein, L. R. 1998. "Evaluating the Potential of Interactive Media through a New Lens:
Search Versus Experience Goods," Journal of Business Research (41:3), pp.
195-203.
Kober, S. E., and Neuper, C. 2013. "Personality and Presence in Virtual Reality:
Does Their Relationship Depend on the Used Presence Measure?,"
International Journal of Human-Computer Interaction (29:1), pp. 13-25.
Kollock, P., and Smith, M. 1996. "Managing the Virtual Commons," Computer-
mediated communication: Linguistic, social, and cross-cultural perspectives),
pp. 109-128.
Krippendorff, K. 2004. Content Analysis: An Introduction to Its Methodology, (2nd
ed.). Thousand Oaks: Sage Publications, Inc.
Kroeber-Riel, W., Weinberg, P., and Gröppel-Klein, A. 2009. "Konsumentenverhalten
(9. Auﬂ.) München: Vahlen".
Kruskal, W. H., and Wallis, W. A. 1952. "Use of Ranks in One-Criterion Variance
Analysis," Journal of the American statistical Association (47:260), pp. 583-
621.
Kuan, K. K. Y., Zhong, Y., and Chau, P. Y. K. 2014. "Informational and Normative
Social Influence in Group-Buying: Evidence from Self-Reported and Eeg
Data," Journal of Management Information Systems (30:4), pp. 151-178.
Kuckartz, U. 2007. "Computergestützte Analyse Qualitativer Daten," in Qualitative
Marktforschung. Springer, pp. 713-730.
Kuckartz, U. 2014. Mixed Methods: Methodologie, Forschungsdesigns Und
Analyseverfahren. Springer-Verlag.
Kuhn, T. S. 1962. "The Structure of Scientifi Revolutions," The Un. of Chicago Press
(2), p. 90.
Kunst, K., and Vatrapu, R. 2018. "Understanding Electronic Word of Behavior:
Conceptualization of the Observable Digital Traces of Consumers’ Behaviors,"
Electronic Markets), pp. 1-14.
Kuß, A., and Tomczak, T. 2007. Käuferverhalten (4 ed.). Stuttgart: Lucius & Lucius.
Kwahk, K.-Y., and Ge, X. 2012. "The Effects of Social Media on E-Commerce: A
Perspective of Social Impact Theory,"), pp. 1814-1823.
232
References
__________________________________________________________________________
Kwahk, K.-y., and Kim, B. 2017. "Effects of Social Media on Consumers’ Purchase
Decisions: Evidence from Taobao," Service Business (11:4), pp. 803-829.
Kyle, A. S., and Xiong, W. 2001. "Contagion as a Wealth Effect," The Journal of
Finance (56:4), pp. 1401-1440.
Lammenett, E. 2017. "Preissuchmaschinen Und Vergleichsportale," in Praxiswissen
Online-Marketing: Affiliate- Und E-Mail-Marketing, Suchmaschinenmarketing,
Online-Werbung, Social Media, Facebook-Werbung. Wiesbaden: Springer
Fachmedien Wiesbaden, pp. 263-267.
Larsen, K. R., Eargle, D. (Eds.). 2015. "Theories Used in Is Research Wiki."
Retrieved 05.03.2019, from http://IS.TheorizeIt.org
Lastovicka, J. L., and Gardner, D. M. 1979. "Components of Involvement," Attitude
research plays for high stakes), pp. 53-73.
Latané, B., Williams, K., and Harkins, S. 1979. "Many Hands Make Light the Work:
The Causes and Consequences of Social Loafing," Journal of personality and
social psychology (37:6), p. 822.
Laurent, G., and Kapferer, J.-N. 1985. "Measuring Consumer Involvement Profiles,"
Journal of marketing research (22:1), pp. 41-53.
Le Bon, G. 1897. The Crowd: A Study of the Popular Mind. T. Fisher Unwin.
Lee, A. S., and Baskerville, R. L. 2012. "Conceptualizing Generalizability: New
Contributions and a Reply," MIS quarterly), pp. 749-761.
Lee, F. S. L., Vogel, D., and Limayem, M. 2003. "Virtual Community Informatics: A
Review and Research," The Journal of Information Technology Theory and
Application (JITTA) (5:1), pp. 47-61.
Lee, M., and Youn, S. 2009. "Electronic Word of Mouth (Ewom)," International
Journal of Advertising (28:3), pp. 473-499.
Lee, R. M., Fielding, N. G., and Blank, G. 2017. "Online Research Methods in the
Social Sciences: An Editorial Introduction".
Lee, Y.-J., Hosanagar, K., and Tan, Y. 2015. "Do I Follow My Friends or the Crowd?
Information Cascades in Online Movie Ratings," Management Science (61:9),
pp. 2241-2258.
Leibenstein, H. 1950. "Bandwagon, Snob, and Veblen Effects in the Theory of
Consumers' Demand," The Quarterly Journal of Economics (64:2), pp. 183-
207.
233
References
__________________________________________________________________________
Leonardi, P. M. 2015. "Ambient Awareness and Knowledge Acquisition: Using Social

Media to Learn "Who Knows What" and "Who Knows Whom"," MIS Q. (39:4),
pp. 747-762.
Leshed, G. 2005. "Posters, Lurkers, and in Between: A Multidimensional Model of
Online Community Participation Patterns," Poster presented at HCI
International, Las Vegas, NV).
Levy, D. A., and Nail, P. R. 1993. "Contagion: A Theoretical and Empirical Review
and Reconceptualization," Genetic, social, and general psychology
monographs).
Li, F., and Whalley, J. 2002. "Deconstruction of the Telecommunications Industry:
From Value Chains to Value Networks," Telecommunications Policy (26:9),
pp. 451-472.
Li, Q., Liang, N., and Li, E. Y. 2018. "Does Friendship Quality Matter in Social
Commerce? An Experimental Study of Its Effect on Purchase Intention,"
Electronic Commerce Research (18:4), pp. 693-717.
Li, X., and Wu, L. 2018. "Herding and Social Media Word-of-Mouth: Evidence from
Groupon," MIS Quarterly (42:4), pp. 1331-1351.
Lin, C., Wu, Y. S., and Chen, J.-C. V. 2013. "Electronic Word-of-Mouth: The
Moderating Roles of Product Involvement and Brand Image," TIIM 2013
Proceedings), pp. 39-47.
Liu, Q., Huang, S., and Zhang, L. 2016. "The Influence of Information Cascades on
Online Purchase Behaviors of Search and Experience Products," Electronic
Commerce Research (16:4), pp. 553-580.
Liu, Q. B., and Karahanna, E. 2017. "The Dark Side of Reviews: The Swaying Effects
of Online Product Reviews on Attribute Preference Construction," MIS
Quarterly (41:2), pp. 427-A426.
Liu, Y. 2006. "Word of Mouth for Movies: Its Dynamics and Impact on Box Office
Revenue," Journal of marketing (70:3), pp. 74-89.
Liu, Y., Feng, J., and Liao, X. 2017. "When Online Reviews Meet Sales Volume
Information: Is More or Accurate Information Always Better?," Information
Systems Research (28:4), pp. 723-743.
Lizardo, O., Mowry, R., Sepulvado, B., Stoltz, D. S., Taylor, M. A., Van Ness, J., and
Wood, M. 2016. "What Are Dual Process Models? Implications for Cultural
Analysis in Sociology," Sociological Theory (34:4), pp. 287-310.
234
References
__________________________________________________________________________
Luo, B., and Lin, Z. 2013. "A Decision Tree Model for Herd Behavior and Empirical
Evidence from the Online P2p Lending Market," Information Systems & e-
Business Management (11:1), pp. 141-160.
Luo, X., and Zhang, J. 2013. "How Do Consumer Buzz and Traffic in Social Media
Marketing Predict the Value of the Firm?," Journal of Management Information
Systems (30:2), pp. 213-238.
Lux, T. 1995. "Herd Behaviour, Bubbles and Crashes," The Economic Journal
(105:431), pp. 881-896.
Maas, P., Schlager, T., Steiner, P. H., and Taborelli, R. 2014. "Die Rolle Von Social
Media Im Kaufentscheidungsprozess - Eine Internationale Studie in Der
Tourismus-, Elektronik- Und Versicherungsindustrie," in: I.VW-HSG. St.
Gallen.
Madhavi, N. B., and Rao, T. 2018. "Personality Traits Influencing the Purchase
Decisions of Passenger Car Buyers of Toyota in Krishna District," ICRTEMMS
Conference Proceedings: Swarna Bharathi lnstitute of Science and
Technology, pp. 603-608.
Majchrzak, A., Faraj, S., Kane, G. C., and Azad, B. 2013. "The Contradictory
Influence of Social Media Affordances on Online Communal Knowledge
Sharing," Journal of Computer-Mediated Communication (19:1), pp. 38-55.
Mangold, W. G., and Faulds, D. J. 2009. "Social Media: The New Hybrid Element of
the Promotion Mix," Business Horizons (52:4), pp. 357-365.
Martin, S. P., and Robinson, J. P. 2007. "The Income Digital Divide: Trends and
Predictions for Levels of Internet Use," Social problems (54:1), pp. 1-22.
Mason, B. 1999. "Issues in Virtual Ethnography," Ethnographic studies in real and
virtual environments: Inhabited information spaces and connected
communities), pp. 61-69.
Matook, S., Brown, S. A., and Rolf, J. 2015. "Forming an Intention to Act on
Recommendations Given Via Online Social Networks," European Journal of
Information Systems (24:1), pp. 76-92.
Matute, J., Polo-Redondo, Y., and Utrillas, A. 2016. "The Influence of Ewom
Characteristics on Online Repurchase Intention," Online Information Review
(40:7), pp. 1090-1110.
Mayring, P. 2000. "Qualitative Inhaltsanalyse [28 Absätze]," Forum Qualitative
Sozialforschung/Forum: Qualitative Social Research, pp. 2-00.
235
References
__________________________________________________________________________
Mayring, P. 2015. Qualitative Inhaltsanalyse. Grundlagen Und Techniken, (12 ed.).

Weinheim und Basel: Beltz Verlag.
McCrae, R. R., and Costa, P. T. 2008. "The Five-Factor Theory of Personality,").
McPherson, M., Smith-Lovin, L., and Cook, J. M. 2001. "Birds of a Feather:
Homophily in Social Networks," Annual review of sociology (27:1), pp. 415-
444.
Meiseberg, B. 2014. "The Effectiveness of E-Tailers’ Communication Practices in
Stimulating Sales of Slow-Selling Versus Best-Selling Products".
Mertens, P. 2010. "Anspruchsgruppen Der Gestaltungsorientierten
Wirtschaftsinformatik," Wirtschaftsinformatik: Ein Plädoyer für Rigor und
Relevanz), p. 19.
Miller, G. A. 1956. "The Magical Number Seven, Plus or Minus Two: Some Limits on
Our Capacity for Processing Information," Psychological review (63:2), p. 81.
Miller, V. 2011. Understanding Digital Culture. Sage Publications.
Milliken, F. J. 1987. "Three Types of Perceived Uncertainty About the Environment:
State, Effect, and Response Uncertainty," Academy of Management review
(12:1), pp. 133-143.
MindTake. 2017. "Welche Informationsquellen Nutzen Sie, Bevor Sie Sich Ein Neues
Handy Kaufen?" Retrieved 5.2.2020, from https://de.statista.com/statistik/
daten/studie/631535/umfrage/umfrage-in-oesterreich-zu-informationsquellen-
bei-der-neuanschaffung-eines-handys/
MindTake. 2019. "Consumer-Check Zur Generation Z." Retrieved 13.2.2020, from
https://retailreport.at/sites/default/files/2019-07/ConsumerCheck%20
Generation%20Z.pdf
Mingers, J. 2001. "Combining Is Research Methods: Towards a Pluralist
Methodology," Information systems research (12:3), pp. 240-259.
Miranda, S. M., Young, A., and Yetgin, E. 2016. "Are Social Media Emancipatory or
Hegemonic? Societal Effects of Mass Media Digitization in the Case of the
Sopa Discourse," MIS quarterly (40:2), pp. 303-329.
Mitchell, A. A. 1979. "Involvement: A Potentially Important Mediator of Consumer
Behavior," ACR North American Advances).
Mittal, B. 1989. "Measuring Purchase‐Decision Involvement," Psychology &
Marketing (6:2), pp. 147-162.
236
References
__________________________________________________________________________
Moon, B.-J. 2004. "Consumer Adoption of the Internet as an Information Search and
Product Purchase Channel: Some Research Hypotheses," International
Journal of Internet Marketing and Advertising (1:1), pp. 104-118.
Moran, G., and Muzellec, L. 2017. "Ewom Credibility on Social Networking Sites: A
Framework," Journal of Marketing Communications (23:2), pp. 149-161.
Morone, A., and Samanidou, E. 2008. "A Simple Note on Herd Behaviour," Journal of
Evolutionary Economics (18:5), pp. 639-646.
Morris, R. 1994. "Computerized Content Analysis in Management Research: A
Demonstration of Advantages & Limitations," Journal of Management (20:4),
pp. 903-931.
Moscovici, S., Lage, E., and Naffrechoux, M. 1969. "Influence of a Consistent
Minority on the Responses of a Majority in a Color Perception Task,"
Sociometry), pp. 365-380.
mpfs. 2018. "Jim-Studie 2018" Retrieved 17.02.2020, from https://www.mpfs.de
/fileadmin/files/Studien/JIM/2018/Studie/JIM2018_Gesamt.pdf
Mudambi, S. M., and Schuff, D. 2010. "What Makes a Helpful Online Review? A
Study of Customer Reviews on Amazon.Com," MIS Quarterly (34:1), pp. 185-
200.
Myers, M. D. 1997. "Qualitative Research in Information Systems," MIS Quarterly), p.
241.
Myers, M. D., and Avison, D. E. 2002. Qualitative Research in Information Systems :
A Reader. London: Sage.
Myers, M. D., and Klein, H. K. 2011. "A Set of Principles for Conducting Critical
Research in Information Systems," MIS quarterly), pp. 17-36.
Nakayama, M., Wan, Y., and Sutcliffe, N. G. 2010. "Wom or Ewom or Something
Else: How Does the Web Affect Our Dependence on Shopping Information
Sources?," AMCIS, p. 446.
Nelson, P. 1974. "Advertising as Information," Journal of Political Economy (82:4),
pp. 729-754.
Nelson, R. J. 1975. "Behaviorism, Finite Automata, and Stimulus Response Theory,"
Theory and Decision (6:3), pp. 249-267.
Ngai, E. W. T., Tao, S. S. C., and Moon, K. K. L. 2015. "Social Media Research:
Theories, Constructs, and Conceptual Frameworks," International Journal of
Information Management (35:1), pp. 33-44.
237
References
__________________________________________________________________________
Nielsen. 2015. "Global Trust in Advertising Survey."

Nielsen, J. 2006. "Participation Inequality: Lurkers Vs. Contributors in Internet
Communities," Jakob Nielsen's Alertbox (107), p. 108.
Nock, S. L., and Guterbock, T. M. 2010. "Survey Experiments," Handbook of survey
research (2), pp. 837-865.
Nonnecke, B., Andrews, D., and Preece, J. 2006. "Non-Public and Public Online
Community Participation: Needs, Attitudes and Behavior," Electronic
Nonnecke, B., and Preece, J. 2000. "Lurker Demographics: Counting the Silent,"
Proceedings of the SIGCHI conference on Human Factors in Computing
Systems: ACM, pp. 73-80.
Nonnecke, B., Preece, J., and Andrews, D. 2004. "What Lurkers and Posters Think
of Each Other [Online Community]," 37th Annual Hawaii International
Conference on System Sciences, 2004. Proceedings of the: IEEE, p. 9 pp.
Nosek, B. A., Banaji, M. R., and Greenwald, A. G. 2002. "E‐Research: Ethics,
Security, Design, and Control in Psychological Research on the Internet,"
Journal of Social Issues (58:1), pp. 161-176.
Okoli, C. 2015. A Guide to Conducting a Standalone Systematic Literature Review.
Orlikowski, W. J., and Baroudi, J. J. 1991. "Studying Information Technology in
Organizations: Research Approaches and Assumptions," Information systems
research (2:1), pp. 1-28.
Österle, H., Becker, J., Frank, U., Hess, T., Karagiannis, D., Krcmar, H., Loos, P.,
Mertens, P., Oberweis, A., and Sinz, E. J. 2011. "Memorandum on Design-
Oriented Information Systems Research," European Journal of Information
Systems (20:1), pp. 7-10.
Outlook, S. D. M. 2019a. "Anzahl Der Nutzer Von Sozialen Netzwerken in
Ausgewählten Ländern Weltweit Im Jahr 2018 (in Millionen)." Retrieved
24.3.2020, from https://de.statista.com/statistik/daten/studie/554504/umfrage
/anzahl-der-monatlich-aktiven-nutzer-von-sozialen-netzwerken-in-
ausgewaehlten-laendern-weltweit/
Outlook, S. D. M. 2019b. "Ausgaben Für Social-Media-Werbung in Deutschland in
Den Jahren 2017 Und 2018 Sowie Eine Prognose Bis 2023 (in Millionen
Euro)." Retrieved 20.3.2020, from https://de.statista.com/statistik/daten/
studie/456177/umfrage/umsaetze-mit-social-media-werbung-in-deutschland/
238
References
__________________________________________________________________________
Paquette, H. 2013. "Social Media as a Marketing Tool: A Literature Review,").

Park, D.-H., and Kim, S. 2008. "The Effects of Consumer Knowledge on Message
Processing of Electronic Word-of-Mouth Via Online Consumer Reviews,"
Electronic commerce research and applications (7:4), pp. 399-410.
Park, D.-H., Lee, J., and Han, I. 2007. "The Effect of on-Line Consumer Reviews on
Consumer Purchasing Intention: The Moderating Role of Involvement,"
International Journal of Electronic Commerce (11:4), pp. 125-148.
Park, M.-S., Shin, J.-K., and Ju, Y. 2014. "The Effect of Online Social Network
Characteristics on Consumer Purchasing Intention of Social Deals," Global
Economic Review (43:1), pp. 25-41.
Parker, W. D., and Prechter, R. R. 2005. "Herding: An Interdisciplinary Integrative
Review from a Socionomic Perspective," Available at SSRN 2009898).
Pavlou, P. A., and Dimoka, A. 2006. "The Nature and Role of Feedback Text
Comments in Online Marketplaces: Implications for Trust Building, Price
Premiums, and Seller Differentiation," Information Systems Research (17:4),
pp. 392-414.
Pavlovic, N. 2018. "Factors Affecting Herd Behaviour in Buying Decisions Influenced
by Online Communities," Americas Conference on Information Systems.
Pearson, K. 1900. "X. On the Criterion That a Given System of Deviations from the
Probable in the Case of a Correlated System of Variables Is Such That It Can
Be Reasonably Supposed to Have Arisen from Random Sampling," The
London, Edinburgh, and Dublin Philosophical Magazine and Journal of
Science (50:302), pp. 157-175.
Penrose, L. S. 1952. On the Objective Study of Crowd Behaviour. HK Lewis.
Pérez-Latre, F. J., Portilla, I., and Sánchez-Blanco, C. 2012. "Social Networks, Media
and Audiences: A Literature Review".
Peters, K., Chen, Y., Kaplan, A. M., Ognibeni, B., and Pauwels, K. 2013. "Social
Media Metrics—a Framework and Guidelines for Managing Social Media,"
Journal of interactive marketing (27:4), pp. 281-298.
Pornpitakpan, C. 2004. "The Persuasiveness of Source Credibility: A Critical Review
of Five Decades' Evidence," Journal of applied social psychology (34:2), pp.
243-281.
239
References
__________________________________________________________________________
Powers, T., Advincula, D., Austin, M. S., Graiko, S., and Snyder, J. 2012. "Digital and
Social Media in the Purchase Decision Process," A Special Report from the
Advertising Research Foundation (52:4), pp. 479-489.
Prechter, R. R. 1999. The Wave Principle of Human Social Behavior and the New
Science of Socionomics. New Classics Library.
Prechter, R. R. 2001. "Unconscious Herding Behavior as the Psychological Basis of
Financial Market Trends and Patterns," The Journal of Psychology and
Financial Markets (2:3), pp. 120-125.
Prechter, R. R. 2003. Pioneering Studies in Socionomics. Elliott Wave International.
Preece, J., and Maloney-Kirchmar, D. 2005. "Online Communities: Design, Theory,
and Practice," Journal of Computer-Mediated Communication (10:4).
Preece, J., Nonnecke, B., and Andrews, D. 2004. "The Top Five Reasons for
Lurking: Improving Community Experiences for Everyone," Computers in
human behavior (20:2), pp. 201-223.
Purnawirawan, N., Dens, N., and De Pelsmacker, P. 2014. "Expert Reviewers
Beware! The Effects of Review Set Balance, Review Source and Review
Content on Consumer Responses to Online Reviews," Journal of Electronic
Raafat, R. M., Chater, N., and Frith, C. 2009a. "Herding in Humans," Trends Cogn
Sci (13:10), pp. 420-428.
Raafat, R. M., Chater, N., and Frith, C. 2009b. "Herding in Humans," Trends in
cognitive sciences (13:10), pp. 420-428.
Rammstedt, B., and John, O. P. 2007. "Measuring Personality in One Minute or Less:
A 10-Item Short Version of the Big Five Inventory in English and German,"
Journal of Research in Personality (41:1), pp. 203-212.
Rasmussen, K. B. 2017. "Data Quality in Online Environments," in: The SAGE
Handbook of Online Research Methods. 55 City Road, London: SAGE
Publications Ltd.
Rheingold, H. 1993. The Virtual Community. Homesteading in the Electronic Frontier.

Reading, MA.: Addison-Wesley Publishing Company.
Richins, M. L., and Bloch, P. H. 1991. "Post-Purchase Product Satisfaction:
Incorporating the Effects of Involvement and Time," Journal of Business
Research (23:2), pp. 145-158.
240
References
__________________________________________________________________________
Richter, D., Riemer, K., and vom Brocke, J. 2011. "Internet Social Networking,"
Wirtschaftsinformatik (53:2), pp. 89-103.
Ridings, C., Gefen, D., and Arinze, B. 2006. "Psychological Barriers: Lurker and
Poster Motivation and Behavior in Online Communities," Communications of
the association for Information Systems (18:1), pp. 329-354.
Rogers, E. 1962. "Diffusion of Innovations. New York: Free Press".
Rook, L. 2006. "An Economic Psychological Approach to Herd Behavior," Journal of
Economic Issues (40:1), pp. 75-95.
Rosen, P. A., and Kluemper, D. H. 2008. "The Impact of the Big Five Personality
Traits on the Acceptance of Social Networking Website," AMCIS 2008
proceedings).
Rössler, P., and Geise, S. 2013. "Standardisierte Inhaltsanalyse: Grundprinzipien,
Einsatz Und Anwendung," in Handbuch Standardisierte Erhebungsverfahren
in Der Kommunikationswissenschaft, W. Möhring and D. Schlütz (eds.).
Wiesbaden: Springer Fachmedien Wiesbaden, pp. 269-287.
Ryan, T., and Xenos, S. 2011. "Who Uses Facebook? An Investigation into the
Relationship between the Big Five, Shyness, Narcissism, Loneliness, and
Facebook Usage," Computers in Human Behavior (27:5), pp. 1658-1664.
Salazar, H. A., Oerlemans, L., and van Stroe‐Biezen, S. 2013. "Social Influence on
Sustainable Consumption: Evidence from a Behavioural Experiment,"
International Journal of Consumer Studies (37:2), pp. 172-180.
Salehan, M., Kim, D. J., and Changsu, K. 2017. "Use of Online Social Networking
Services from a Theoretical Perspective of the Motivation-Participation-
Performance Framework," Journal of the Association for Information Systems
(18:2), pp. 141-172.
Samuelson, W., and Zeckhauser, R. 1988. "Status Quo Bias in Decision Making,"
Journal of Risk and Uncertainty (1:1), pp. 7-59.
Scharfstein, D. S., and Stein, J. C. 1990. "Herd Behavior and Investment," The
American economic review), pp. 465-479.
Schlosser, A. E. 2005. "Posting Versus Lurking: Communicating in a Multiple
Audience Context," Journal of Consumer Research (32:2), pp. 260-265.
Schreier, M. 2014. "Varianten Qualitativer Inhaltsanalyse: Ein Wegweiser Im Dickicht
Der Begrifflichkeiten," Forum Qualitative Sozialforschung/Forum: Qualitative
Social Research: DEU.
241
References
__________________________________________________________________________
Schreiner, M., Hess, T., and Benlian, A. 2015. "Gestaltungsorientierter Kern Oder
Tendenz Zur Empirie? Zur Neueren Methodischen Entwicklung Der
Wirtschaftsinformatik," Arbeitsbericht, Institut für Wirtschaftsinformatik und
Neue Medien, Fakultät für Betriebswirtschaft, Ludwig-Maximilians-Universität
(Arbeitsbericht 1/2015), Retrieved 18.11.2018, from:https://www.econstor.eu/
bitstream/10419/109029/1/ 821607464.pdf
Schwaiger, M., and Meyer, A. 2011. Theorien Und Methoden Der Betriebswirtschaft:
Handbuch Für Wissenschaftler Und Studierende. Vahlen.
Scott, D., Meerman. 2015. The New Rules of Marketing and Pr, (5 ed.). Hoboken:
John Wiley & Sons, Inc.
Selwyn, N. 2006. "Digital Division or Digital Decision? A Study of Non-Users and
Low-Users of Computers," Poetics (34:4-5), pp. 273-292.
Sen, S., and Lerman, D. 2007. "Why Are You Telling Me This? An Examination into
Negative Consumer Reviews on the Web," Journal of interactive marketing
(21:4), pp. 76-94.
Shapiro, S. S., and Wilk, M. B. 1965. "An Analysis of Variance Test for Normality
(Complete Samples)," Biometrika (52:3/4), pp. 591-611.
Shen, G. C.-C., Chiou, J.-S., Hsiao, C.-H., Wang, C.-H., and Li, H.-N. 2016.
"Effective Marketing Communication Via Social Networking Site: The
Moderating Role of the Social Tie," Journal of Business Research (69:6), pp.
2265-2270.
Shen, X., Zhang, K. Z. K., and Zhao, S. J. 2014. "Understanding Information
Adoption in Online Review Communities: The Role of Herd Factors," 2014
47th Hawaii International Conference on System Sciences, pp. 604-613.
Sherif, M. 1936. "The Psychology of Social Norms".
Shiller, R. J. 1990. "Speculative Prices and Popular Models," Journal of Economic
perspectives (4:2), pp. 55-65.
Shiller, R. J. 2001. "Bubbles, Human Judgment, and Expert Opinion, Yale University-
Cowles Foundation." National Bureau of Economic Research (NBER).
Shiller, R. J. 2015. Irrational Exuberance: Revised and Expanded Third Edition.
Princeton university press.
Shiller, R. J., Fischer, S., and Friedman, B. M. 1984. "Stock Prices and Social
Dynamics," Brookings papers on economic activity (1984:2), pp. 457-510.
242
References
__________________________________________________________________________
Shiue, Y.-C., Chiu, C.-M., and Chang, C.-C. 2010. "Exploring and Mitigating Social
Loafing in Online Communities," Computers in Human Behavior (26:4), pp.
768-777.
Short, J., Williams, E., and Christie, B. 1976. The Social Psychology of
Telecommunications. John Wiley & Sons.
Simmel, G. 1949. "The Sociology of Sociability," American journal of sociology (55:3),
pp. 254-261.
Simon, H. A. 1955. "A Behavioral Model of Rational Choice," The quarterly journal of
economics (69:1), pp. 99-118.
Simon, H. A. 1990. "Invariants of Human Behavior," Annual review of psychology
(41:1), pp. 1-20.
Simonsohn, U., and Ariely, D. 2008. "When Rational Sellers Face Nonrational
Buyers: Evidence from Herding on Ebay," Management science (54:9), pp.
1624-1637.
Sinha, R., and Swearingen, K. 2001. "Comparing Recommendations Made by Online
Systems and Friends," In: DELOS Workshop: Personalisation and
Recommender Systems in Digital Libraries).
Skowronski, J. J., and Carlston, D. E. 1989. "Negativity and Extremity Biases in
Impression Formation: A Review of Explanations," Psychological bulletin
(105:1), p. 131.
Smith, A. D., and Rupp, W. T. 2003. "Strategic Online Customer Decision Making:
Leveraging the Transformational Power of the Internet," Online information
review).
Social, W. A. 2018. "Anzahl Der Facebook-Nutzer Nach Altersgruppen Und
Geschlecht in Deutschland Im Januar 2018 (in Millionen)." Retrieved
28.3.2020, from https://de.statista.com/statistik/daten/studie/512316/umfrage
/anzahl-der-facebook-nutzer-in-deutschland-nach-alter-und-geschlecht/
Solomon, M. R. 2010. Consumer Behaviour: A European Perspective. Pearson
education.
Sornette, D., and Andersen, J. V. 2002. "A Nonlinear Super-Exponential Rational
Model of Speculative Financial Bubbles," International Journal of Modern
Physics C (13:02), pp. 171-187.
243
References
__________________________________________________________________________
Souza de, C. S., and Preece, J. 2004. "A Framework for Analyzing and
Understanding Online Communities," Interacting with Computers (16:3), pp.
579–610.
Sparks, B. A., and Browning, V. 2011. "The Impact of Online Reviews on Hotel
Booking Intentions and Perception of Trust," Tourism Management (32:6), pp.
1310-1323.
Spyrou, S. 2013. "Herding in Financial Markets: A Review of the Literature," Review
of Behavioral Finance).
Stewart, D. W., and Pavlou, P. A. 2002. "From Consumer Response to Active
Consumer: Measuring the Effectiveness of Interactive Media," Journal of the
Academy of Marketing Science (30:4), pp. 376-396.
Stigler, G. J. 1961. "The Economics of Information," Journal of political economy
(69:3), pp. 213-225.
Stigler, S. M. 1986. The History of Statistics: The Measurement of Uncertainty before
1900. Harvard University Press.
Straub, E. T. 2009. "Understanding Technology Adoption: Theory and Future
Directions for Informal Learning," Review of Educational Research (79:2), pp.
625-649.
Strout, H. 2011. "Why Lurkers Are Valuable to Your Online Community," Farland
Group).
Sue, V. M., and Ritter, L. A. 2012. Conducting Online Surveys. Sage.
Sun, H. 2013. "A Longitudinal Study of Herd Behavior in the Adoption and Continued
Use of Technology," MIS Quarterly (37:4), pp. 1013-1041.
Sun, M. 2012. "How Does the Variance of Product Ratings Matter?," Management
Science (58:4), pp. 696-707.
Sundar, S. S., Oeldorf-Hirsch, A., and Xu, Q. 2008. "The Bandwagon Effect of
Collaborative Filtering Technology," in Chi'08 Extended Abstracts on Human
Factors in Computing Systems. pp. 3453-3458.
Sunstein, C. R. 2005. "Group Judgments: Statistical Means, Deliberation, and
Information Markets," NYUL Rev. (80), p. 962.
Suppes, P. 1993. "From Behaviorism to Neobehaviorism," in Models and Methods in
the Philosophy of Science: Selected Essays. Springer, pp. 341-355.
244
References
__________________________________________________________________________
Surowiecki, J. 2004. The Wisdom of Crowds: Why the Many Are Smarter Than the
Few and How Collective Wisdom Shapes Business, Economies, Societies,
and Nations. New York, NY, US: Doubleday & Co.
Svendsen, G. B., Johnsen, J.-A. K., Almås-Sørensen, L., and Vittersø, J. 2013.
"Personality and Technology Acceptance: The Influence of Personality Factors
on the Core Constructs of the Technology Acceptance Model," Behaviour &
Information Technology (32:4), pp. 323-334.
Tang, Q., Gu, B., and Whinston, A. B. 2012. "Content Contribution for Revenue
Sharing and Reputation in Social Media: A Dynamic Structural Model," Journal
of Management Information Systems (29:2), pp. 41-76.
Tarde, G. 1903. "The Laws of Imitation (H. Holt and Company)".
Thelwall, M. 2009. "Homophily in Myspace," Journal of the American Society for
Information Science and Technology (60:2), pp. 219-231.
Trommsdorff, V. 2004. Konsumentenverhalten, (6 ed.). Stuttgart: Kohlhammer.
Truong, Y., and Simmons, G. 2010. "Perceived Intrusiveness in Digital Advertising:
Strategic Marketing Implications," Journal of strategic marketing (18:3), pp.
239-256.
Tsai, F. f., and Reis, H. T. 2009. "Perceptions by and of Lonely People in Social
Networks," Personal Relationships (16:2), pp. 221-238.
Tsao, W.-c., and Hsieh, M.-t. 2015. "Ewom Persuasiveness: Do Ewom Platforms and
Product Type Matter?," Electronic Commerce Research (15:4), pp. 509-541.
Tseng, S.-L., Lu, S., Grover, V., and Weathers, D. 2017. "The Effect of Herding
Behavior on Online Review Voting Participation".
Tversky, A., and Kahneman, D. 1983. "Extensional Versus Intuitive Reasoning: The
Conjunction Fallacy in Probability Judgment," Psychological review (90:4), pp.
293-315.
Tversky, A., and Kahneman, D. 1991. "Loss Aversion in Riskless Choice: A
Reference-Dependent Model," The quarterly journal of economics (106:4), pp.
1039-1061.
Van Campenhout, G., and Verhestraeten, J.-F. 2010. "Herding Behavior among
Financial Analysts: A Literature Review".
Van Ginneken, J. 1992. Crowds, Psychology, and Politics, 1871-1899. Cambridge
University Press.
245
References
__________________________________________________________________________
Veblen, T. 1899. "The Theory of the Leisure Class: An Economic Study of

Institutions. ," New York: The Modern Library.
Veeraraghavan, S. K., and Debo, L. G. 2011. "Herding in Queues with Waiting Costs:
Rationality and Regret," Manufacturing & Service Operations Management
(13:3), pp. 329-346.
Velasquez, A., Wash, R., Lampe, C., and Bjornrud, T. 2014. "Latent Users in an
Online User-Generated Content Community," Computer Supported
Cooperative Work (CSCW) (23:1), pp. 21-50.
Vithayathil, J., Dadgar, M., and Osiri, J. 2017. "Social Media Usage and Shopping
Preferences: An Empirical Investigation," Proceedings of the 50th Hawaii
International Conference on System Sciences.
VuMA. 2019. "Anteile Von Prepaid- Und Vertragskunden Im Mobilfunk in
Deutschland Nach Alter 2019." Retrieved 19.3.2020, from
https://de.statista.com/statistik/daten/studie/154166/umfrage/vertrags--und-
prepaid-mobilfunkkunden-nach-alter-in-deutschland/
Walden, E. A., and Browne, G. J. 2009. "Sequential Adoption Theory: A Theory for
Understanding Herding Behavior in Early Adoption of Novel Technologies,"
Journal of the Association for Information Systems (10:1), pp. 31-62.
Walsham, G. 1993. "Interpreting Information Systems in Organizations." John Wiley
& Sons, Inc.
Weathers, D., Sharma, S., and Wood, S. L. 2007. "Effects of Online Communication
Practices on Consumer Perceptions of Performance Uncertainty for Search
and Experience Goods," Journal of Retailing (83:4), pp. 393-401.
Webb, E. J., Campbell, D. T., Schwartz, R. D., and Lee, S. 1966. Unobtrusive
Measures: Nonreactive Research in the Social Sciences. Chicago: IL: Rand
McNally.
Webb, E. J., Campbell, D. T., Schwartz, R. D., and Lee, S. 2000. Unobtrusive
Measures. Thousand Oaks: CA: SAGE Publications, Inc.
Webster, J., and Watson, R. T. 2002. "Analyzing the Past to Prepare for the Future:
Writing a Literature Review," MIS Quarterly (26:2), pp. xiii-xxiii.
Wei, L. 2009. "Filter Blogs Vs. Personal Journals: Understanding the Knowledge
Production Gap on the Internet," Journal of computer-mediated
communication (14:3), pp. 532-558.
246
References
__________________________________________________________________________
Welker, M., Taddicken, M., Schmidt, J.-H., and Jackob, N. 2014. Handbuch Online-
Forschung.
Sozialwissenschaftliche Datengewinnung Und -Auswertung in Digitalen Netzen, (12

ed.). Köln: Herbert von Halem Verlag.
Welker, M., and Wünsch, C. 2010. "Methoden Der Online-Forschung," in Handbuch
Online-Kommunikation, W. Schweiger and K. Beck (eds.). Wiesbaden: VS
Verlag für Sozialwissenschaften, pp. 487-517.
Whiting, A., and Williams, D. 2013. "Why People Use Social Media: A Uses and
Gratifications Approach," Qualitative Market Research: An International
Journal).
Wilcoxon, F. 1945. "Individual Comparisons by Ranking Methods. Biometrics Bulletin,
1 (6), 80-83".
Wilde, T., and Hess, T. 2007. "Forschungsmethoden Der Wirtschaftsinformatik,"
Wirtschaftsinformatik (49:4), pp. 280-287.
Wilson, K., Fornasier, S., and White, K. M. 2010. "Psychological Predictors of Young
Adults' Use of Social Networking Sites," Cyberpsychology, behavior, and
social networking (13:2), pp. 173-177.
Wirtz, B. W., and Göttel, V. 2016. "Technology Acceptance in Social Media: Review,
Synthesis and Directions for Future Empirical Research," Journal of Electronic
Witt, E. A., Massman, A. J., and Jackson, L. A. 2011. "Trends in Youth’s Videogame
Playing, Overall Computer Use, and Communication Technology Use: The
Impact of Self-Esteem and the Big Five Personality Factors," Computers in
Xia, L. 2013. "Effects of Companies' Responses to Consumer Criticism in Social
Media," International Journal of Electronic Commerce (17:4), pp. 73-100.
Xiao, B., and Benbasat, I. 2011. "Product-Related Deception in E-Commerce: A
Theoretical Perspective," Mis Quarterly (35:1), pp. 169-196.
Xu, S. X., and Zhang, X. 2013. "Impact of Wikipedia on Market Information
Environment: Evidence on Management Disclosure and Investor Reaction,"
Mis Quarterly), pp. 1043-1068.
Yi-Fen, C., and Ya-Ju, W. 2010. "Effect of Herd Cues and Product Involvement on
Bidder Online Choices," CyberPsychology, Behavior & Social Networking
(13:4), pp. 423-428.
247
References
__________________________________________________________________________
Zaichkowsky, J. L. 1985. "Measuring the Involvement Construct," Journal of

consumer research (12:3), pp. 341-352.
Zeelenberg, M., and Pieters, R. 1999. "Comparing Service Delivery to What Might
Have Been: Behavioral Responses to Regret and Disappointment," Journal of
Service Research (2:1), pp. 86-97.
Zeelenberg, M., and Pieters, R. 2007. "A Theory of Regret Regulation 1.0," Journal of
Consumer psychology (17:1), pp. 3-18.
Zhang, J. 2010. "The Sound of Silence: Observational Learning in the Us Kidney
Market," Marketing Science (29:2), pp. 315-335.
Zhang, J., and Liu, P. 2012. "Rational Herding in Microloan Markets," Management
science (58:5), pp. 892-912.
Zhang, J., Liu, Y., and Chen, Y. 2015. "Social Learning in Networks of Friends
Versus Strangers," Marketing Science (34:4), pp. 573-589.
Zhang, K. Z. K., Zhao, S. J., Cheung, C. M. K., and Lee, M. K. O. 2014. "Examining
the Influence of Online Reviews on Consumers' Decision-Making: A Heuristic–
Systematic Model," Decision Support Systems (67), pp. 78-89.
Zhang, W., and Watts, S. A. 2008. "Capitalizing on Content: Information Adoption in
Two Online Communities," Journal of the Association for Information Systems
(9:2), pp. 72-93.
Zhou, T. 2011. "Understanding Online Community User Participation: A Social
Influence Perspective," Internet Research (21:1), pp. 67-81.
Zhou, W., and Duan, W. 2016. "Do Professional Reviews Affect Online User Choices
through User Reviews? An Empirical Study," Journal of Management
Information Systems (33:1), pp. 202-228.
Zou, H., Sun, H., and Fang, Y. 2015. "Understanding Post-Adoption Regret from
the Perspectives of Herding and Mindfulness," 2015 36th International

Conference on Information Systems, Fort Worth.
248
Appendix A: Supplement to Chapter 3
__________________________________________________________________________

A.1: Introduction to the IS research methods (section 3.1.2)
The methods illustrated in Figure 16 (Method Profile on WI in the years 2007 - 2012)
are briefly described below based on the description of Schreiner et al. (2015, pp. 5-
6).
Method Description
Formal/Conceptual/Argumentative Logical–deductive reasoning as a research method

Deductive can take place at various levels of formalisation: either
within the framework of mathematical–formal models,
in semi-formal models (conceptual) or purely linguisti-
cally (argumentative).
Simulation The simulation formally depicts the behaviour of the

system under investigation in a model and simulates
environmental conditions by means of certain assign-
ments of the model parameters. Knowledge can be
gained both through model construction and through
the observation of endogenous model parameters.
Reference Modelling Reference modelling creates inductively (based on ob-

servations) or deductively (e.g. from theories or mod-
els) mostly simplified and optimised images (ideal con-
cepts) of systems in order to deepen existing knowl-
edge and to generate design templates.
Action Research A practical problem is solved by a mixed circle of sci-

ence and practice. Several cycles of analysis, action
and evaluation steps are carried out, each of which
provides for low-structured instruments such as group
discussions or planning games.
Prototyping A preliminary version of an application system is de-

veloped and evaluated. Both steps can generate new
insights.
Case Study The case study usually examines complex, difficult-to-
249
__________________________________________________________________________
Method Description
define phenomena in their natural context. It repre-

sents a special form of qualitative-empirical methodol-
ogy that intensively examines a few feature holders.
The focus is either on the most objective possible in-
vestigation of theses (behavioural–scientific approach)
or the interpretation of behavioural patterns as pheno-
types of the realities constructed by the test persons
(construction-oriented approach).
Grounded Theory Grounded theory ("object-anchored theory formation")

aims at the inductive acquisition of new theories
through intensive observation of the object of investiga-
tion in the field. The various procedures for coding and
evaluating the predominantly qualitative data are speci-
fied exactly.
Quantitative and Qualitative These two methods combine survey techniques such
Cross-Sectional Analyses as questionnaires, interviews, Delphi method, content
analysis etc. into two aggregates. They comprise a
one-time survey across several individuals, which is
then coded and evaluated quantitatively or qualita-
tively. The result is a cross-sectional picture across the
sample participants, which usually allows conclusions
to be drawn about the population.
Laboratory and Field Experiments The experiment investigates causal relationships in a

controlled environment by manipulating an experimen-
tal variable in a repeatable manner and measuring the
effect of the manipulation. The object of investigation is
examined either in its natural environment (in the 'field')
or in an artificial environment (in the 'laboratory'), which
significantly influences the possibilities of environ-
mental control.
250
Appendix B: Supplement to Chapter 4
__________________________________________________________________________

B.1: Tabular overview (concept matrix) of literature set identified (chapter 4)
B.2: Theoretical lenses of literature set identified (section 4.2.3)
251
____________________________________________________________________________________________________________________
B.1: Tabular overview (concept matrix) of literature set identified (Chapter 4)
Key Topics
Aspects affecting SM
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Author Research ogy Theoreti-
(year of publication) Title Source Stream (primary) cal lens
Huang, Jen-Hung Herding in online prod- Psychology Psychology Experiment Information x x x x

Chen, Yi-Fen uct choice & Market- cascade
(2008) ing theory
Yi-Fen, Chen Herd behaviour in pur- Computers Psychology Experiment Social in- x x x
(2008) chasing books online in Human fluence
Behaviour theory
Christy M.K. The impact of elec- Internet Information Survey Dual- x x

Cheung, Matthew tronic word-of-mouth: Research & Knowl- process
K.O. Lee, The doption of online edge Man- theory
Neil Rabjohn opinions in online cus- agement (Elabora-
(2008) tomer communities tion likeli-
hood
model)
252
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Walden, Eric A. Sequential Adoption JAIS Information Literature Information x x

Browne, Glenn J. Theory: A Theory for Systems review cascade
(2009) Understanding Herding theory
Behaviour in Early
Adoption of Novel
Technologies
Duan, Wenjing Informational cascades MISQ Information Quantita- Information x x x

Gu, Bin and software adoption Systems tive analy- cascade
Whinston, An- on the internet: an em- sis theory
drew B. pirical investigation
(2009)
Mira, Lee Electronic word of Interna- Marketing Survey Attribution x x x

Seounmi, Youn mouth (eWOM) tional Jour- Experiment theory
(2009) nal of
Advertising
253
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Mudambi, Susan What makes a helpful MISQ Information Quantita- Information x x

M. Schuff, David online review? A study Systems tive analy- economics
(2010) of customer reviews on sis theory
Amazon.com
Makoto, Naka- WOM or eWOM or AMCIS Information Survey Attribution x x x x

yama Yun, Wan Something Else: How Systems theory
Norma, Sutcliffe Does the Web Affect
(2010) Our Dependence on
Shopping Information
Sources?
Chen, Y., Online social interac- Journal of Marketing Experiment Social x x x

Wang, Q. tions: A natural experi- Marketing learning
Xie, J. ment on word of mouth Research theory; In-
(2011) versus formation
observational learning cascade th.
254
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Kee-Young, The Effects of Social HICSS Information Survey Social in- x x

Kwahk Media on E-commerce: Systems fluence
Xi, Ge A Perspective of Social theory/
(2012) Impact Theory Social im-
pact theory
Christy M.K., The Impact of Observa- HICSS Information Quantita- Information x x

Cheung tional Learning and Systems tive analy- processing
Bo, Xiao Electronic Word of sis theory
Ivy L.B., Liu Mouth
(2012) on Consumer Purchase
Decisions: The Moder-
ating Role of Consumer
Expertise and
Consumer Involvement
255
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Salazar Helen, Social influence on sus- Interna- Economics Survey Social x x

Arce Oerlemans, tainable consumption: tional Jour- Experiment learning
Leon van evidence from a behav- nal of Con- theory
Stroe‐Biezen, ioural experiment sumer Herding
Saskia (2013) Studies theory
Heshan, Sun A longitudinal study of MISQ Information Experiment - x x x x

(2013) herd behavior in the Systems
adoption and continued
use of technology
Lukas P., Forbes Does Social Media In- Journal of Economics Survey - x x
Eve M., Vespoli fluence Consumer Business &
(2013) Buying Behavior? An Economics
Investigation Of Rec- Research
ommendations And
Purchases
256
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Kuan, Kevin K. Y. Informational and Nor- JMIS Information Experiment Social in- x x
Zhong, Yingqin mative Social Influence Systems fluence
Chau, Patrick in Group-Buying: Evi- theory
Y.K. dence from Self-
(2014) Reported and EEG
Data
Godinho de Ma- Peer influence in the MISQ Information Mathemati- Social in- x x
tos, Miguel; diffusion of iPhone 3G Systems cal model- fluence
Ferreira, Pedro; over a large social net- ling theory
Krackhardt, David work
(2014)
257
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Jabr, Wael Know yourself and MISQ Information Mathemati- Prospect x x

Zheng, hiqiang know your enemy: An Systems cal model- theory
(2014) analysis of firm recom- ling
mendations and con-
sumer reviews in a
competitive environ-
ment
Brinja, Meiseberg The Effectiveness of ICIS Information Quantita- Long tail x x

(2014) E-tailers’ Communica- Systems tive analy- theory
tion Practices in Stimu- sis
lating Sales of Slow-
Selling versus Best-
Selling Products
258
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Purnawirawan, Expert reviewers be- Journal of Economics Experiment - x x x

Nathalia Dens, ware! The effects of re- Electronic
Nathalie view set balance, re- Commerce
De Pelsmacker, view source, and re- Research
Patrick view content on con-
(2014) sumer responses to
online reviews
Xiao-Liang, Shen Understanding Informa- HICSS Information Survey Adoption x x

Kem Z.K., Zhang tion Adoption in Online Systems theory
Sesia J., Zhao Review Communities: Herding
(2014) The Role of Herd Fac- theory
tors
Kem Z.K., Zhang Examining the influ- Decision Information Survey Dual- x x x
Sesia J., Zhao ence of online reviews Support Systems process
Christy M.K., on consumers' deci- Systems theory
259
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Cheung sion-making: A heuris- (Heuristic–

Matthew K.O., tic–systematic model systematic
Lee model)
(2014)
Haiyun, Zou Understanding Post- ICIS Information Survey Herding x

Heshan, Sun Adoption Regret from Systems Conceptual theory
Yulin, Fang the Perspectives of
(2015) Herding and
Mindfulness
Matook, Sabine Forming an intention to European Information Survey - x x

Brown, Susan A. act on recommenda- Journal of Systems
Rolf, Johanna tions given via online Information
(2015) social networks Systems
260
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Baek, Hyunmi Normative social influ- Journal of Economics Quantita- Social in- x x x
Lee, Saerom ence and online review Electronic tive analy- fluence
Oh, Sehwan helpfulness: Polynomial Commerce sis theory
Ahn, JoongHo modeling and response Research Mathemati-
(2015) surface analysis cal model-
ling
Tsao, Wen-chin eWOM persuasive- Electronic Economics Experiment Social in- x x x x

Hsieh, Ming- ness: do eWOM plat- Commerce fluence
tsang forms and product type Research theory
(2015) matter?
Rodney Graeme, Facebook advertising’s Internet Information Survey - x x

Duffett influence on intention- Research & Knowl-
(2015) to-purchase and pur- edge Mgmt.
chase amongst Millen-
nials
261
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Lee, Young-Jin Do I Follow My Friends Manage- Economics Mathemati- Information x x

Hosanagar, Kartik or the Crowd? Informa- ment Sci- cal model- cascade
Tan, Yong tion Cascades in Online ence ling theory
(2015) Movie Ratings
Jurui, Zhang; Social Learning in Net- Marketing Marketing Mathemati- Social x x

Yong, Liu; Yubo works of Friends versus Science cal model- learning
Chen Strangers ling theory
(2015)
Im, Il Deal-seeking versus MISQ Information Mathemati- - x x

Jun, Jongkun brand-seeking: search Systems cal model-
Oh, Wonseok behaviors and pur- ling
Jeong, Seok-Oh chase propensities in
(2016) sponsored search plat-
forms
262
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Xuan, Tan Herding Behavior in AMCIS Information Multiple Herding x

Karlene, Cousins Social Media Networks Systems case study theory
(2016) in China
Stefan, Hirsch- Information quality ECIS Information Survey - x x

meier needs throughout the Systems
Roman, Tilly purchase process
Detlef, Schoder
(2016)
Matute, Jorge The influence of Online In- Information Survey - x x x

Polo-Redondo, EWOM characteristics formation & Knowl-
Yolanda Utrillas, on online repurchase Review edge Man-
Ana intention agement
(2016)
263
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Liu, Qihua The influence of infor- Electronic Economics Mathemati- Information x x x

Huang, Shan mation cascades on Commerce cal model- cascade
Zhang, Liyi online purchase behav- Research ling theory
(2016) iors of search and ex-
perience products
Wenqi, Zhou Do Professional Re- Journal of Information Quantita- - x x x

Wenjing, Duan views Affect Online Manage- Systems tive analy-
(2016) User Choices Through ment IS sis; Mathe-
User Reviews? An Em- matical
pirical Study odel
Dewan, Sanjeev Popularity or Proximity: ISR Information Experiment Social x x x

Ho, Yi-Jen (Ian) Characterizing the Na- Systems learning
Ramaprasad, Jui ture of Social Influence theory
(2017) in an Online Music
Community
264
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Liu, Qianqian Ben The dark side of re- MISQ Information Experiment - x x x x
Karahanna, Elena views: The swaying ef- Systems
(2017) fects of online product
reviews on attribute
preference construction
Shih-Lun, Tseng The Effect of Herding AMCIS Information Experiment Information x x x

Shuya, Lu Behavior on Online Re- Systems Mathemati- cascade
Varun, Grover view Voting Participa- cal model- theory
Danny, Weathers tion ling
(2017)
Joseph, Vitha- Social Media Usage HICSS Information Survey Media rich- x x
yathil and Shopping Prefer- Systems ness theory
Majid, Dadgar ences: an Empirical In- Social net-
John, Osiri vestigation work theory
(2017)
265
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Liu, Yang When Online Reviews Information Information Mathemati- - x x x

Feng, Juan Meet Sales Volume In- Systems Systems cal model-
Liao, Xiuwu formation: Is More or Research ling
(2017) Accurate Information
Always Better?
Kwahk, Kee- Effects of social media Service Economics Survey Social in- x x
young on consumers’ pur- Business fluence
Kim, Byoungsoo chase decisions: evi- theory
(2017) dence from Taobao
Xitong, Li Herding and Social MISQ Information Quantita- Herding x x

Lynn, Wu Media Word-of-Mouth: Systems tive analy- theory
(2018) Evidence from Groupon sis
266
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Peter Gordon, Information overload in Business Economics Literature - x

Roetzel the information age: a Research review
(2018) review of the literature
from business admini-
stration, business psy-
chology, and related
disciplines with a bibli-
ometric approach and
framework develop-
ment
Amy Wenxuan, Herding in the con- Journal of Marketing Mathemati- Herding x

Ding sumption and purchase the Acad- cal model- theory
Li, Shibo of digital goods and emy of ling
(2018) moderators of the herd- Marketing
ing bias Science
267
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Li, Qi, Does friendship quality Electronic Economics Experiment - x x x

Liang, Ni, matter in social com- Commerce
Li, Eldon Y. merce? An experimen- Research
(2018) tal study of its effect on
purchase intention
Kunst, Katrine Understanding elec- Electronic Information Conceptual - x x

Vatrapu, Ravi tronic word of behavior: Markets Systems
(2018) conceptualization of the
observable digital
traces of consumers’
behaviors
Shen, Yongchao Influence of aggregated European Marketing Experiment - x x

Wei, Shan; Luan, ratings on purchase Journal of
Jing decisions: an event- Marketing
(2018) related potential study
268
____________________________________________________________________________________________________________________
Key Topics
reg. Herd
impact
Behaviour
Technology
Information
Information
Information
Popularity
Homohily
Adoption
Methodol-
Valence
Product
Source
Quality
Others
Type
Type
Su Jung, Kim Understanding the ef- Interna- Marketing Quantita- Dual- x x x

Ewa, Maslowska fects of different review tional Jour- tive analy- process
Edward C., features on purchase nal of Ad- sis theory
Malthouse probability vertising (Elabora-
(2018) tion likeli-
hood
model)
Ismail, Erkan Social media or shop- Journal of Marketing Survey Adoption x x x

Chris, Evans ping websites? The in- Marketing theory
(2018) fluence of eWOM on Communi-
consumers’ online pur- cations
chase intentions
269
__________________________________________________________________________
B.2: Theoretical lenses of literature set identified (section 4.2.3)
The literature analysis presented in Chapter 4 has shown that researchers in the field
of herd behaviour in buying decisions use a variety of theoretical lenses to frame
their research. These are briefly explained below and serve as a supplement to Fig-
ure 22 (Applied theoretical lenses of literature set investigated) in section 4.2.3. The
'herding theory' is not explained, as it is explained in section 2.1. Most of the descrip-
tion, originating area and level of analysis were taken from (Larsen 2015).
Information cascade theory
Originating area Behavioural Economics
Level of analysis Individual
Description Information cascade theory describes the conditions which lead to the
mechanism of herd behaviour; the presence of uncertainty about a deci-
sion, and the observed actions of other individuals (Banerjee 1992;
Bikhchandani et al. 1992). In behavioural economics and network the-
ory, the informational cascade describes phenomenon in which a num-
ber of people make the same decision one after another (Duan et al.
2009). An information cascade is considered a two-stage process. First,
an individual must face a decision, typically a binary decision. Second,
external factors can influence this decision (usually by observing the ac-
tions of others in similar scenarios (Bikhchandani et al. 1992).
Social influence theory
Originating area Psychology
Description In 1958 psychologist Herbert Kelman proposed in the central theme of

social influence theory that an individual’s attitudes, beliefs, and subse-
quent actions or behaviours are influenced by referent others. Thus
people might also alter their attitudes and behaviours in response to
what they perceive others might do or think. Kelman (1958) points out
270
__________________________________________________________________________
three varieties of social influence: compliance, identification, and inter-

nalisation.
 It is assumed that compliance occurs when individuals accept

influence and adopt their behaviour in order to avoid punish-
ment and gain rewards at the same time. For this reason, “the
satisfaction derived from compliance is due to the social effect
of accepting influence” (Kelman 1958, p. 53).
 The variety of identification is meant when people/ individuals
get influenced by someone else or rather adopt their behav-
iour to establish a desired relationship to another group or
person. Thus, satisfaction occurs due to “the act of conform-
ing” (Kelman 1958, p. 53).
 Internalisation is supposed to take place when individuals ac-
cept influence after they perceive the content of the induced
behaviour is rewarding. It also helps people adopt the behav-
iour when they realize that it is congruent with their own value
system. In this case, the satisfaction occurs due to “the con-
tent of the new behaviour” (Kelman 1958, p. 53).
Social learning theory
Originating area Psychology & Criminology
Level of analysis Individual, Group
Description Learning theories try to point out the way of thinking of individuals and
the factors which determine their behaviour. Social learning theories are
learning theories based on the belief that the way humans act (human
behaviour) is ascertained by a combination of three aspects: cognitive
factors, environmental influences and behaviour. The theory is based on
the work of Gabriel Trade (1843–1904), who argues that social learning
arises out of four main stages: close contact, imitation of superiors, un-
derstanding of concepts and role model behaviour (Bandura 1969).
271
__________________________________________________________________________
Dual-process theory
Level of analysis n.a.
Description The dual-process theory in psychology is based on the fact that

thoughts can arise in two different ways or as the result of two different
processes. The two processes can consist of an implicit (automatic)
unconscious process and an explicit (controlled) conscious process.
Conscious processes and actions can change with persuasion or edu-
cation, but implicit processes or attitudes usually take a long time to
change with the formation of new habits (Chaiken and Trope 1999). In
addition to psychology, dual-process theory has also been linked to
economics via prospective theory and behavioural economics, and in-
creasingly to economics via cultural analysis in sociology (Lizardo et al.
2016).
Attribution theory
Originating area Social psychology
Description Attribution theories state how individuals evaluate and determine the
cause of other people's behaviour. Hence it explains the observed attri-
bution processes, which human beings use to understand why an event
or behaviour occurred (Kelley and Michela 1980). There are two main
types of attribution: internal and external attribution. Internal attribution
refers to the interpretation of a person's behaviour caused by internal
characteristics such as altruism, motivation, etc. External attribution re-
fers to the interpretation of behaviour caused by external characteristics
such as compliance, cohesion, etc (Larsen 2015).
272
__________________________________________________________________________
Adoption theory
Description Adoption theory deals with the decisions an individual makes to accept
or reject a particular innovation. In some models, adoption also de-
scribes the extent to which this innovation is integrated into the relevant
context. Adoption theory provides a micro-perspective of change, i.e. it
does not focus on the whole, but rather on the parts that make up the
whole. The results of adoption theory are measured in terms of behav-
ioural changes; the signs of these behavioural changes can be under-
stood through contextual, cognitive and affective factors. (Straub 2009)
The following six theories are grouped into the category 'others' in section 4.2.3:
'information economics theory', 'information processing theory', 'long tail theory', 'me-
dia richness theory', 'prospect theory', and 'social network theory'.
The following theories are not outlined below as they are explained in previous sec-
tions: 'information processing theory' in section 2.2.2, and 'media richness theory' in
section 2.3.3
Information economics theory
Originating area Microeconomics
Level of analysis Micro-, macro-level, individual level
Description Information economics is formally related to game theory, and two dif-
ferent types of games can be applied: games with complete information
and games with incomplete information. Experimental and game theory
methods have been developed to model and test the theories of the in-
formation economy, including potential applications in public policy (Bal-
tes and Smelser 2004).
273
__________________________________________________________________________
Long tail theory
Originating area n.a.
Level of analysis n.a.
Description The long tail theory was developed by Anderson (2006), according to his
theory in the Internet age, mass markets will lose importance compared
to niche markets. This is due to the possibility of being able to offer
products with no or comparatively low costs and to the fact that these
products can be found by potential customers. Rarely sold or very exotic
products, whose trade is not worthwhile for the conventional retail shop,
find their marketplace on the Internet.
Prospect theory
Level of analysis Individual, Group, Organisation, Inter-organisational
Description Kahneman and Tversky (1979) developed the prospect theory, which
describes how individuals assess their loss and gain perspectives. One
of the most fundamental principles is that individuals could seek risks if
they perceive the current outlook as negative, and individuals could
avoid risks if they perceive their current situation or outlook as positive.
Thereby, risky decisions (prospects) are segregated into two phases:
editing (defining gains and losses in relation to a reference point) and
evaluation of prospects (Larsen 2015).
274
__________________________________________________________________________
Social network theory
Originating area Social psychology, social anthropology, mathematical sociology, psy-

chometrics
Level of analysis Individual, group, network
Description The social network theory analyses different social relationships in terms
of nodes and ties. Accordingly, individual actors within the networks are
so-called nodes, while ties do describe the relationships between the ac-
tors. In general, there can be various types of ties between the nodes.
Looking at all the connections between the nodes with several ties, a
social network can be seen as a map of linkages. The theory describes
how several people, organisations or groups interact with others inside
their network. However, in theory, the characteristics of individuals are
less important than their relationships (ties) to others. (Larsen 2015)
275
Appendix C: Supplement to Chapter 5
__________________________________________________________________________

C.1: Observation protocol (section 5.3.2)
C.2: Coding guide (section 5.3.3)
276
__________________________________________________________________________
C.1: Observation protocol (section 5.3.2)
The following observation protocol was used to take notes and describe the three so-
cial media platforms form the study. The observation protocol was translated into
English by the author and formatted for its presentation in the Appendix.
Part I - General topics
 Name of social media

 Observation period
 Type of collected data
 Number of users
 Description of conversations
 Main target of the community
 Character of contributions
Part II - Special features
 Description of community
o entry rules
o rules of treatment
 Functionalities
o chat
o own postings
o search function/archive
o individualisation
 Others
o privacy protection
o user-friendliness
277
____________________________________________________________________________________________________________________
C.2: Coding guide (section 5.3.3)
The following coding guide represents an excerpt (top five categories) from the complete coding catalogue which was used for coding
the non-reactive data of the three social media platforms. The codes and descriptions were translated into English by the author and
formatted for its presentation in the Appendix.
Main Category Subcategory Description/Definition Anchor Example Coding Rule
Satisfaction Mobile network This category describes the "Many of my friends also have o2 and are Code, if satisfaction with
(MC 1) (SC 1.1) satisfaction with the mobile quite happy with it. Yes, the net has gaps network is mentioned.
network. now and then, but I can live with that".
Satisfaction Customer ser- This category describes the "The service is good. Fast processing etc. I Code, if satisfaction with
(MC 1) vice satisfaction with the customer am fully satisfied, the whole communication, customer service is men-
(SC 1.2) service. also with the chat — a great thing! We are to- tioned.
tally enthusiastic!!!"
Satisfaction Offer This category describes the " Offer was the best it was on the market at Code, if satisfaction with
(MC 1) (SC 1.3) satisfaction with the prod- the time. I am very satisfied so far". offer is mentioned.
uct/offer.
Satisfaction Mobile phone This category describes the "Hello! My dream phone is the S6...Active! Code, if satisfaction with
(MC 1) (SC 1.4) satisfaction with a mobile de- Because it is waterproof and has a bigger mobile device is men-
vice (smartphone). battery, for me as a sportsman this is very tioned.
very important! Apart from that I love the S6.
278
____________________________________________________________________________________________________________________
The screen, the camera. everything great".
Satisfaction Provider This category describes the "I am satisfied with o2 and have never had Code, if satisfaction with
(MC 1) (SC 1.5) satisfaction with a mobile any problems. I have always been offered mobile provider is men-
provider. reasonable deals". tioned.
Dissatisfaction Mobile network This category describes the "No network availability within buildings. Code, if dissatisfaction
(MC 2) (SC 2.1) dissatisfaction with the mo- Once you are indoors, reception is no longer with mobile network is
bile network. usable. In the middle of a conversation, it mentioned.
breaks down because there is no network
available. As soon as you are in the fresh air,
the reception is fully available. I will cancel
my contract immediately and never order or
sign anything with this provider again!"
Dissatisfaction Customer ser- This category describes the "Employees usually have no idea, are com- Code, if dissatisfaction
(MC 2) vice dissatisfaction with the cus- pletely unqualified. Especially in the shops. with customer service is
(SC 2.2) tomer service. Appointments with O2 are often not kept and mentioned.
you can't reach anybody on the hotline for
ages".
Dissatisfaction Offer This category describes the "Instead of talking about roaming tariffs, you Code, if dissatisfaction
(MC 2) (SC 2.3) dissatisfaction with the prod- should first offer reasonable domestic tariffs, with offer is mentioned.
279
____________________________________________________________________________________________________________________
uct/offer. I'm totally dissatisfied with my second-class

tariff, but with the others there are no better
ones".
Dissatisfaction Mobile phone n/a n/a n/a

(MC 2) (SC 2.4)
Dissatisfaction Provider This category describes the "Is there actually anyone who is completely Code, if dissatisfaction
(MC 2) (SC 2.5) dissatisfaction with a mobile satisfied with o2? Well, I don't know anyone! I with mobile provider is
provider. have quit my contract and to my sorrow they mentioned.
ignore some things here, so that you have to
give notice which will cost you another year".
Comment and Mobile network This category describes "I haven't had coverage since yesterday and Code, if comments or
Opinion (SC 3.1) comments and opinions with not today either, otherwise I can't complain, opinions with regard to
(MC 3) regard to mobile network. fortunately we still have WLAN via Telekom". mobile network are men-
tioned.
Comment and Customer ser- n/a n/a n/a

Opinion vice
(MC 3) (SC 3.2)
Comment and Offer This category describes "I don't understand why you do LTE when it is Code, if comments or
280
____________________________________________________________________________________________________________________
Opinion (SC 3.3) comments and opinions with not yet widely available". opinions with regard
(MC 3) regard to product/offer. product/offer are men-
tioned.
Comment and Mobile phone This category describes "Most people are not interested in the con- Code, if comments or
Opinion (SC 3.4) comments and opinions with tract in the real sense, but only in the latest opinions with regard to
(MC 3) regard to mobile devices smartphones. Guys, if you can't afford a con- mobile devices (smart-
(smartphone). tract, then you don't need to show off with a phone) are mentioned.
class-A smartphone".
Comment and Provider This category describes "Is there a human sitting there or is that a Code, if comments or
Opinion (SC 3.5) comments and opinions with machine that comments against it? You bet- opinions with regard to a
(MC 3) regard to a mobile provider. ter invest the power into the net as soon as mobile provider are men-
possible, otherwise the people will run tioned.
through the holes in the net to the competi-
tion in no time!"
Support Mobile network This category describes of- "Please send us your request in a PN. We'll Code, if asked or offered
(asked/offered) (SC 4.1) fered or asked support with gladly take care of it". support with regard to
(MC 4) regard to mobile network. mobile network is men-
tioned.
Support Customer ser- This category describes of- "It's a pity that you are currently not com- Code, if asked or offered
281
____________________________________________________________________________________________________________________
(asked/offered) vice fered or asked support with pletely satisfied with our service. We would support with regard to
(MC 4) (SC 4.2) regard to customer service. be happy to check your request again more customer service is men-
closely. Please send us your request and tioned.
your data via PN to our service page".
Support Offer This category describes of- "Now I wanted to know whether someone Code, if asked or offered
(asked/offered) (SC 4.3) fered or asked support with could recommend a tariff to me or whether support with regard to a
(MC 4) regard to product/offer. someone would advise me directly against product/offer is men-
my plan". tioned.
Support Mobile phone This category describes of- "Is it worth buying an older model of the Code, if asked or offered
(asked/offered) (SC 4.4) fered or asked support with iPhone series? Which iPhone would that be support with regard to a
(MC 4) regard to a mobile device and which advantages and disadvantages mobile device (smart-
(smartphone). does this model have? phone) is mentioned.
thanks in advance :)".
Support Provider This category describes of- "Total rip-off! What's the best way to proceed Code, if asked or offered
(asked/offered) (SC 4.5) fered or asked support with now?" support with regard to a
(MC 4) regard to a mobile provider. mobile provider is men-
tioned.
Experience Mobile network This category describes ex- "I also hardly ever have a net and almost al- Code, if experiences with
periences with regard to mo- ways the 'E'. And I am on the road in the Co- regard to mobile network
282
____________________________________________________________________________________________________________________
(MC 5) (SC 5.1) bile network. logne/Bonn urban area. What a pity! With the are mentioned.
company mobile phone it seems to be a little
better."
Experience n/a n/a n/a n/a

(MC 5)
Experience Offer This category describes ex- "I've only just done that, got myself the new Code, if experiences with
(MC 5) (SC 5.3) periences with regard to Samsung, with a contract and a PlayStation regard to a product/offer
product/offer. 4, and I wanted to have both, so this bundle are mentioned.
offer came in pretty handy. Compared to buy-
ing them separately you save a bit".
Experience Mobile phone This category describes ex- "An acquaintance of mine had his smart- Code, if experiences with
(MC 5) (SC 5.4) periences with regard to a phone repaired at fixxoo and said that every- regard to a mobile device
mobile device (smartphone). thing worked fine there. From the first im- (smartphone) are men-
pression it seems to me also quite respect- tioned.
able, so try it".
Experience Provider This category describes ex- "If I read all the answers here, I can only con- Code, if experiences with
(MC 5) (SC 5.5) periences with regard to a firm poor network coverage with the opera- regard to a mobile pro-
mobile provider. tor". vider are mentioned.
283
Appendix D: Supplement to Chapter 6
__________________________________________________________________________

D.1: Interview guide of expert interviews (section 6.3.2)
D.2: List of interviews (section 6.3.2)
D.3: Coding guide (section 6.3.3)
284
__________________________________________________________________________
D.1: Interview guide of expert interviews (section 6.3.2)
The following interview guide was used for the expert interviews. Depending on the
function and background of the individual interviewee, questions were deepened,
modified and/or dropped. The questionnaire was translated into English by the author
and formatted for its presentation in the Appendix.
Introductory notes to the survey
• The conversation is recorded and then transcribed. Do you agree with this?
(yes/no)
• Please do not hesitate to ask if anything is unclear.
• If you do not want to/cannot answer questions, just let me know.
• May I contact you after the interview for individual questions? (yes/no; mail/phone)
Remarks on the questionnaire
• The questionnaire is used as rough guideline for the interview (devia-

tions/prioritisation are possible).
• An interview preparation of the interviews participants is unnecessary.
Interviewee
• Name:
• Position in the company:
• Company/Sector:
• Day, Date:
• Duration:
Notes:
Part I - Introduction
1.1. Presentation, framework and objectives of the interview (brief information on the in-
terview background)
1.2. Introduction of the interviewee
285
__________________________________________________________________________
Part II - General
2.1 How would you describe the influence of /social media/online communities on the
mobile market in the last years?
2.2 People tend to follow the opinion of others — is this effect in the mobile market rein-
forced by social media? If so, to what extent have you already observed this?
2.3 What opportunities and risks do you see for the mobile industry through /social me-
dia/online communities?
2.4 How personal interest in new digital media and technologies is related to the influ-
ence of /social media/online communities on mobile product decisions?
Part III - Influencing factors on herd behaviour
3.1. How do you see the impact of /social media/online communities on herd behaviour?
(Do people make different decisions as a result?)
3.2. How do /social media/online communities influence our decision- or buying behav-
iour? (Does our decision behaviour change to the effect that less individual deci-
sions are made?)
3.3. In which phase of the buying decision process do /social media/online communities
have the highest influence?
3.4. How do you evaluate the connection between existing use or buying intention and
the influence of /social media/online communities?
3.5. Does a higher popularity of the social media/online community have a stronger in-
fluence on herd behaviour?
3.6. What significance do the number of likes or the number of comments have on herd
behaviour? Number of members?
286
__________________________________________________________________________
Part IV - Reasons for the influence of social media
4.1. What are the main reasons for the influence of /social media/online communities?
(What are the most significant reasons?)
4.2. How do you value the importance of emotions 'soft facts' (e.g. dissatisfaction,
praise, criticism) compared to 'hard facts' (e.g. product comparisons, reviews) in
/social media/online communities?
4.3. How do you assess the connection between a high level of activity (writing content;
high use frequency) in /social media/online communities and herd behaviour?
4.4. How do you assess the connection between a low level of activity (passive informa-
tion consumption; low use frequency) in /social media/online communities and herd
behaviour?
4.5. How do you assess the effects of active vs. passive using of /social media/online
communities on herd behaviour?
4.6. What connection do you see between low and passive using of online communities
and herd behaviour?
Part V - Social Media vs. other sources of influence
5.1 What other influencing factors are important on herd behaviour besides /social me-
dia/online communities?
5.2 What influence have /social media/online communities on herd behaviour compared
to the influence of the private surroundings (family, friends, acquaintances)?
5.3 In the case of a buying decision, what influence is decisive and why? (e.g. mobile
phone recommended by friends and afterwards a comment seen at FB vs. vice
versa)
5.4 Are people more likely to be influenced by online or offline advice (e.g. expert online
vs. expert offline) and what does this depend on?
5.5 How do you rate the importance of the community type (Facebook vs. expert forum
vs. rating pages) in terms to the strengths of influence on herd behaviour?
5.6 To what extent does it make a difference on herd behaviour whether members of an
/social media/online communities have a profile or not?
287
__________________________________________________________________________
Part VI - Mobile communications specifics
6.1 What do you see as the main factors influencing mobile communication decisions?
6.2 On which mobile communications decisions do /social media/online communities

have the greatest influence and on which ones the least?
6.2 How do you see the influence of /social media/online communities of:
 Image
 Price decision
 Recommendation
 New contract
 VVL
 New services
 Mobile phone
6.4 Which community type (Facebook vs. expert forum vs. rating pages) has the
strongest influence on which mobile communications decision?
6.5 How do you see the influence of the private environment compared to /social me-
dia/online communities on:
 Image
 Price decision
 Recommendation
 New contract
 VVL
 New services
 Mobile phone
6.6 On which mobile communication decisions does the private environment have the
strongest influence?
6.7 Optional: depending on question 5.1: On which mobile communication decision

does (input from question 5.1) have the strongest influence?
Additional information and conclusion
• Further questions, comments, suggestions

• On which contents would in-depth discussions be helpful in your opinion?
• Which documents could support the information from the interview?
288
__________________________________________________________________________
Appendix to the interview guide:
Example of herd behaviour based on an experiment: Decision for a restaurant

(Banerjee 1992).
100 people go in them two unknown restaurants A and B next to each other. The
probability which restaurant is better is almost equally distributed in advance. Arriving
in front of the restaurants, 99 people get the information that restaurant B is the bet-
ter one. One person gets the information that restaurant A is better, this person is al-
lowed to choose a restaurant first. This person goes to restaurant A. The one behind
has the choice to either rely on his own information and decide for restaurant B or to
follow the first person. This person decides to go to restaurant A as well. All other
persons behind join and end up in restaurant A.
289
__________________________________________________________________________
D.2: List of interviews (section 6.3.2)
No. Role Relevant Background Date Duration Type

(min.)
1 Freelance Social media expert June, 2016 68 Pre-test
2 Manager Strategy & Telecommunication ex- July, 2016 72 Pre-test

Innovation pert
3 Manager Online Customer behaviour, so- July, 2016 59 Personal

Sales & Marketing cial media, marketing interview
4 CEO Online targeting expert September, 55 Personal

2016 interview
5 Manager Social Me- Social media expert with October, 56 Personal

dia telecommunication 2016 interview
knowledge
6 Manager Business Analytical tools, customer November, 62 Personal

Analytics behaviour 2016 interview
7 Data Scientist Data, customer behaviour November, 57 Personal

2016 interview
8 Manager Online Customer behaviour, so- January, 56 Personal

Sales & Marketing cial media, marketing 2017 interview
290
____________________________________________________________________________________________________________________
D.3: Coding guide (section 6.3.3)
The following coding guide represents an excerpt (top five categories) from the complete coding catalogue which was used for coding
the content of the conducted expert interviews. The codes and descriptions were translated into English by the author and formatted
for its presentation in the Appendix.
Drivers of social Credibility This category describes credi- "So people use it especially if the content Code, if credibility as a
media usage (SC 1.1) bility as one driver of social is not complete nonsense now. And if the driver of social media us-
(MC 1) media usage. texts, evaluation and commentary also fit age is mentioned.
together and are good, then for me it is a
sign of credibility, of what has been ex-
pressed there".
Drivers of social Wisdom of This category describes the "There is a certain frequency in the opin- Code, if 'wisdom of
media usage crowds phenomenon 'wisdom of ion, where I then say: This is an opinion of crowds' as one driver for
(MC 1) (SC 1.2) crowds' as one driver for social several people, several people must have social media usage is
media usage. dealt with it". mentioned.
Drivers of social Convenience This category describes con- "And one then makes it easier for oneself Code, if convenience as
media usage (SC 1.3) venience as one driver for so- and, if it is not completely wrong, one one driver for social me-
(MC 1) cial media usage. agrees with this opinion. It saves time and dia usage is mentioned.
makes life easier".
291
____________________________________________________________________________________________________________________
Drivers of social Trustworthiness This category describes trust- "If there is a name or you even have to log Code, if trustworthiness
media usage (SC 1.4) worthiness as one driver for in before you can comment. There are as one driver for social
(MC 1) social media usage. simply the thresholds much higher that media usage is men-
you can say nonsense". tioned.
Drivers of social Independency This category describes inde- "If one wants an independent opinion be- Code, if independency as
media usage (SC 1.5) pendency as one driver for so- fore deciding on a mobile phone product, one driver for social me-
(MC 1) cial media usage. one simply goes online and looks at differ- dia usage is mentioned.
ent reviews. Then you have a good inde-
pendent cross-sectional opinion in a short
time".
Drivers of social Transparency This category describes trans- "Communities have had a very strong im- Code, if transparency as
media usage (SC 1.6) parency as one driver for so- pact on transparency and have thus in- one driver for social me-
(MC 1) cial media usage. creased their use". dia usage is mentioned.
Drivers of social Popularity of This category describes popu- "The more popular the platform, the more Code, if popularity of
media usage community larity of community as one often you will come across it and use it community as one driver
(MC 1) (SC 1.7) driver for social media usage. more often in case of doubt". for social media usage is
mentioned.
General influence Review/rating This category describes re- "I believe that more and more such rating Code, if review-/and rat-
groups platforms view/rating platforms as a sites have a very strong influence, being platforms as general
292
____________________________________________________________________________________________________________________
(MC 2) (SC 2.1) general influence group that cause ninety percent of decisions made by influence group are men-
may influence herd behaviour buyers are made through price compari- tioned.
in consumers' buying deci- son portals".
sions.
General influence Friends and fam- This category describes "The private environment, like your friends Code, if friends and fam-
groups ily friends and family as a general or family will still be crucial". ily as general influence
(MC 2) (SC 2.2) influence group that may influ- group are mentioned.
ence herd behaviour in con-
sumers' buying decisions.
General influence Search engine This category describes "The Google search slot. But even then it Code, if search engine as
groups (SC 2.3) search engine as a general in- influences the prioritization via search en- general influence group
(MC 2) fluence group that may influ- gine algorithms. And it flushes offers, de- is mentioned.
ence herd behaviour in con- pending on how well the algorithm has
sumers' buying decisions. been designed, more strongly upwards".
General influence Social networks This category describes social "Whereas on a platform I have, in the best Code, if social networks
groups (SC 2.4) network as a general influence case, a lot of returning people who ex- as general influence
(MC 2) group that may influence herd change and advise each other". group are mentioned.
behaviour in consumers' buy-
ing decisions.
293
____________________________________________________________________________________________________________________
General influence Independent test- This category describes inde- "Tests, such as Stiftung Warentest or oth- Code, if independent
groups ing foundation pendent testing foundation as ers, simply per se as an institution should testing foundation as
(MC 2) (SC 2.5) a general influence group that provide a neutral view on products". general influence group
may influence herd behaviour is mentioned.
in consumers' buying deci-
sions.
Product specifics Low-involvement This category describes low- "So decisions with a high impact on one's Code, if low- vs. high-
(MC 3) vs. high- vs. high-involvement products own life are not only based on online rec- involvement products as
involvement as product specifics that may ommendations or opinions. Differently e.g. product specifics are
products influence herd behaviour in with a vacation, there one can make once mentioned.
(SC 3.1) consumers' buying decisions. two weeks a bad experience, of it the
world does not go at the end of the day
however, annoys one perhaps over it".
Product specifics Hardware as This category describes hard- "So I believe that most people have a Code, if hardware as
(MC 3) high-involvement ware as high-involvement strong emotional connection to their high-involvement product
product products as a product specific smartphone and that the experiences of as product specifics is
(SC 3.2) that may influence herd behav- others have a strong impact on their own mentioned.
iour in consumers' buying de- decisions".
cisions.
294
____________________________________________________________________________________________________________________
Product specifics Connectivity as This category describes con- "Mobile telephony is simply a commodity Code, if connectivity as
(MC 3) low-involvement nectivity as low-involvement product for the end customer, it just has to low-involvement product
product product as a product specific be there and work, there is no longer high as product specifics is
(SC 3.3) that may influence herd behav- involvement in it. You simply make deci- mentioned.
iour in consumers' buying de- sions more or less independently".
cisions.
Role of social Information pro- This category describes infor- "I think communities are heavily used Code, if information pro-
media in buying vider in pre- mation provider in the pre- when you want to gather information first. vider in the pre-buying
decision process buying phase buying phase as the role of Or one wants to orientate oneself". phase as the role of so-
(MC 4) (SC 4.1) social media in the buying de- cial media in the buying
cision process. decision process is men-
tioned.
Role of social Confirmation in This category describes the "I am in the process and want to secure Code, if confirmation in
media in buying buying phase confirmation of decisions in the my decision once again". the buying phase as the
decision process (SC 4.2) buying phase as the role of role of social media in the
(MC 4) social media in the buying de- buying decision process
cision process. is mentioned.
Role of social Optimisation in This category describes the "If you are VVL-authorised and depending Code, if optimisation in
media in buying post-buying optimisation in the post-buying on the satisfaction, check again for price the post- buying phase
295
____________________________________________________________________________________________________________________
decision process phase phase as the role of social or any optimization possibilities". as the role of social me-
(MC 4) (SC 4.3) media in the buying decision dia in the buying decision
process. process is mentioned.
Customer seg- Technology affin- This category describes the "Some people are supertechis, early Code, if technology affin-
ment specifics ity technology affinity as a specific adopters who want to be cool". ity as a specific charac-
(MC 5) (SC 5.1) characteristic of a customer teristic of a customer
segment. segment is mentioned.
Customer seg- Price sensitivity This category describes price "I still think it depends on the target group, Code, if price sensitivity
ment specifics (SC 5.2) sensitivity as a specific charac- so I think it's really segment-specific. If I'm as a specific characteris-
(MC 5) teristic of a customer segment. price sensitive now, I'm going to look at tic of a customer seg-
price comparison sites". ment is mentioned.
Customer seg- Service affinity This category describes ser- "And service affinity of consumers can Code, if service affinity as
ment specifics (SC 5.3) vice affinity as a specific char- perhaps also have an influence on the a specific characteristic
(MC 5) acteristic of a customer seg- buying decision". of a customer segment is
ment. mentioned.
Customer seg- Culture This category describes cul- "There are also target groups with a dif- Code, if culture as a spe-
ment specifics (SC 5.4) ture as a specific characteristic ferent cultural background, such as people cific characteristic of a
(MC 5) of a customer segment. of Turkish origin, who make the mobile customer segment is
mentioned.
296
____________________________________________________________________________________________________________________
phone decision differently".
Customer seg- Age This category describes age "There are enough studies that show that Code, if age as a specific
ment specifics (SC 5.5) as a specific characteristic of a the age group plays a role in the buying characteristic of a cus-
(MC 5) customer segment. decision". tomer segment is men-
tioned.
297
Appendix E: Supplement to Chapter 7
__________________________________________________________________________

E.1: Questionnaire (section 7.4.2.1)
E.2: Screenshots Unipark (section 7.4.2.1)
298
__________________________________________________________________________
E.1: Questionnaire (section 7.4.2.1)
The following questionnaire was used for the online survey. The original survey was
conducted in German, the following questionnaire is translated into English by the
author and formatted for its presentation in the Appendix.
299
__________________________________________________________________________
Index
1 Welcome ............................................................................................................. 304
2 Socio-demographic Data ................................................................................... 304

2.1 Age ............................................................................................................. 304
2.2 Gender .......................................................................................................... 305
2.3 Household..................................................................................................... 305
2.3.1 Household Size .................................................................................... 305
2.3.2 Household Type ................................................................................... 305
2.4 Profession ..................................................................................................... 306
2.5 Educational Level ......................................................................................... 306
3 Segmentation Questions ................................................................................... 307

3.1 Personality .................................................................................................... 307
3.2 Interests ........................................................................................................ 307
3.3 Social Media Activity ..................................................................................... 308
3.3.1 Daily Social Media Content Creation ................................................... 308
3.3.2 Social Media Content Creation per Week ............................................ 309
3.3.3 Participation in Social Media per Day .................................................. 309
3.3.4 Participation in Social Media per Week ............................................... 309
3.3.5 Social Media Consumption per Day..................................................... 309
3.3.6 Social Media Consumption per Week .................................................. 309
4 Comparison Group Selection (Automated Classification) ............................. 310
5 Smartphone ........................................................................................................ 310

5.1 Smartphone - Group 1 (Control group) ......................................................... 310
5.1.1 Product Involvement ............................................................................ 310
5.1.2 Pre-Buying Phase - Smartphone ......................................................... 310
5.1.2.1 Main Information Source (1)..................................................... 310
5.1.2.2 Reasons for Main Information Source (1) ................................ 311
5.1.2.3 Second Most Important Information Source (2) ....................... 311
5.1.2.4 Reasons for Second Most Important Information Source (2) ... 312
5.1.3 Buying Phase - Smartphone ................................................................ 312
5.1.3.1 Smartphone - No External Information ..................................... 312
5.1.4 Post-Buying Phase - Smartphone........................................................ 313
5.1.4.1 "Love it or exchange it"- (At Neutral) ........................................ 313
5.1.4.2 "Love it or exchange it"- Decision (A preferred) ....................... 313
5.1.4.3 Love it or exchange it"- Decision (B preferred) ........................ 314
300
__________________________________________________________________________
5.2 Smartphone - Group 2 (FB Input in Buying Phase) ...................................... 314

5.2.2.5 Smartphone - Input Customer Rating ...................................... 317
5.2.2.6 Smartphone - Input Expert Review .......................................... 318
5.2.3.1 Smartphone - Input FB ............................................................. 318
5.2.4.1 "Love it or exchange it"- (B preferred; A recom. friends) .......... 319
5.2.4.2 "Love it or exchange it"- (A preferred; B recom. friends) .......... 319
5.3 Smartphone - Group 3 (Friends Input in Buying Phase) ............................... 320
5.3.2.4 Smartphone - Input Expert Review .......................................... 323
5.3.2.5 Smartphone – Customer Rating............................................... 323
5.3.3.1 Smartphone - Input Friends ..................................................... 324
5.3.4.1 "Love it or exchange it"- (B preferred; A recommendation FB) 324
5.3.4.2 "Love it or exchange it"- (A preferred; B recommendation FB) 325
6 Mobile Contract .................................................................................................. 325

6.1 Mobile Contract - Group 1 (Control Group) ................................................... 325
6.1.2 Pre-Buying Phase – Mobile Contract ................................................... 326
6.1.2.2 Reason for Main Information Source (1) .................................. 327
6.1.3 Buying Phase – Mobile Contract.......................................................... 329
6.1.3.1 Mobile Contract - No External Information ............................... 329
6.1.4 Post-Buying Phase – Mobile Contract ................................................. 329
6.1.4.1 "Love it or exchange it"- (at Neutral) ........................................ 329
301
__________________________________________________________________________
6.2 Mobile Contract - Group 2 (FB Input in Buying Phase) ................................. 330
6.2.2.5 Mobile Contract - Input Customer Rating ................................. 333
6.2.2.6 Mobile Contract - Input Expert Review ..................................... 334
6.2.3.1 Mobile Contract - Input FB ....................................................... 334
6.2.4.1 "Love it or exchange it"- (B preferred; A recom. of Friends) ..... 335
6.2.4.2 "Love it or exchange it"- (A preferred; B recom. of Friends) ..... 335
6.3 Mobile Contract - Group 3 (Friends Input in Buying Phase).......................... 336
6.3.2.5 Mobile Contract - Input Expert Review ..................................... 339
6.3.2.6 Mobile Contract - Input Customer Reviews .............................. 340
6.3.3.1 Mobile Contract - Input Friends ................................................ 340
7 Entertainment Service ....................................................................................... 342

7.1 Entertainment Service - Group 1 (Control Group) ......................................... 342
7.1.2 Pre-Buying Phase - Entertainment Service.......................................... 343
7.1.3 Buying Phase - Entertainment Service ................................................ 345
7.1.3.1 Entertainment Service - No External Information ..................... 345
7.1.4 Post-Buying Phase – Entertainment Service ....................................... 345
7.1.4.1 "Love it or exchange it"- (at Neutral) ........................................ 345
302
__________________________________________________________________________
7.2 Entertainment Service - Group 2 (FB Input in Buying Phase) ....................... 347
7.2.2.5 Entertainment Service - Input Customer Review ..................... 350
7.2.2.6 Entertainment Service - Input Expert Review........................... 351
7.2.3.1 Entertainment Service - Input FB ............................................. 351
7.2.4 Post-Buying Phase - Entertainment Service ........................................ 352
7.2.4.1 "Love it or exchange it"- (B preferred; A recom. friends) .......... 352
7.2.4.2 "Love it or exchange it"- (A preferred; B recom. friends) .......... 352
7.3 Entertainment Service - Group 3 (Friends Input in Buying Phase) ............... 353
7.3.2.5 Entertainment Service - Input Expert Review........................... 356
7.3.2.6 Entertainment Service - Input Customer Review ..................... 357
7.3.3.1 Entertainment Service - Input Friends...................................... 357
7.3.4 Post-Buying Phase - Entertainment Service ........................................ 358
8 End ...................................................................................................................... 359
303
__________________________________________________________________________
1 Welcome
Welcome to the survey!
Thank you for participating in this survey!
The survey is part of a dissertation at the Friederich Alexander University. All your details will
be completely anonymous.
The general aim of the survey is to identify different behaviours in the buying decision proc-
ess.
The contact person for the survey is Natascha Pavlovic (natascha.pavlovic@fau.de).
At the end of the survey you have the opportunity to take part in a lottery. 5 Amazon vouch-
ers worth 20 € will be raffled.
Please click on the arrow below to start the survey!

• Duration of survey: approx. 5–10 minutes
2 Socio-demographic Data
→ In the programmed questionnaire the whole chapter 2 was moved to the end. It is only still
listed here in the original order as there are various references in the questionnaire that
refer to the origin numbering.
2.1 Age
How old are you?
Question type: Single choice Next

1  < 18 years
2  18 - 24
3  25 - 29
4  30 - 39
5  40 - 49
6  50 - 59
7  60 - 69
8  > 69
304
__________________________________________________________________________
2.2 Gender
Are you…?

1  …female
2  …male
2.3 Household
2.3.1 Household Size
How many people — children and adults together — live in your household, including
yourself?
Please always consider the household in which you currently see the centre of your life and
add up all the people who live in this household permanently.

3  I live alone
4  2 persons
5  3 persons
6  4 persons
7  5 persons
8  6 or more persons
2.3.2 Household Type

Who is living with you in your household?
Please always think of the household in which you currently see the centre of your life.
Question type: Multiple Choice Next
1  Spouse or partner
2  Spouse or partner & children
3  Mother/father/legal guardian
4  Mother/father/legal guardian & sibling
5  Other family members (e.g. grandparents, aunt, uncle)
6  Flatmates
7  Other Person
305
__________________________________________________________________________
2.4 Profession
Which professional group do you currently belong to?

1  Pupil
2  Volunteering Year/Federal Voluntary Service
3  Trainee
4  Student
5  Worker/skilled worker
6  Employee/Official without management function
7  Employee/Official with management function
8  Self-employed/Entrepreneur
9  Currently looking for work/registered unemployed
10  On parental leave/parental leave/other leave
11  Housewife/househusband
12  In (early) retirement/pensioner
13  Not specified and/or other
2.5 Educational Level

What is your highest educational level?

1  No graduation
2  Secondary School Certificate/Comparable
3  Intermediate maturity/ Comparable
4  A-level/Comparable
5  University or college degree (Diploma, Master, Bachelor, etc.)
6  Doctoral Studies
7  Other:____________
306
__________________________________________________________________________
3 Segmentation Questions
3.1 Personality
Now the focus is on a general description of your personality.
How do the following statements apply to you? I see myself as someone who:
1 = Disagree strongly, 4 = Neither nor, 7 = Agree strongly
Question type: Single choice per row, Items rotate 1 2 3 4 5 6 7 Nex

Disagree Nei- Agree t
strongly ther strongly
nor
1 8 is reserved       
2 9 is generally trusting       
3 10 tends to be lazy       
4 11 is relaxed, handles stress well       
5 12 has few artistic interests       
6 13 is outgoing, sociable       
7 14 tends to find fault with others       
8 15 does a thorough job       
9 16 gets nervous easily       
10 17 has an active imagination       
3.2 Interests
This section about your general interest in mobile communications, digital topics and general
preferences in buying decisions.
How do the following statements apply to you?
1 = Disagree strongly, 4 = Neither nor, 7 = Agree strongly
307
__________________________________________________________________________
Question type: Single choice per row, Items rotate 1 2 3 4 5 6 7 Nex

Disagree Nei- Agree t
strongly ther strongly
nor
1       
I am very interested in technical things
2 Digital media (Internet, computer, mobile phone,       
etc.) are an essential part of my everyday life
3 Digital entertainments (movies, music, games, etc.)       
are an essential part of my everyday life
4       
I value brand products a lot
5 In my circle of friends and acquaintances we often       
talk about our smartphones and the latest apps
6 I am up on new trends at an early stage       
7 The price is usually the most important criterion for       

me when choosing a mobile communications prod-
uct
8 A good personal support and a good service are       
very important to me
9 High quality is more important to me than a lower       
price
10 When I like something, I don't spend a lot of time       
thinking about it and buy it.
3.3 Social Media Activity

Now it's about your average social media activity. Please think of the entire spectrum of
social media, for example social networks (e.g. Facebook, Xing), comparison portals (e.g.
Check24, Amazon), video portals (e.g. YouTube), blogs (e.g. Twitter) or forums.
3.3.1 Daily Social Media Content Creation

How often do you create content (e.g. texts, videos, photos, detailed reviews) and publish it
on social media on average per day?

1  ____ (number between 1 - 1000 enter); 3.3.3
2  less than 1 time per day 3.3.2
308
__________________________________________________________________________
3.3.2 Social Media Content Creation per Week

How often do you create content (e.g. texts, videos, photos, detailed reviews) and publish it
on social media on average per week?

3.3.3 Participation in Social Media per Day

How often do you participate/take part in social media on average per day? (By sharing,
commenting on (e.g. 'Like', 'Dislike') or rating content.)

3.3.4 Participation in Social Media per Week

How often do you participate/take part in social media on average per week? For exam-
ple, by sharing, commenting on (e.g. 'Like', 'Dislike') or rating (e.g. Rating) content.

3.3.5 Social Media Consumption per Day

How often do you consume content (e.g. read, watch, listen) from social media on average
per day?

1  ____ (number between 1 - 1000 enter);
3.3.6 Social Media Consumption per Week

How often do you consume content (e.g. read, watch, listen) from social media on aver-
age per week?

1  ____ (number between 0 - 1000 enter);
309
__________________________________________________________________________
4 Comparison Group Selection (Automated Classification)

Assignment of the test persons to the groups takes place automatically 5-6-7-, 5-6-7, etc. to
achieve equal distribution of the groups and no 'bias' via preferences (programming via Uni-
park).
→ Automatic assignment!
Question type: Single choice, Items rotate
1  Smartphones 5
2  Mobile Contract 6
3  Entertainment Service/for (movies & series) 7
5 Smartphone
5.1 Smartphone - Group 1 (Control group)

In the following, we will focus on information concerning the decision for a new smartphone.
5.1.1 Product Involvement
If you would buy a new smartphone for yourself, how important would be the following
facts to you regarding the smartphone?
1 = Very unimportant, 4 = Neutral, 7 = Very important
Question type: Single choice per row (items rotating) 1 2 3 4 5 6 7
Very Neu- Very
unim- tral impor-
portant
tant
1 Interest in product       
2 Fun to use       
3 Personal expression possibilities, Identification       

with the product
4 Perceived degree of risk; 'probability of falling into       
it'
5 Perceived risk in terms of risk costs. e.g. in the       
event of loss or damage
5.1.2 Pre-Buying Phase - Smartphone

5.1.2.1 Main Information Source (1)
Which source would be the most important for you to inform yourself before buying a
smartphone?
310
__________________________________________________________________________
Question type: Single choice (Answers rotating) Next

1  Social Networks (e.g. Facebook, Twitter, LinkedIn)
2  Expert reviews (e.g. Stiftung Warentest)
3  Online forums
4  Review and rating sites (e.g. CCheck24, Amazon)
5  Blogs
6  Family
7  Video portals (e.g. YouTube)
8  Friends
9  Acquaintances/workmates
10  Others: _________________________
5.1.2.2 Reasons for Main Information Source (1)

Why would you consider this source as most important? Question type: Next
Maximum 2/(Answers rotating)
1  Transparency/complexity reducing
2  Credibility
3  Objectivity/independence
4  Trustworthiness
5  Personal relationship
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
5.1.2.3 Second Most Important Information Source (2)

Which source would be the second most important for you to inform yourself before buy-
ing a smartphone?
311
__________________________________________________________________________

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
5.1.2.4 Reasons for Second Most Important Information Source (2)

Why would you consider this source as second most important?
Question type: Maximum 2/(Answers rotating) Next
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
5.1.3 Buying Phase - Smartphone

5.1.3.1 Smartphone - No External Information
Now you can choose between two smartphones: Smartphone A and Smartphone B. You
have no further information about the smartphones.
312
__________________________________________________________________________
If you had to decide on a purchase now, which smartphone would you choose?
Question type: Single choice per row
Smartphone A Neutral Smartphone B
1 2 3
1   
Next 5.1.4.2 5.1.4.1 5.1.4.3
5.1.4 Post-Buying Phase - Smartphone

5.1.4.1 "Love it or exchange it"- (At Neutral)
Now you will be informed that there is a special offer that you can take up when you make
your purchase. The promotion is called "Love it or exchange it", which means that you have
the opportunity to exchange the smartphone within four weeks.
In case of a purchase - how likely would you take advantage of such an exchange?
1 = Very unlikely, 4 = Neutral, 7 = Very likely
Question type: Single choice per row 1 2 3 4 5 6 7 Next

Very Neutral Very
unlikely likely
1 Using the right of exchange        7.2
5.1.4.2 "Love it or exchange it"- Decision (A preferred)

Now you will be informed that there is a special offer that you can take advantage of when
you complete the purchase. The promotion is called "Love it or exchange it", which means
that you have the opportunity to exchange the smartphone within four weeks.
Now imagine that you have bought Smartphone A and have been using it for three weeks.
You are very happy with your decision because the smartphone has completely fulfilled
your expectations.
Now you receive a reminder message from the provider that you have the possibility to ex-
change your Smartphone A for Smartphone B within one week.
313
__________________________________________________________________________
How likely would you use the right of exchange?


Very Neutral Very
unlikely likely
5.1.4.3 Love it or exchange it"- Decision (B preferred)

that you have the opportunity to exchange the smartphone within four weeks
Now suppose that you have bought Smartphone B and have been using it for three weeks.
You are very happy with your decision because the Smartphone has fully fulfilled your ex-
pectations.
Now you receive a reminder from the provider that you have the possibility to exchange
your Smartphone B for Smartphone A for another week.
How likely would you take advantage of the exchange?

Very Neutral Very
unlikely likely
5.2 Smartphone - Group 2 (FB Input in Buying Phase)

In the following, we will discuss all the information you need to know to decide on a new
smartphone.
If you would buy a new smartphone for yourself, how important would be the following as-
pects regarding the smartphone for you?
1 = Very important, 4 = Neutral, 7 = Very unimportant
314
__________________________________________________________________________
Question type: Single choice per row (items rotated) 1 2 3 4 5 6 7
Very Neu- Very
impor- tral unim-
tant por-
tant
2 Fun to use       
3 Personal expression possibilities, Identification with       

the product
it'

smartphone?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

Why would you consider this source as the most important?
315
__________________________________________________________________________
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

Which source would be the second most important for you to inform yourself before buy-
ing a smartphone?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

Why would you consider this source as the second most important?
316
__________________________________________________________________________
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
5.2.2.5 Smartphone - Input Customer Rating

There are two smartphones to choose from: Smartphone A and Smartphone B and you
have the customer ratings from one of the largest and best-known comparison portals
(Check24):
Smartphone A: 4 of 5 Stars (4.100 Customer ratings)

Smartphone B: 2 of 5 Stars (4.500 Customer ratings)
Question type: Single choice per row Next
4 of 5 Stars 2 of 5 Stars
(Customer ratings) (Customer ratings)
1 2 3
1   
317
__________________________________________________________________________
5.2.2.6 Smartphone - Input Expert Review

Now you get even more information about the smartphones from a renowned independ-
ent consumer organisation (Stiftung Warentest) based on expert reviews:
Smartphone A: 2 of 5 Stars (independent expert review)

Smartphone B: 4 of 5 Stars (independent expert review)
(expert review) (expert review)
1 2 3
1   

5.2.3.1 Smartphone - Input FB
Now you have received some information and/or ratings for the two smartphones. Right be-
fore your final purchase decision, you will receive the following additional information
from Facebook.
Smartphone A: 20.000 "Likes" (Likes information) and numerous negative com-

ments as e.g. "Regret the purchase", "Does not fulfil its prom-
ises", etc.
Smartphone B: 40.000 "Likes" (Likes information) and numerous positive com-
ments as e.g. "Best smartphone I've ever had", "I highly recom-
mend it to anyone", etc.
If you had to make a purchase decision now, which smartphone would you choose?
318
__________________________________________________________________________
2.000 Likes + negative com- 4.000 Likes + positive com-

ments ments
(Facebook) (Facebook)
1 2 3
1   
Next 5.2.4.2 5.1.4.1 5.2.4.1

5.2.4.1 "Love it or exchange it"- (B preferred; A recommendation friends)
you make your purchase. The promotion is called "Love it or exchange it", which means that
you have the opportunity to exchange your smartphone within four weeks.
Now assume that you have bought Smartphone B and you have been using it for three
weeks. You are very happy with your decision because the Smartphone has fulfilled all
your expectations.
You are now informed that some of your friends have also bought a new smartphone in
the last few weeks. The majority of them have chosen Smartphone A; they are totally en-
thusiastic about their new Smartphone and would definitely recommend it to others.
Now you receive a reminder message from the provider that you have the possibility to ex-
change your Smartphone B for Smartphone A for another week.

Very Neutral Very
unlikely likely
1 Using the right of exchange       
5.2.4.2 "Love it or exchange it"- (A preferred; B recommendation friends)

that you have the opportunity to exchange the smartphone within four weeks.
319
__________________________________________________________________________
Now suppose that you have bought Smartphone A, and have been using it for three weeks.
pectations.
You are now informed that some of your friends have also bought a new smartphone in
the last few weeks. The majority of them have chosen Smartphone B; they are totally en-
thusiastic about their new Smartphone and would definitely recommend it to others.
Now you will receive a reminder from the provider that you have the possibility to exchange
your Smartphone A for Smartphone B for another week.

Very Neutral Very
unlikely likely
5.3 Smartphone - Group 3 (Friends Input in Buying Phase)

In the following, we will focus on information concerning the decision for a new smartphone.
If you would buy a new smartphone for yourself, how important would be the following as-
pects regarding the smartphone to you?
Vrey Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

with the product
it'

smartphone?
320
__________________________________________________________________________

3  Online forums
4  Review and rating sites (e.g. Check24, Amazon)
5  Blogs
6  Family
8  Friends
10  Others: _________________________
Next
Question type: Maximum 2/(Answers rotating)
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
321
__________________________________________________________________________

Which source would be the second most important for you to inform yourself Next
before buying a smartphone? Question type: Single choice (Answers rotat-
ing)
3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

Why would you consider this as the second most important source?
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
322
__________________________________________________________________________
5.3.2.4 Smartphone - Input Expert Review

There are two smartphones to choose from: Smartphone A and Smartphone B and you
have expert reviews from a renowned independent consumer organisation (Stiftung
Warentest):
Smartphone A: 2 of 5 Stars (independent expert review)
Smartphone B: 4 of 5 Stars (independent expert review)
2 of 5 Stars 4 von 5 Stars

1 2 3
1   
5.3.2.5 Smartphone – Customer Rating

Now you get even more information about the smartphones from one of the largest and
most well-known comparison portals (Check24) based on customer ratings:
Smartphone A: 4 of 5 Stars (4.100 customer ratings)

Smartphone B: 2 of 5 Stars (4.500 customer ratings)
(Expert review) (Expert review
1 3 4
1   
323
__________________________________________________________________________

5.3.3.1 Smartphone - Input Friends
Now you have received some information and/or ratings of the two smartphones. Right be-
fore you make your final purchase decision, you notice that some of your friends have
also bought a new smartphone in the last few weeks.
Significantly fewer have chosen Smartphone A and they are less satisfied with their new
smartphone. They mention some points of criticism and would not recommend it.
The majority of them have chosen Smartphone B. They are totally enthusiastic about
their new Smartphone and would definitely recommend it.
(Expert review) (Expert review)
No recommendation from Clear recommendation from

your friends your friends
1 2 3
1   
Next 5.3.4.2 5.1.4.1 5.3.4.1

5.3.4.1 "Love it or exchange it"- (B preferred; A recommendation FB)
Now you get the information that there is a special offer that you can use when making a
purchase. The promotion is called "Love it or exchange it", which means that you have the
opportunity to exchange your smartphone within four weeks.
Now assume that you have bought Smartphone B, and have been using it for three weeks.
pectations.
You also get the following information about smartphones from Facebook:
Smartphone A: 40.000 "Likes" (Like Information) and numerous positive comments

such as "Best Smartphone I ever had", "I can recommend it to
everyone", etc.
Smartphone B: 20.000 "Likes" (Like Information) and numerous negative comments
such as "Regret the purchase", "Does not fulfil its promises", etc.
324
__________________________________________________________________________
Now you get a reminder message from the provider that you have the possibility to ex-
change your Smartphone B for Smartphone A for another week.
How likely would you make use of the exchange?

Very Neutral Very
unlikely likely
5.3.4.2 "Love it or exchange it"- (A preferred; B recommendation FB)

Now you get the information that there is a special offer that you could take advantage of
when making a purchase. The promotion is called "Love it or exchange it", which means that
you have the opportunity to exchange your smartphone within four weeks.
Now assume that you have bought Smartphone A, and have been using it for three weeks.
pectations.
You get the following information about the smartphones from Facebook:
Smartphone A: 20.000 "Likes" (Like Information) and numerous negative comments

such as "Regret the purchase", "Does not fulfil its promises", etc.
Smartphone B: 40.000 "Likes" (Like Information) and numerous positive comments
such as "Best Smartphone I ever had", "I can recommend it to
everyone", etc.
Now you get a reminder message from the provider that you have the possibility to ex-
change your Smartphone A for Smartphone B for another week.
1 = Very unlikely, 4 = Neutral, 7 = Very likely 1 2 3 4 5 6 7 Next

Very Neutral Very
unlikely likely
6 Mobile Contract
6.1 Mobile Contract - Group 1 (Control Group)

In the following, the focus is on information concerning the decision for a new mobile con-
tract.
325
__________________________________________________________________________

If you would conclude a new mobile contract for yourself, how important would be the fol-
lowing aspects to you regarding the mobile phone contract
Very Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

the product
it'
6.1.2 Pre-Buying Phase – Mobile Contract

Which source would be the most important for you to inform yourself before concluding
a mobile contract?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
326
__________________________________________________________________________
6.1.2.2 Reason for Main Information Source (1)

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

Which source would be the second most important for you to inform yourself before
concluding a mobile contract?
327
__________________________________________________________________________

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
328
__________________________________________________________________________
6.1.3 Buying Phase – Mobile Contract

6.1.3.1 Mobile Contract - No External Information
Now there are two mobile contracts to choose from: Mobile contract A and mobile con-
tract B, you have no further information about the mobile contracts.
If you had to decide on a purchase now, which mobile contract would you choose?
Mobile Contract A Neutral Mobile Contract B
1 2 3
1   
Next 6.1.4.2 6.1.4.1 6.1.4.3
6.1.4 Post-Buying Phase – Mobile Contract

6.1.4.1 "Love it or exchange it"- (at Neutral)
Now you get the information that there is an offer that you could take advantage of by signing
the contract. The promotion is called "Love it or exchange it", which means that you have the
possibility to change your mobile contract within four weeks.
In the case of signing a contract, how likely would you make use of the exchange option?

Very Neutral Very
unlikely likely

possibility to change your mobile phone contract within four weeks.
Now assume that you have concluded mobile contract A, and have been using it for three
weeks. You are very happy with your decision because the mobile contract has fully ful-
filled your expectations.
Now you receive a reminder message from the provider that you have the possibility to
change your mobile contract A to mobile contract B for another week.
329
__________________________________________________________________________

Very Neutral Very
unlikely likely

change your mobile contract B against mobile contract A for another week.

Very Neutral Very
unlikely likely
6.2 Mobile Contract - Group 2 (FB Input in Buying Phase)

In the following, the focus is on information concerning the decision for a new mobile phone
contract.
If you would conclude a new mobile contract for yourself - how important would be the fol-
lowing aspects to you regarding the mobile phone contract
330
__________________________________________________________________________
1 = Very unimportant, 4 = Neutral, 7 = Very important 1 2 3 4 5 6 7
Very Neu- Very

Question type: Single choice per row (items rotating)
unim- tral impor-
portant
tant
2 Fun to use       

the product
it'

a mobile contract?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

331
__________________________________________________________________________
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

before concluding a mobile contract? Question type: Single choice (Answers
rotating)
3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
332
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
6.2.2.5 Mobile Contract - Input Customer Rating

Now there are two mobile contracts to choose from: Mobile contract A and mobile con-
tract B and you have the customer ratings from one of the largest and best-known compari-
son portals (Check24):
Mobile Contract A: 4 of 5 Stars (4.100 Customer Ratings)

Mobile Contract B: 2 of 5 Stars (4.500 Customer Ratings)
1 2 3
1   
333
__________________________________________________________________________
6.2.2.6 Mobile Contract - Input Expert Review

Now you receive even more information on mobile phone tariffs from a renowned inde-
pendent consumer organisation (Stiftung Warentest) based on expert reviews:
Mobile Contract A: 2 of 5 Stars (independent expert review)
Mobile Contract B: 4 of 5 Stars (independent expert review)
1 2 3
1   

6.2.3.1 Mobile Contract - Input FB
Now you have received some information and/or ratings for the two Mobile Contracts. Right
before your final purchase decision you will receive the following additional information
from Facebook.
Mobile Contract A: 20.000 "Likes" (Like information) and numerous negative comments
as e.g. " Regret the conclusion of the contract", "Does not fulfil its
promises", etc.
Mobile Contract B: 40.000 "Likes" (Like information) and numerous positive comments
such as "Best tariff I have ever had", "I can recommend it to eve-
ryone", etc.
334
__________________________________________________________________________

ments ments
1 2 3
1   
Next 6.2.4.2 6.1.4.1 6.2.4.1

6.2.4.1 "Love it or exchange it"- (B preferred; A recommendation of Friends)
signing the contract. The promotion is called "Love it or exchange it", which means that you
have the opportunity to change the mobile contract within four weeks.
Now assume that you have concluded mobile contract B, you have just decided on and
have been using it for three weeks. You are very happy with your decision, because the mo-
bile contract has fully fulfilled your expectations.
Now you are informed that some of your friends have also signed a new mobile contract in
the last few weeks. The majority of them have chosen mobile contract A, they are totally
enthusiastic about their new tariff and would definitely recommend it to others.
change your mobile contract B for mobile contract A for another week.
How likely would you take advantage of the exchange option?

Very Neutral Very
unlikely likely
6.2.4.2 "Love it or exchange it"- (A preferred; B recommendation of Friends)

335
__________________________________________________________________________
Now assume that you have concluded mobile contract A, you have just decided on and
have been using it for three weeks. You are very happy with your decision, because the mo-
bile contract has fully fulfilled your expectations.
Now you are informed that some of your friends have also signed a new mobile contract in
the last few weeks. The majority of them have chosen mobile contract B, they are totally
enthusiastic about their new tariff and would definitely recommend it.

Very Neutral Very
unlikely likely
6.3 Mobile Contract - Group 3 (Friends Input in Buying Phase)

The following section is about information concerning the decision for a new mobile phone
tariff.
If you would sign a new mobile contract for yourself, how important would be the following
aspects of the mobile contract to you?
Very Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

with the product
it'
336
__________________________________________________________________________

a mobile contract?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________

337
__________________________________________________________________________
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

Which source would be the second most important for you to inform yourself
before concluding a mobile contract?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
338
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
6.3.2.5 Mobile Contract - Input Expert Review

There are now two mobile contracts to choose from: Mobile contract A and mobile
contract B and you have expert reviews from a renowned independent consumer
organisation (Stiftung Warentest):
Mobile Contract A: 2 of 5 Stars (independent expert review)
Mobile Contract B: 4 of 5 Stars (independent expert review)
1 2 3
1   
339
__________________________________________________________________________
6.3.2.6 Mobile Contract - Input Customer Reviews

Now you receive further information on mobile contracts from one of the largest and
best-known comparison portals (Check24) based on customer reviews:
Mobile Contract A: 4 of 5 Stars (4.100 Customer reviews)
Mobile Contract B: 2 of 5 Stars (4.500 Customer review)
(Expert review (Expert review)
(Customer reviews) (Customer reviews)
1 2 3
1   

6.3.3.1 Mobile Contract - Input Friends
Now you have received some information or ratings of the two mobile phone tariffs.
Shortly before making your final purchase decision, you notice that some of your friends
have also signed a new mobile contract in the last few weeks.
Significantly fewer have chosen mobile contract A and they are less satisfied with their
new tariff. They name a few points of criticism and would not recommend it.
The majority of them have chosen mobile contract B. They are totally enthusiastic
about their new tariff and would definitely recommend it.
340
__________________________________________________________________________
(Customer review) (Customer reviews)

1 2 3
1   
Next 6.3.4.2 6.1.4.1 6.3.4.1

Now assume that you have concluded mobile contract B, and have been using it for three
You get the following information about mobile contracts from Facebook:
Mobile Contract A: 40.000 "Likes" (Like information) and numerous positive comments
ryone", etc.
Mobile Contract B: 20.000 "Likes" (Like information) and numerous negative comments
such as "Regret the contracting ", "Does not keep its promises",
etc.
change your mobile contract B to mobile contract A for another week.

Very Neutral Very
unlikely likely
341
__________________________________________________________________________

You get the following information about mobile contracts from Facebook:
Mobile Contract A: 20.000 "Likes" (Like information) and numerous negative comments
such as "Regret the contracting ", "Does not keep its promises",
etc.
Mobile Contract B: 40.000 "Likes" (Like information) and numerous positive comments
ryone", etc.

Very Neutral Very
unlikely likely
7 Entertainment Service
7.1 Entertainment Service - Group 1 (Control Group)

This section provides information about the decision for a new entertainment service (access
to exclusive movies and series).
If you were to subscribe a new entertainment service for yourself, how important are the
following aspects of the entertainment service to you?
342
__________________________________________________________________________
Very Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

the product
it'
7.1.2 Pre-Buying Phase - Entertainment Service

Which source would be the most important for you to inform yourself before deciding on
an entertainment service?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
343
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

Which source would be the second most important for you to inform yourself before de-
ciding on an entertainment service?

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
344
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
7.1.3 Buying Phase - Entertainment Service

7.1.3.1 Entertainment Service - No External Information
Now there are two entertainment services to choose from: Entertainment service A and
Entertainment service B. There is no further information about the entertainment services.
If you had to decide on a purchase now, which entertainment service would you choose?
Entertainment Service A Neutral Entertainment Service B
1 3 4
1   
Next 7.1.4.2 7.1.4.1 7.1.4.3
7.1.4 Post-Buying Phase – Entertainment Service

7.1.4.1 "Love it or exchange it"- (at Neutral)
have the opportunity to change the entertainment service within four weeks.
In the case of concluding a contract, how likely would you make use of this exchange op-
tion?
345
__________________________________________________________________________

Very Neutral Very
unlikely likely

Now assume that you have concluded entertainment service A, and have been using it for
three weeks. You are very happy with your decision because the Entertainment service has
fully fulfilled your expectations.
change your entertainment service A to entertainment service B for another week.

Very Neutral Very
unlikely likely

have the opportunity to change the Entertainment service within four weeks.
Now assume that you have concluded entertainment service B, and have been using it for
change your entertainment service B to entertainment service A for another week.

Very Neutral Very
unlikely likely
346
__________________________________________________________________________
7.2 Entertainment Service - Group 2 (FB Input in Buying Phase)

In the following, the focus is on information concerning the decision for a new entertainment
service (access to exclusive movies and series).
If you subscribe a new entertainment service for yourself, how important would be the
Very Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

the product
it'
347
__________________________________________________________________________

3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
348
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

Which source would be the second most important for you to inform yourself before
deciding on an entertainment service?
3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
349
__________________________________________________________________________

2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
7.2.2.5 Entertainment Service - Input Customer Review

There are now two entertainment services to choose from: Entertainment service A and
entertainment service B and you have the customer reviews from one of the largest and
best-known comparison portals (Check24):
Entertainment Service A: 4 of 5 Stars (4.100 Customer reviews)

Entertainment Service B: 2 of 5 Stars (4.500 Customer reviews)
1 2 3
1   
350
__________________________________________________________________________
7.2.2.6 Entertainment Service - Input Expert Review

Now you receive further information about the entertainment services from a renowned
independent consumer organisation (Stiftung Warentest) based on expert reviews:
Entertainment Service A: 2 of 5 Stars (independent Expert review)

Entertainment Service B: 4 of 5 Stars (independent Expert review)
1 2 3
1   

7.2.3.1 Entertainment Service - Input FB
Now you have received a few information and ratings of the two entertainment offers. Right
before your making your final purchase decision, you will receive the following addi-
tional information from Facebook.
Entertainment Service A: 20.000 "Likes" (Like information) and numerous negative

comments such as "Regret the contracting", "Does not
keep its promises", etc.
Entertainment Service B: 40.000 "Likes" (Like information) and numerous positive
comments such as "Best offer I have ever had", "I can rec-
ommend it to everyone", etc.
351
__________________________________________________________________________

ments ments
1 2 3
1   
Next 7.2.4.2 7.1.4.1 7.2.4.1
7.2.4 Post-Buying Phase - Entertainment Service

7.2.4.1 "Love it or exchange it"- (B preferred; A recommendation friends)
Now you are noticing that some of your friends have also decided on a new entertainment
service in the last few weeks. The majority of them have chosen entertainment service A,
they are totally enthusiastic about it and would definitely recommend it to others.

Very Neutral Very
unlikely likely
7.2.4.2 "Love it or exchange it"- (A preferred; B recommendation friends)

352
__________________________________________________________________________
Now you are noticing that some of your friends have also decided on a new entertainment
service in the last few weeks. The majority of them have chosen entertainment service B,
they are totally enthusiastic about it and would definitely recommend it to others.

Very Neutral Very
unlikely likely
7.3 Entertainment Service - Group 3 (Friends Input in Buying Phase)

In the following, the focus is on information concerning the decision for a new entertainment
service (access to exclusive movies and series).
If you subscribe a new entertainment service for yourself, how important would be the
Very Neu- Very
unim- tral impor-
portant
tant
2 Fun to use       

with the product
it'
353
__________________________________________________________________________


3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
354
__________________________________________________________________________

Why would you consider this source as the most important?Question type: Next
Maximum 2/(Answers rotating)
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________

before deciding on an entertainment service?Question type: Single choice
(Answers rotating)
3  Online forums
5  Blogs
6  Family
8  Friends
10  Others: _________________________
355
__________________________________________________________________________
2  Credibility
6  Speed
7  Actuality
8  Mass opinion
9  Convenience
10  Others: _________________________
7.3.2.5 Entertainment Service - Input Expert Review

Now there are two entertainment services to choose from: entertainment service A and
entertainment service B and you have expert reviews from a renowned independent
consumer organisation (Stiftung Warentest):
Entertainment Service A: 2 of 5 Stars (independent expert review)

Entertainment Service B: 4 of 5 Stars (independent expert review)
1 2 3
1   
356
__________________________________________________________________________
7.3.2.6 Entertainment Service - Input Customer Review

Now you receive further information about the entertainment services from one of the larg-
est and best-known comparison portals (Check24) based on customer ratings:
Entertainment Service A: 4 of 5 Stars (4.100 Customer reviews)

Entertainment Service B: 2 of 5 Stars (4.500 Customer reviews)
1 2 3
1   

7.3.3.1 Entertainment Service - Input Friends
Now you have received a few information and ratings of the two entertainment services.
Right before making your final purchase decision, you notice that some of your friends
have also decided on a new entertainment service in the last few weeks.
Significantly fewer have chosen entertainment service A and they are less satisfied with
the service. They mention some points of criticism and would not recommend it.
The majority of them have chosen entertainment service B. They are totally enthusias-
tic about the service and would definitely recommend it to others.
357
__________________________________________________________________________

1 2 3
1   
Next 7.3.4.2 7.1.4.1 7.3.4.1
7.3.4 Post-Buying Phase - Entertainment Service

You get the following information about the entertainment services from Facebook:
Entertainment Service A: 40.000 "Likes" (Like information) and numerous positive

comments such as "Best offer I've ever had", "I recommend
it to everyone", etc.
Entertainment Service B: 20.000 "Likes" (Like information) and numerous negative
comments such as "Regret the conclusion of the contract ",
"Does not keep its promises", etc.

Very Neutral Very
unlikely likely
358
__________________________________________________________________________

You get the following information about the entertainment services from Facebook:
Entertainment Service A: 20.000 "Likes" (Like information) and numerous negative
comments such as "Regret the conclusion of the contract ",
"Does not keep its promises", etc.
Entertainment Service B: 40.000 "Likes" (Like information) and numerous positive
comments such as "Best offer I've ever had", "I recommend
it to everyone", etc.

Very Neutral Very
unlikely likely
8 End
Now you have reached the end of the survey.
Thank you very much for your participation!
You are welcome to use the opportunity to comment on the survey conducted.
Insert text field for free text
If you want to participate in the lottery (5 x 20€ Amazon voucher) please enter your email ad-
dress:
Insert text field for email address
359
__________________________________________________________________________
E.2: Screenshots Unipark (section 7.4.2.1)

The survey was programmed with the Unipark software. The following screenshots
are used to illustrate the survey approach.
360
__________________________________________________________________________
361

Dissertation Natascha Pavlovic

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dissertation Natascha Pavlovic

Uploaded by

Copyright:

Available Formats

Herd Behaviour in Consumer Buying

Decisions in the Age of Social Media

Der Rechts- und Wirtschaftswissenschaftlichen Fakultät /

der Friedrich-Alexander-Universität Erlangen-Nürnberg

Dipl.-Kffr. Nataša Pavlović-Höck,

von der Rechts- und Wirtschaftswissenschaftlichen Fakultät /

Promotionstermin: 7. Juli 2020

Tag der mündlichen Prüfung: 22. Juni 2020

Vorsitzende/r des Promotionsorgans Prof. Dr. Markus Beckmann

Gutachter/in: Prof. Dr. Michael Amberg

Based on conducted expert interviews, determining factors of herd behaviour in buy-

To develop a deeper understanding on herd behaviour along the buying decision

Durch die Analyse von nutzergenerierten Inhalten dreier verschiedener 'Social-

Auf der Grundlage von durchgeführten Experteninterviews werden entscheidende

Um ein tieferes Verständnis für das Herdenverhalten entlang der Kaufentschei-

Zusammenfassend bietet diese Dissertation einen umfassenden Einblick in das Ge-

Acknowledgements .................................................................................................. III

Abstract (German) ................................................................................................... VI

Overview of Contents ........................................................................................... VIII

Table of Contents .................................................................................................... IX

List of Figures........................................................................................................ XIV

List of Tables ......................................................................................................... XVI

List of Abbreviations ........................................................................................... XVIII

2 Research Background ......................................................................................... 14

3 Research Design .................................................................................................. 45

4 Study I: State-of-the-Art Analysis ....................................................................... 64

8 Overall Discussion ............................................................................................. 200

9 Conclusion ......................................................................................................... 207

References ............................................................................................................. 218

Appendix A: Supplement to Chapter 3 ................................................................ 249

Appendix B: Supplement to Chapter 4 ................................................................ 251

Appendix C: Supplement to Chapter 5 ................................................................ 276

Appendix D: Supplement to Chapter 6 ................................................................ 284

Appendix E: Supplement to Chapter 7 ................................................................ 298

Acknowledgements .................................................................................................. III

Abstract (German) ................................................................................................... VI

Overview of Contents ........................................................................................... VIII

Table of Contents .................................................................................................... IX

List of Figures........................................................................................................ XIV

List of Tables ......................................................................................................... XVI

List of Abbreviations ........................................................................................... XVIII

1.1 Motivation ......................................................................................................... 2

1.2 Research Focus ................................................................................................ 5

1.3 Objective of the Dissertation ............................................................................. 6

1.4 Structure of the Dissertation ............................................................................. 9

1.5 Bibliographic Notes ......................................................................................... 12

1.6 Conventions .................................................................................................... 13

2 Research Background ......................................................................................... 14

2.1 Herd behaviour ............................................................................................... 15

2.2 Consumer Buying Decision Process ............................................................... 24

2.3 Social media ................................................................................................... 35

3 Research Design .................................................................................................. 45

3.1 IS Research Paradigms and Methods ............................................................ 46

3.2 Applied Research Methodologies ................................................................... 50

3.3 Research Strategy .......................................................................................... 58

4 Study I: State-of-the-Art Analysis ....................................................................... 64

4.1 Literature Review Approach............................................................................ 65

4.2 Results of the Quantitative Analysis ............................................................... 71

4.3 Results of the Qualitative Analysis of the Literature Set ................................. 78

4.3.1.2 Information Sources ................................................................... 82

4.4 Summary and Interim Discussion ................................................................... 98

5.1 Objectives of the Study ................................................................................. 104

5.2 Theoretical Foundation ................................................................................. 106

5.3 Methodical Approach .................................................................................... 109