You are on page 1of 22

This article was downloaded by: [Dalhousie University]

On: 08 August 2013, At: 01:00


Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Marketing Communications


Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/rjmc20

Word-of-mouth marketing influence


on offline and online communications:
Evidence from case study research
a a
Lars Groeger & Francis Buttle
a
Macquarie Graduate School of Management, Macquarie
University , North Ryde , NSW , 2109 , Australia
Published online: 10 Jun 2013.

To cite this article: Journal of Marketing Communications (2013): Word-of-mouth marketing


influence on offline and online communications: Evidence from case study research, Journal of
Marketing Communications, DOI: 10.1080/13527266.2013.797736

To link to this article: http://dx.doi.org/10.1080/13527266.2013.797736

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Journal of Marketing Communications, 2013
http://dx.doi.org/10.1080/13527266.2013.797736

Word-of-mouth marketing influence on offline and online


communications: Evidence from case study research
Lars Groeger* and Francis Buttle1

Macquarie Graduate School of Management, Macquarie University, North Ryde, NSW 2109,
Australia

This case study reports results from three research studies conducted over 12 weeks as
part of a product seeding campaign. Partnering with a word-of-mouth marketing
(WOMM) agency for this research, studies 1 and 2 report agency-conducted surveys of
campaign participants’ online and offline word-of-mouth (WOM) behaviors. Study 3
Downloaded by [Dalhousie University] at 01:00 08 August 2013

deployed an innovative web-based methodology to map and visualize WOM


communication patterns, to reveal how campaign-related conversations spread within
and across offline friendship networks and the role played by tie strength in that
process. We find that agency reports of WOMM campaign results overstate reach and
understate frequency. Our results have implications for the measurement of reach and
frequency of WOMM campaigns.
Keywords: word-of-mouth; word-of-mouth marketing; brand ambassador; Facebook;
social network analysis

Introduction
Word-of-mouth marketing (WOMM) campaigns are an increasingly popular component
of the marketing communications mix, being referenced not only in the popular business
press (Rosen 2009; Sernovitz 2009) but also in contemporary marketing management
texts (Kotler, Keller and Burton 2009). Often, WOMM campaigns are associated with
influencer strategies in which products are placed with persons who are expected to use,
share, and talk about the product with their friends and family; these persons are known as
brand ambassadors, buzz agents, or product seeds. Armstrong and Kotler (2011, 170), for
example, note that an increasing number of businesses are applying ‘buzz marketing by
enlisting or even creating opinion leaders to serve as “brand ambassadors” who spread the
word about their products’. These products may be completely new-to-market or new to a
segment that is being targeted. Product sampling is not new (Holmes and Lett 1977; Jain,
Mahajan, and Muller 1995). Being a well-established sales promotion tool, it is widely
used to introduce new products and generate positive word-of-mouth (WOM) or buzz.
However, the association of product sampling with a dedicated WOMM campaign is a
relatively new phenomenon.
A brand ambassador campaign provides the context for our research, which explores
problematic issues of campaign effectiveness, in particular the measurement of WOMM
campaign reach and frequency. We present a case study of one particular WOMM
campaign undertaken as part of a product seeding strategy for a new product launch.
Overall, we report three studies. The first two were conducted by the agency and
investigated the online and offline WOM behaviors of campaign participants. The third
study was conducted by the authors of this paper and explores how the social-structural

*Corresponding author. Email: lars.groeger@mgsm.edu.au

q 2013 Taylor & Francis


2 L. Groeger and F. Buttle

characteristics of participants’ social networks influence the spread of the campaign


message within and across generations of conversation partners.

Literature review
WOM researchers, from the 1950s (Katz and Lazarsfeld 1955) to the late 1990s
(Sundaram, Mitra, and Webster 1998), focused on offline dyadic interactions, such as
how opinion leaders influence followers, ignoring the socio-structural context within
which these interactions take place. In this century, however, researchers have shifted their
interest to online environments, where WOM has become styled as word-of-web and
word-of-mouse (Breazeale 2009). People can now be connected in many electronically
enabled ways, including online social networks such as Facebook, Twitter and LinkedIn,
blogs, wikis, file-sharing services, chat rooms, and online communities.
The idea of motivating customers to spread pro-brand messaging within their social
Downloaded by [Dalhousie University] at 01:00 08 August 2013

networks is well established (Mancuso 1969), but the work by Reingen and Kernan (1986)
helped marketers understand the significant impact that social networks could have on
message dissemination. They called for research into the influence of social networks on
consumer behaviors and were critical of previous WOM research for ‘its failure to capture
the social-structural context within which such communication is embedded’ (Reingen
and Kernan 1986, 370). Twenty years later, Van Den Bulte and Wuyts (2009) again urged
researchers to develop an improved understanding of the location of individuals in social
networks as this could offer considerable insights into how WOM is disseminated.
Recently, marketing researchers have begun to explore the significance of consumer-
generated content in online environments. For example, Liu (2006) finds that the volume
of online WOM mentions, rather than positive or negative valence, best predicts box office
success for movies. Other studies have examined the effects of recommendation behaviors
from customers on Amazon.com and similar shopping websites (Chevalier and Mayzlin
2006), social networking platforms (Trusov, Bucklin, and Pauwels 2009), bulletin boards
(Huang 2010), chat rooms (Godes and Mayzlin 2004), and online communities
(Mathwick, Wiertz, and De Ruyter 2008; Zhu and Zhang 2010). Of these, only Trusov,
Bucklin, and Pauwels (2009) investigated the effects of social networks on message
dissemination.
Very few researchers have focused specifically on product sampling and WOMM
campaigns. Carl (2007) set out to understand what he calls the ‘generational relay’ of
WOMM messaging – the spread of WOM from the original WOMM campaign
participants (Generation Zero or Gen0) to their conversation partners (Generation One or
Gen1), and then from those Gen1 conversation partners to the next generation of
conversation partners (Gen2), and then to Gen3. In their later work, Carl, Libai, and Ding
(2008) analyzed inter-generational WOM relay data from a WOMM campaign for a low
involvement consumer good that had 5000 original Gen0 participants. Ahuja et al. (2007)
explored whether buzz agents feel any ethical tensions about exerting commercial
influence on their friends and family. Kozinets et al. (2010) studied an online WOMM
campaign in which consumers were seeded with a new technology device and encouraged
to stimulate WOM by writing about it in their personal blogs. Hinz et al. (2011) compared
four product seeding strategies in two small-scale field experiments and one real-life viral
marketing campaign, and found evidence that the best seeding strategies, those that focus
on seeding hubs (people with many ties) or bridges (people who connect two or more
otherwise disconnected clusters of people), can be eight times more successful than the
least successful ones.
Journal of Marketing Communications 3

Problem definition
In conventional advertising research, reach and frequency are two commonly used
effectiveness measures (Katz 2010). Reach refers to the number of a defined target
audience (e.g., people or households) exposed at least once to a firm’s advertising message
in a defined period of time. Reach is broadly synonymous with ‘cumulative audience’
(AMA 2012). Frequency is the average number of times a member of that audience is
exposed to the message during the defined period of time.
Media planners may make assumptions about the number of times an audience
member needs to be exposed to a message in order for the desired cognitive, affective, or
behavioral outcome to be achieved. For example, low involvement products are often
advertised with high frequency. McDonald (1995, 1) suggests that media planners should
strive for ‘effective frequency’, the underlying concept of which he describes thus: ‘If
there is too little exposure, the advertising will fail to be noticed; on the other hand, if there
Downloaded by [Dalhousie University] at 01:00 08 August 2013

is too much, the recipient will be ‘saturated’ and the surplus will be redundant’. Writing in
pre-digital days, Krugman (1972) suggested that three exposures might be enough to
arouse action, while Kamin (1978) suggested that it was not unusual for a planner to
recommend a media mix that reaches over 90% of prospects an average of eight times
each. Cheong, de Gregorio, and Kim’s (2010, 403) contemporary survey of US advertising
agencies report that ‘traditional exposure-based criteria such as reach-and-frequency
distribution remain important [in the age of digital advertising] and often are used in
evaluations of offline media schedules’.
Thus, it is still accepted that target audiences may need to be exposed to an ad several
times for the message to have its desired effect (Katz 2010). The potential for multiple
exposures of an audience member to the campaign message, whether offline or online, is
often known or estimated from syndicated media usage research (Cheong, de Gregorio,
and Kim 2010). No such insight into frequency is currently available for WOMM
campaigns. Our interactions with our partner agency, Soup2 (http://www.thesoup.com.au),
and investigations of other WOMM agencies such as BzzAgent, SheSpeaks, Tremor, trnd,
and Vocalpoint indicate that frequency is usually not reported by the agencies that create
and run WOMM campaigns. In the absence of data about frequency, it is assumed that
each person reached in a WOMM campaign is a unique identity. Thus, the total number of
campaign-related conversations is used as a proxy for campaign reach.
Moreover, the integration of reach generated by WOMM campaigns with equivalent
data from associated offline and online marketing campaigns to provide a unified
assessment of the reach and frequency of an integrated multi-channel marketing campaign
can be very troublesome. Agencies currently have little understanding of the incremental
ad message exposures of persons reached by WOMM campaigns. Even if the offline and
online data were available and of good quality, integration of the data to provide a single
picture of campaign reach and frequency would be conceptually and technically difficult.
Measurement of the effectiveness of WOMM campaigns can be difficult,
inconvenient, and costly. Our WOMM agency partner measures WOMM campaign
reach by asking Gen0, the brand ambassadors, in an online survey how many people they
spoke to about the product or campaign. The ambassadors in turn invite the members of
Gen1 via email to also participate in the survey. In the absence of specifically identified
persons, researchers tend to rely on estimations of reach based on the number of reported
conversations, as outlined in Figure 1.
For example, if the product is given to 500 brand ambassadors (Gen0) and each reports
talking to an average of 10 people, Gen0-to-Gen1 reach is assumed to be 5000 (500 £ 10)
4 L. Groeger and F. Buttle

500 Generation 0
(Gen0)
5,000 Generation 1
10 conversations (Gen1)

20,000 Generation 2
4 conversations (Gen2)

Figure 1. Example of reach estimation across multiple generations.

persons. If a sample of Gen1 is subsequently contacted and reports talking to a further four
people on average about the campaign, then Gen1-to-Gen2 reach is estimated at 20,000
(5000 £ 4) persons. Estimates of WOMM campaign reach therefore assume that each person
Downloaded by [Dalhousie University] at 01:00 08 August 2013

spoken to is a unique person. However, there is clearly some possibility that these 20,000
persons are not unique individuals; in other words that there may be multiple exposures of
some persons (illustrated in light gray in Figure 1) to the campaign-related messaging.
Lazarsfeld and Merton (1954) introduced the term ‘homophily’ to refer to the tendency
of persons to affiliate with others who have similar attributes such as age, education, or
ethnicity. Whether that does happen in the WOMM context will depend on the social-
structural characteristics of the networks of Gen0 participants and subsequent generations.
If homophily is influential, then it may be inappropriate for agencies and clients to assume
that the total number of campaign-related conversations is the same as the campaign’s
reach. Homophily leads to the formation of relatively homogenous groups or network
clusters (McPherson, Smith-Lovin, and Cook 2001). Also, because the friends of any
single person are often friends with each other, it also is very likely that these friends also
speak to each other (Christakis and Fowler 2009). This property of network relationships is
called transitivity (see Figure 2).
If there is a friendship tie between person A and person B, and another between person
A and person C, then in a transitive network persons B and C will also be connected. This
idea is captured by the expression ‘friends of my friends are my friends’ (Davis 1970).
Strong friendship ties, characterized by high clustering, are more often transitive than
weak ties, and thus transitivity is often seen as evidence for the existence of strong ties
(Shi, Adamic, and Strauss 2007). In principle, therefore, if campaign-related WOM
diffuses along strong, transitive ties from Gen0 to Gen1, it is possible that when Gen1
passes on the message it will be to a Gen2 who could also be friends with Gen0.

?
A

Figure 2. Transitivity of ties.


Journal of Marketing Communications 5

Figure 3 shows how transitivity could lead to multiple exposures. The left-hand side of
the figure shows that there are nine conversations in total between Gen0 and Gen1, and
Gen1 and Gen2. As shown on the right-hand side of the figure, the transitive ties between
Gen1 and Gen2 mean that only seven people are reached by the campaign. Both of
the people spoken to by person B (Gen1) are also spoken to by person C (also Gen1).
This illustrates that the true number of people reached by a WOMM campaign can only be
measured if conversations are associated with individuals embedded within a network of
friendship ties.
The purpose of our research is to present a case study of one particular WOMM
campaign in order to shed light on how agencies may report WOMM campaign reach, and
to subject agency practices and assumptions to critical reflection. Our research partner
joins us in striving to improve the quality of reporting not only of WOMM campaigns
per se, but of multi-channel integrated communication strategies that include WOMM
campaigns. Our analysis of the social-structural attributes of WOMM campaign
Downloaded by [Dalhousie University] at 01:00 08 August 2013

participants’ social networks takes us outside the regular scope of agency-conducted


WOMM campaign measurement to identify unique conversation partners and to
differentiate explicitly between reach and frequency.

The case of Hahn White


The brand ambassador campaign we report and analyze supported the launch of a new
beer, Hahn White, into the Australian market. The Australian beer market is in the mature
stage of its life cycle. In the 5 years through to 2011 –2012, industry value added fell by an
annualized rate of 0.7%, compared to an annualized rate of national GDP growth of 2.6%.

Reach based on conversations Reach based on unique individuals


Nine conversations in total within transitive network
Seven individuals exposed to message
(Nine conversations)

C C

A B A B
Gen 0 Gen 0

Conversation

Figure 3. How transitivity may lead to multiple exposures.


6 L. Groeger and F. Buttle

In those 5 years, new products were introduced both by the market leaders, Fosters and
Lion-Nathan, and by smaller boutique competitors (IBISWorld 2012). Our case study is a
Lion-Nathan innovation.
The WOMM campaign’s goal was to drive momentum for the adoption of Hahn
White during summer by creating tasting occasions, encouraging brand ambassadors
(the agency called them summer ambassadors) to talk to friends about Hahn White, and
increasing their propensity to purchase. The campaign was run by our WOMM agency
partner.
A total of 2000 members of the agency’s panel of potential brand ambassadors were
invited to an on-premise tasting. On the basis of their relative enjoyment of the beer and
their fit with the brand’s target market profile, 800 of the 2000 attendees were invited to
become summer ambassadors. The participants were aged 25 –35, male and female, city
dwellers, drank beer at least fortnightly, and were organizers of social events for their
Downloaded by [Dalhousie University] at 01:00 08 August 2013

friendship networks. The timeline of the WOMM campaign and the research we report
here is outlined in Figure 4.
The summer ambassadors were sent two cases of the beer, coasters, and branded
glasses for two key dates, Christmas Day (25 December) and Australia Day (26 January).
They were expected and encouraged to incorporate the new beer into social get-togethers,
parties, and barbeques.
During the course of the WOMM campaign, three studies were conducted, all of which
were principally designed to develop a better understanding of campaign reach. We report
those studies below, and later reflect on what the results mean for agencies and their
clients. Studies 1 and 2 were conducted by our partner and follow the agency’s routine
research and reporting practices for all the campaigns they run. Study 3 is a non-agency
investigation conducted by authors of this article.

Research conducted by WOMM agency as Academic


part of typical campaign management research

Brand
Campaign Study 3
Ambassador Study 1 Study 2
Start Facebook App
Selection

800 panel Two cases of beer Online Survey Online Survey Facebook app
members selected are sent to
to be summer Gen0 Gen0 report on Gen0 report on Pass on function to
ambassadors the number of the number of invite Gen1
Tasting and online and offline additional offline
(Gen0) sharing events Collection of FB
conversations conversations
take place friendship network
since campaign since study 1
data from Gen0 and
start
Gen1
Gen1 report on Gen0 and Gen1
the number of identify unique
offline conversation partners
conversations and define the
strength of their
friendship tie

Campaign set-up Campaign Start 4 weeks 12 weeks

Figure 4. Campaign and data collection set-up.


Journal of Marketing Communications 7

Study 1: first agency-conducted census survey of Gen0


One month after the campaign began, as is customary practice, the agency conducted an
online census survey of the summer ambassadors to investigate the reach of the campaign
and the WOM behaviors of participants; 630 of the 800 summer ambassadors responded
(78.8%). The agency did not investigate the possibility of non-response bias influencing
the results reported here. Results showed that an average of 16.6 people attended the
first sharing event and 15.4 people attended the second, resulting in 25,600 people
experiencing a Hahn White event. Summer ambassadors reported sharing Hahn White
bottles with 26.4 people on average each, resulting in 21,120 event-based tastings of Hahn
White as a direct result of the campaign. Because the agency did not collect identification
details of attendees, it had no idea how many individuals attended these events. Given the
influence of homophily and transitivity, it is possible that some attendees participated in
more than one event, meaning that the 25,600 attendees are not unique individuals.
Downloaded by [Dalhousie University] at 01:00 08 August 2013

In this first census survey, participants were asked about their offline campaign-related
conversations since the start of the campaign 4 weeks before. To reduce the complexity of
the recall task (Nelson-Field, Riebe, and Sharp 2012), participants were asked to think about
the social events during which they shared the product. Further, they were asked to think
about their family and friends and recollect whether they had discussed Hahn White with
them during the past 4 weeks. Summer ambassadors reported speaking offline to an average
of 44 people (Gen1). Another question asked whether any ambassador had mentioned Hahn
White online in any way. The question listed a variety of online communication channels
and asked Gen0 respondents to check the channels they had used: 43% reported using
email, 73% Facebook, and 6% Twitter; 11% had commented on forums or blogs, and 2%
mentioned the brand on their own blog. Clearly, since these percentages total more than
100%, some ambassadors had mentioned the product or campaign in more than one online
channel. Based on the agency’s experience of over 100 previous WOMM campaigns, they
estimated the total online campaign reach at 56,614 persons (Table 1), which is equivalent
to about 71 persons per ambassador. The agency does not know whether these are unique
persons or whether there is some degree of multiple exposures; for example, some Gen1
might receive emails from two different Gen0, or might see a Facebook item and receive a
Tweet. Equally, the agency has no insight into whether the offline Gen1 population is
different from the online Gen1 population. Given our earlier remarks about homophily and

Table 1. Agency’s estimation of online reach.

Forums, message
boards, Participant’s
Email Facebook Twitter blogs, etc. own blog
Percentage of participants com- 43 (271) 73 (459) 6 (38) 11 (69) 2 (13)
municating by each channel
(n ¼ 630)
Average audience size for mem- 20 248 301 30 915/month
bers of the Soup community
Estimated percent of online 95 30 10 30 30
messages with receiver
response (e.g., email opened,
blog read)
Agency’s estimate of channel 6536 43,450 1445 792 4392
reach (n ¼ 800)
Total online reach (n ¼ 800) 56,614
8 L. Groeger and F. Buttle

transitivity, there is certainly a possibility of overlap between the two populations, meaning
that there is potential for multiple exposures to the campaign messaging from both online
and offline sources.
Table 1 shows the agency’s computation of online reach. For example, 271 of the
ambassadors reported using email to ‘talk’ about Hahn White. Internal studies conducted
by the agency led them to estimate that each email from a panel member reached about 20
Gen1. They also estimated that 95% of these emails are opened and read by the recipients
(Soup 2012). Assuming that the online communication behaviors of the 630 participants
are representative of all 800 ambassadors, email messaging about the brand or campaign
reached 6536 persons (800 £ 0.43 £ 20 £ 0.95). Applying the same computational logic
to each online channel produces the numbers shown on the ‘channel reach’ row of the
table.
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Study 2: second agency-conducted survey of Gen0


Following the agency’s routine campaign research and reporting practices, a second
agency-conducted survey was undertaken 12 weeks after the campaign began. This online
survey was only sent to the 630 participants in study 1. In total, 304 completed responses
were received (48% of the 630 who participated in study 1; 38% of the original 800
summer ambassadors). This survey found continued growth in offline campaign-related
conversations. Gen0 participants were first presented with their response from the study 1.
They were then asked if they had spoken to any additional people, clearly stating that they
should only report incremental conversations. Summer ambassadors reported having a
further 11 offline conversations, on average, resulting in a total of 55 offline Gen0-to-Gen1
conversations for each ambassador in 3 months. Ambassadors were invited to pass this
second online survey to the Gen1 to whom they had spoken. Gen0 forwarded the survey to
665 Gen1; 89 G1 clicked on it, 66 attempted the survey and 52 Gen1 respondents
completed the survey successfully, which represents a response rate for Gen1 of 7.5%.
These 52 Gen1 survey participants reported speaking on average to another 7.3 persons
about the campaign. In this particular campaign, Gen2 was not contacted, but previous
studies by the agency suggest that number of offline conversations drop by 50% from
Gen1 to Gen2 (Soup 2012). Thus, it is estimated that on average another 3.5 conversations
were held between Gen2 and Gen3. Figure 5 summarizes the number of total offline
conversations based on the agency’s routine campaign measurement practices. It shows
that the campaign reached an estimated 1,240,787 people offline over three generations.
The agency cannot tell from its research whether these are unique persons or whether some
of those reached were exposed more than once to campaign-related messaging. However,

1,240,787
800
Total number of Conversations
Generation 0

55.4 conversations 44,320


Gen 0 to Gen 1 Conversations

7.3 conversations 319,845


Gen 1 to Gen 2 Conversations

876,622
Gen 2 to Gen 3 Conversations
3.5 conversations

Figure 5. Estimate of offline reach 12 weeks after campaign start.


Journal of Marketing Communications 9

once again, given the presence of homophily and transitivity, we believe there is a
possibility of multiple exposures.

Study 3: social network investigation


The third study steps outside the agency’s customary research and reporting processes and
uses an alternative approach, grounded on social network analysis (SNA), to estimate the
reach of the Hahn White WOMM campaign. Figure 4 illustrates how our investigation fits
into the timeline of research conducted for this campaign and summarizes how it differs
from Soup’s routine campaign measurement practices in terms of both methodology and
data collected. Invitations to participate in study 2 and study 3 were sent out at the same
time, because study 3 took place immediately after study 2.
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Methodology
SNA is a sociological methodology that identifies an individual’s role in a group or
community and maps the network of connections between that individual and others
(Moreno 1934). One of SNA’s leading principles is ‘that dyadic relationships do not occur
in isolation, but rather form a complex structural pattern of connectivity and cleavage
beyond the dyad’ (Kilduff and Brass 2010, 317).
We used SNA to identify the nodes (friends) and friendship ties that make up the social
networks of campaign participants: we identified ego (participant), ego’s alters
(participant’s Facebook friends), alter – alter relationships (ego’s friends who are friends
with each other on Facebook), and the strength of the ties that connected these participants
in the WOMM campaign. By applying SNA to map the network connections of campaign
participants and then identifying specific conversation partners within this network, we
were able to link each campaign-related conversation to a unique individual. SNA has
generally been limited to examining small, well-bounded populations involving a small
number of snapshots of interaction patterns (Eagle, Pentland, and Lazer 2009). This is
mostly due to the inconvenience and high costs of capturing the names of an individual’s
friends and subsequently mapping their relationships through the usual data collection
methodologies of interview, survey, or observation. However, we apply SNA to this much
larger and potentially less tightly bounded network of WOMM campaign participants by
tapping into existing network data on Facebook.com.

Facebook
Facebook is a social networking phenomenon, with 845 million monthly active users by
the end of 2011 (Facebook 2012). Facebook allows users to download their own ‘ego’
friendship network and to learn which of these people are friends with each other. Hence,
Facebook provides a convenient and low-cost platform for exploring social-structural
network effects. For our research, participants granted us permission to access their ego
friendship networks, as enabled by Facebook for all their members. Although the
Facebook friendship network is a digital construct, we were particularly interested to know
who had talked to whom in real, ‘offline’ life about the Hahn White WOMM campaign.
We used Facebook social network data as a proxy for the offline social network(s)
within which the WOMM campaign is embedded. Clearly, it is only legitimate to use
Facebook data as a proxy if the online and offline social worlds of participants align.
Previous research suggests that Facebook users tend to interact online with people with
10 L. Groeger and F. Buttle

whom they have already an offline relationship (Ellison, Steinfield, and Lampe 2007;
Lewis et al. 2008) and that, although online friends are not a precise mirror of offline
friendships, they are a reasonable proxy (Hogan 2008; Subrahmanyam et al. 2008). We
further tested this assumption in our own research.
Our results did confirm the alignment of the online and offline social networks of our
participants: 83% of all Gen0 and Gen1 participants in our research reported that at least
half of their regular offline social interaction partners were also their friends on Facebook.
Indeed, 51% stated that ‘most’ or ‘pretty much all’ of their offline friends were in their
online friendship network. Therefore, we are confident that our participants’ Facebook
friendship networks are a reasonable proxy of their offline social networks, though not a
precise mirror.

Data collection through Facebook app


Downloaded by [Dalhousie University] at 01:00 08 August 2013

Data about WOMM campaign participants’ social networks were collected using a
proprietary Facebook application (app) that was specially developed and custom built for
this research in partnership with the agency. The app invited respondents to identify the
members of their Facebook friendship network to whom they had talked about Hahn
White, and to report the strength of the ties that bind those friendships.
Each participant’s list of Facebook friends was automatically populated to the screen
by the app, and participants merely had to check boxes to indicate, firstly, that a campaign-
related conversation had taken place with a particular person and, secondly, to report tie
strength. Thus, we were able to model each participant’s friendship network and the
embedded communication ties that connect Facebook members. In addition, summer
ambassadors (Gen0) could also forward the survey onto members of their social network
(Gen1), who could complete it and in turn could send it on to Gen2 directly through
Facebook or via email. Each person was identified with a unique identification number.
This allowed us to protect the privacy and confidentiality of survey participants whilst
simultaneously building a network picture of WOMM campaign participants and their
conversational partners.
Three months after the launch campaign began, our WOMM agency partner sent an
email message to ambassadors inviting them to participate in our Facebook survey.
Participation was incentivized with a gift of two blocks of chocolate. Study 3 was
undertaken immediately after study 2 (see Figure 4). We hoped that study 2 would have
encouraged participants to reconnect with the campaign and bring campaign-related issues
to the front-of-mind.

Results from the Facebook app survey


NodeXL, an open-source template that can be applied to network data stored in Excel 2007
or 2010 workbooks, was used to understand, map, and visualize the social-structural
attributes of participants’ social networks.
We collected friendship networks of 304 ambassadors (Gen0) and 52 Gen1. Figure 6 is
a sample of one friendship network of a summer ambassador (Mr Green). The ambassador
(ego) is at the center of the network, and by definition connected to all other members
(alters) of the network, represented as black dots. SNA practitioners call these ‘nodes’.
Lines (known in SNA as ‘ties’) between the nodes represent friendships between the
summer ambassadors’ friends (alter– alter ties).
In Figure 7, we now integrate the offline conversations that Mr Green had with
members of his friendship network about the campaign. The solid arrows pointing away
Journal of Marketing Communications 11

Facebook Friendship Network of Mr Green, Gen0

Gen0
Mr Green
Gen1
Mr Orange
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Gen1
Ms Blue

Figure 6. Illustrative Gen0 friendship network (Mr Green).

from the central node represent conversations. Two conversation partners (Gen1) are
marked as Mr Orange and Ms Blue as they also participated in our research. Thus, they
accepted an invitation to complete our survey that had come from Mr Green and identified
their respective conversation partners (Gen2).
Participants identified a total of 5294 conversation partners and the strengths of ties
that bound those friendships. Although there are several multidimensional measures of

Facebook Friendship Network of Mr Green, Gen 0

Gen0
Mr Green
Gen1
Mr Orange

Gen1
Ms Blue

Mr Green (Gen0)
to Gen1 Conversations

Figure 7. Mr Green’s conversations to Gen1 embedded within friendship network.


12 L. Groeger and F. Buttle

tie-strength (Petróczi, Nepusz, and Baszo 2006), we used ‘closeness’ a single-item measure
of tie-strength. Marsden and Campbell (1984, 497) state that ‘closeness (the measure of the
emotional intensity of a tie) is the best indicator of tie strength’ and has often been used as
single indicator of tie-strength in previous research.3 Our selection of this single-item
measure was partly due our desire to minimize respondent burden. Our scale item asked
participants to rate the closeness of each friendship on a nine-point Likert-type scale
ranging from 1 ¼ barely know the person to 9 ¼ we are very close friends.
We found that that the stronger the tie, the more likely it was to be activated for Hahn
White-related conversations: 70% of all offline Hahn White conversations travelled along
strong ties (points 7 –9 on the scale) and 30% along weaker ties (points 1 –6). This is
confirmation of Reingen and Kernan’s (1986) and Carl’s (2006) finding that strong ties are
the dominant conduits for referral behaviors.
We then investigated the number of common friends as a function of tie strength. Our
motivation for this was our belief that having a high number of common friends raises the
Downloaded by [Dalhousie University] at 01:00 08 August 2013

possibility of multiple exposures to the campaign message. If this were the case, there
would be a concomitant reduction in campaign reach, because the assumption that each
person spoken to is unique falls away. Our analysis showed that conversation partners who
reciprocally define their friendship tie as strong have on average 41 common friends (20%
of all Gen0’s friends are also friends with Gen1), whereas weaker-tied friendships have
only 8 common friends, or 4% friendship overlap. In SNA terms, a Gen1 whose friendship
tie is defined as strong has an average ‘degree’ of 41 as opposed to a weak tie Gen1 with an
average ‘degree’ of 8, degree being the number of connections to other nodes.
Figure 8 illustrates the friendship overlap using the example of Mr Green and two of
his conversations partners, Mr Orange and Ms Blue, who also provided us with their
network data. Mr Orange and Mr Green defined their friendship tie as strong. Thirty
percent of Mr Green and Mr Orange’s friends are friends with both of them. It is therefore
possible that Mr Green and Mr Orange both spoke about the campaign to their common
friends, thus limiting campaign reach in terms of unique individuals, whilst increasing the
likelihood that their common friends are exposed to the campaign message multiple times.
In contrast, Mr Green and Ms Blue defined their friendship tie as weak. They have only
3% of friends in common. However, this weak tie provides an opportunity for the
campaign to reach into new clusters of the social network, as there is a strong possibility
that Ms Blue could pass campaign-related information on to people who do not know and
have not spoken to Mr Green (see Figure 9). This would be confirmation of Granovetter’s
(1973, 1982) weak tie hypothesis, which suggests that while strong ties are more likely to
be associated with within-cluster influence, weak ties allow communication to flow
between otherwise disconnected clusters of nodes.
Figure 9 now further integrates the conversations of the Gen0, Mr Green, and two
participating Gen1 (Ms Blue and Mr Orange). Solid arrows pointing away from Mr
Orange and Ms Blue represent offline conversations that they had with members of their
friendship network (Gen2). Clearly, Mr Orange’s conversation partners are mostly located
within the cluster of common friends he shares with Mr Green. Ms Blue, however, reaches
out to members of her network that are not even connected to Mr Green. Ms Blue and Mr
Orange are also not directly connected.
Figure 10 looks in more detail now at the conversations that Mr Green and Mr Orange
initiated. Two of Mr Orange’s six conversation partners were also spoken to by Mr Green,
thereby reducing Mr Orange’s reach to four persons instead of six.
Through following the analysis of all generations’ social networks and conversational
behaviors, we found that 20.7% of all offline campaign-related conversations were
Journal of Marketing Communications 13

Facebook Friendship Network of Mr Orange, Gen1

Friendship Overlap
Mr Green & Mr Orange

Gen1
Mr Orange

Gen0
Mr Green
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Friendship Overlap
Ms Blue & Mr Orange

Gen1
Ms Blue

Facebook Friendship Network of Mr Blue, Gen1

Figure 8. Integrating Mr Orange’s and Ms Blue’s friendship network.

Facebook Friendship Network of Mr Orange, Gen1

Mr Orange to Gen2
Conversations

Mr Green (Gen0)
to Gen1 Conversations

Ms Blue to Gen2
Conversations

Facebook Friendship Network of Mr Blue, Gen1

Figure 9. Ms Blue’s and Mr Orange’s conversations embedded.


14 L. Groeger and F. Buttle

Mr Orange to Gen2
Conversations

Multiple Exposures of
Mr Green’s and Mr
Orange’s common
friends

Mr Green (Gen0)
Downloaded by [Dalhousie University] at 01:00 08 August 2013

to Gen1 Conversations

Figure 10. Multiple exposures of Mr Green’s and Mr Orange’s common friends.

Table 2. Total campaign reach and multiple exposures (mean).

Number of Number of unique Multiple


Generation conversations individuals exposure (%)
Strong ties Gen0 þ Gen1 þ Gen2 969,873 731,085 24.6
Weaker ties Gen0 þ Gen1 þ Gen2 270,878 252,893 6.6
Total 1,240,751 983,978 20.7

instances of multiple exposures. By differentiating between the conversations along strong


and weaker ties, only 6.6% of conversations along weaker ties lead to multiple exposures,
as opposed to 24.6% along strong ties. Even when we remove the extreme cases from our
analysis and compute trimmed means, there is no change in the percentages of multiple
exposures (see Table 2).

Discussion
This case study describes three studies associated with a single WOMM campaign,
including two routine agency-conducted analyses of the offline and online reach of the
campaign (studies 1 and 2) and an independent investigation (study 3) that also
investigates campaign reach (and frequency) but using a different methodology, SNA.
We are able to draw several conclusions from this research. First, we consider the
results and implications of study 3 where we used SNA to assess campaign reach and
frequency. Our evidence finds that approximately 21% of offline conversations in the
Hahn White campaign were with persons who also were reached by another member of the
node’s social network. We conclude that agency estimates of WOMM campaign reach are
overestimated. Agencies err when they assume that every conversation is with a unique
person. Multiple exposures, it appears, are a product of the strong ties that exist between
members of social networks, and those ties, in turn, may be associated with homophily and
transitivity. These results provide support for Godes and Mayzlin’s (2009) assertion that
weak-tie acquaintances are important for the spread of WOM campaigns, because
although they are peripheral in a participant’s ego network, they are likely to be central in
Journal of Marketing Communications 15

another social network. Where weak ties connect nodes, those conversational partners
have only on average 4% of friends in common as opposed to 20% for strong ties.
Study 1 revealed that each summer ambassador reached on average 71 persons (56,000
in total) through their online use of email, Twitter, Facebook, chat-rooms, and blogs. It is
certainly possible that Gen1 receivers of those messages forward, re-post, or re-tweet some
of them, thereby further extending online exposure to the message beyond Gen1.
However, the agency does not capture this data. Neither does the agency know whether the
56,000 persons who reached online are unique persons or whether some of them are
exposed to online messaging from more than one conversational partner or in more than
one channel.
The agency did collect some tantalizing evidence of overlap between the offline and
online worlds of campaign participants during study 2, in which participants were asked
open-ended questions about how they had used Hahn White. Alannah (aged 32) wrote:
Downloaded by [Dalhousie University] at 01:00 08 August 2013

For the sharing bottles, everyone I knew was all gathered out after Christmas through
Australia day rush. So I just posted on my Facebook that the first 6 people to visit my house
bringing a snack could have 2 beer glasses and 2 large bottles of beer. I had 8 people show up
and we had a mellow night with music.
This remark suggests two things. First, this is evidence of an alignment of online and
offline social networks, with Facebook friends meeting face-to-face for the mellow
evening; second, this is evidence of multiple exposure to campaign-related messaging –
first online and then at Alannah’s home. Study 1 also found that summer ambassadors
(Gen0) spoke to about 35,000 people offline about Hahn White (average < 44 persons)
during the 4 weeks after the campaign start. Again, the agency does not know whether
these are unique persons. It is certainly possible, given homophily and transitivity, that
some of these 35,000 persons are also represented in the 56,000 persons that were ‘spoken’
to online. Study 2 further explored offline reach with a further 11 persons being spoken to
by each summer ambassador, making a total of about 55 on average or about 44,000 across
the community of summer ambassadors over the course of three months. Yet again, the
agency does not know whether these are unique persons, because the survey methodology
only allowed for reporting of gross numbers, not specific identified people.
The Gen0 survey in agency-conducted study 2 could also be forwarded to Gen1 and
Gen2. Data obtained from these cohorts, sensitized by historical data sourced from our
agency partner, suggest that the campaign reached an estimated 1,240,787 people offline
over three generations (Figure 4). The agency does not know whether these are unique
persons. Evidence from study 3 suggests that approximately 21% of offline WOMM-
messaging are multiple exposures. This implies that true offline reach to unique
individuals may be 980,000 (i.e., 1,240,747 £ 0.79), subject to there being no significant
non-response bias. Regardless of the accuracy of these estimates, it is clear from these data
that WOMM campaigns reach out beyond the initial cohort of Gen0 participants and
expose a large number of people to the messaging across several generations.
While we explore the reach and frequency of one particular WOMM campaign, some
critics may suggest that these measures, adopted from advertising, are inappropriate for
assessing the performance of WOMM campaigns. In recent years, agency practitioners
and academics have begun to conceptualize and operationalize customer engagement
(Verhoef, Reinartz, and Krafft 2010; Hollebeek 2011). Hollebeek (2011, 565) suggests
that three themes capture the complexity of customer-brand engagement: immersion,
passion, and activation. These themes ‘represent the degree to which a customer is
prepared to exert relevant cognitive, emotional and behavioral resources in specific
16 L. Groeger and F. Buttle

interactions with a focal brand’. Gen0 participants in our focal campaign certainly spoke
favorably about the brand and shared it with friends at social events, thus devoting
emotional and behavioral resources to the campaign, so we do accept that there is merit in
these observations. However, our partner agency and others such as trnd and Bzzagent
continue to present campaign reach as one of the key performance indicators of a
campaign rather than engagement. US-based agency ChatThreads (http://www.
chatthreads.com) uses a proprietary methodology to calculate a ‘net conversation value’
for each WOMM campaign they monitor. This metric goes beyond campaign reach and
combines a customer lifetime value model with a WOM referral value model. While
ChatThreads uses an adjustment factor for ‘social network overlap’ (Cuppari et al. 2010,
12), the agency does not publish how this is calculated or whether the structure of
participants’ friendship network is considered.
We also note that our Facebook app methodology was a success. We deployed an
innovative approach to collect offline friendship data, leveraging existing Facebook
Downloaded by [Dalhousie University] at 01:00 08 August 2013

friendship networks at low cost and high convenience for both researcher and participant.
This approach enabled us to apply SNA within the context of a WOMM campaign
producing insights that refine our understanding of WOMM campaign reach. We do not
claim that this methodology is suitable for all WOMM research projects. The
demographics of the Hahn White summer ambassadors and their Facebook use may
make them particularly suitable.

Managerial implications
The WOMM industry is still in its infancy. This research shows that the industry needs to
develop indicators of reach and frequency that are as robust as those used in other media
channels. The difficulty lies in tracking the flow of WOM and thus potential overlap, not
just in offline and online contexts independently, but between these two contexts.
We are not yet in a position to assess the degree of conversation overlap between
online and offline cohorts that are reached during a WOMM campaign. Our research
reports only offline multiple exposures. Neither are we able to make any observations
about multiple exposures that might be achieved in a multimedia, multichannel campaign,
such as when new products are launched with TV, radio, online, and WOMM campaigns
aimed at the same target markets and striving to produce synergistic outcomes.
We do not mean to devalue multiple exposures to marketing communication
managers. Marks and Kamins (1988, 267) noted that
Consumers may encounter either a sample followed by exposure to advertising or exposure to
advertising followed by a product sample usage experience. These possibilities suggest some
interesting issues, including the effect of each exposure sequence on consumers’ product
beliefs, the magnitude of the attitude change created by each exposure sequence, and the
attitude formed after these two different exposure sequences.
We believe there is an opportunity to explore these issues in further depth in the product-
seeding context.
Multiple exposures may be a desirable outcome of a WOMM or multi-media
campaign, as we previously discussed. This will depend on the specific product-market
context and the cognitive, affective, or behavioral responses that are sought. For example,
for a technologically advanced innovation, it may seem reasonable to aim for a larger
proportion of multiple exposures to support a more complex learning process. However,
maximizing reach and minimizing multiple exposures could be appropriate for low
involvement products. We believe that agencies should explicitly consider the question of
Journal of Marketing Communications 17

multiple exposures when planning and evaluating WOMM campaigns on behalf of their
clients. This clearly is an important matter for those who are interested in advertising
effectiveness, especially in the context of integrated marketing communications, and
remains a significant research challenge. Clients would be better advised by agencies
reporting the number of unique persons reached by the campaign and the average exposure
frequency of those persons to the campaign message. We support Carl’s (2011) appeal for
the intelligent integration of WOMM into the Integrated Marketing Communications mix
as opposed to a silo approach.
The estimates from study 3 suggest that about 21% of all the offline campaign-related
messaging, in this particular case study, consists of multiple exposures. WOMM agencies
will overreport reach and underreport multiple exposures if they assume that every
conversation is with a unique person. We recommend that agencies no longer confound
reach with the total number of conversations. The erroneous assumption that every
conversation is with a unique person not only exaggerates WOMM campaign reach but
Downloaded by [Dalhousie University] at 01:00 08 August 2013

also campaign efficiency metrics. If we assume (hypothetically) that our case study
campaign cost $150,000 to mount, the agency’s research would suggest that the campaign
reached 1.2 million persons (see Table 1) at an average cost of 12.1 cents per person
reached ($150,000/1,240,000). The data we collected in study 3 suggest that the true reach
was only 980,000 persons, meaning that the cost per person reached was 15.3 cents.
One solution to account for multiple exposures might be to apply a correction factor to
agency estimates of campaign reach, which on the basis of this evidence would be 0.8.
However, we do not believe we are yet in a position to recommend such a factor. This case
study reports three pieces of research for a single WOMM campaign. We cannot assume
that the results from these participants in this campaign are typical of all campaigns. A
change to the campaign setup could impact the choice of attendees and subsequent
conversations. For example, providing participants with the product on premises and
encouraging them to invite acquaintances rather than close friends could lead to an
activation of fewer transitive ties and entail fewer multiple exposures. Product
characteristics also influence immediate and ongoing WOM. Products that are more
publicly visible (cars, fashion, etc.) receive more immediate and ongoing WOM (Berger
and Schwartz 2011). The duration of campaign-initiated WOM impacts not only the
number of conversations, but also the probability of multiple exposures. Further research
across various campaigns is necessary before any such correction factor could be
confidently asserted.

Limitations and opportunities for future research


Our results suggest that WOMM agencies overreport reach and underreport multiple
exposures if they (incorrectly) assume that every conversation is with a unique person.
However, we do not know whether this estimate also applies to online messaging, a
question that remains to be explored.
One limitation of the agency-conducted studies is the failure to test for non-response
bias. Both studies reported data from a subset of the populations of interest. Because there
was no follow-up of non-respondents, it is not known whether these subsets typify the
larger populations or non-respondents were materially different in their online and offline
WOM behaviors. The 630 who participated became the population for the second agency
study. If the 21% of non-respondents from study 1 are materially different, then the second
study starts from a position of bias. Further bias may have been introduced into the results
because only 48% of the 630 participated. Again, there was no test for non-response bias.
18 L. Groeger and F. Buttle

The research for study 3 was performed on the same population as study 2, and was thus
also subject to potential non-response bias.
This research was conducted in the context of a new product launch. There could be no
pre-existing marketplace WOM about the brand since it was new-to-market. The brand
launch was supported only by print advertising, point-of-sale, and this WOMM campaign.
It is not possible to conclude with absolute certainty that the POS and advertising did not
evoke some of the WOM that were identified in the research.
There is a possibility to conduct further SNA of our data-set. A fuller SNA of our data-
set could consider the influence of, for example, degree, density, and clustering on
WOMM message dissemination (Webster and Morrison 2004). Density is a measure of the
number of ties in a network expressed as a proportion of the maximum possible number of
ties. Denser networks, other things being equal, may indicate a predisposition to multiple
exposures since there is higher number of connections between network members. It
would also be possible to conduct a cluster analysis of each participant’s network. Also
Downloaded by [Dalhousie University] at 01:00 08 August 2013

known as component or subgroup, a cluster is a relatively dense group of nodes that are
only weakly connected to other clusters. Examples of real-world clusters include
university friends, sporting associates, work colleagues, family members, and partners.
Campaign participants select which persons they communicate to. If all those persons are
members of one particular cluster of an actor’s network – say, family – but not other
clusters, such as college friends and workmates, and then other things being equal, the
message may have constrained reach. We intend to continue this line of partnered research
and test a number of these propositions.

Conclusion
In summary, we believe that our research makes a number of contributions. First, it makes
an original contribution to an emerging area of academic study – the measurement of
WOMM campaign effects. Second, it uses an innovative methodology to track how
WOMM messages spread in social networks, an approach that to our knowledge has never
been used before. Third, it is a case study that contrasts industry practices with academic
standards, showing the insights that SNA can deliver. Fourth, it is an exemplar of industry –
academy collaboration. Finally, it has significant practical implications for agencies’
methods for measuring the effectiveness and efficiency of their WOMM campaigns.
WOM marketing is still far from being a line item in every company’s marketing
communication budget. There are still obstacles to greater acceptance of WOM marketing,
in particular regarding widely accepted ways to measure WOM. Agencies and their clients
can use this research to enhance their understanding of the influence of social networks on
message dissemination and on reach and frequency. Whilst acknowledging that there is
still scope for improvement, we apply a more rigorous approach to campaign measurement
and, in so doing, provide ample evidence that a WOMM campaign can work.

Notes
1. Email: francis.buttle@mgsm.edu.au
2. Soup works with a panel of more than 100,000 influencers who sample, advocate and help shape
its clients’ products. Soup works across many categories such as technology, FMCG, retail,
finance, pharmaceuticals, automotive, and government in Australia, New Zealand, and the UK.
3. This measure of closeness as an indicator of tie strength is not to be confused with the SNA
centrality metric of closeness that is defined as the sum of graph-theoretic distances from one
actor to all other actors (Freeman 1979).
Journal of Marketing Communications 19

Notes on contributors
Lars Groeger, PhD, is Lecturer in Management at Macquarie Graduate School of Management
where he teaches MBA students and Executives in Sydney and Hong Kong. His research aims to
help managers gain an increased understanding of the various forms of social interaction between
customers and the firm’s power to stimulate these interactions in a systematic and value-creating
way. Lars has been educated at leading international business schools in Germany, France and the
US. Prior to joining academia, Lars gained extensive management consulting experience in the
private and public sectors.
Francis Buttle, PhD, FCIM has over 30 years‘ experience in marketing and customer relationship
management. He is Visiting Professor at Macquarie Graduate School of Management, Sydney, and
was previously Professor of Marketing at Manchester Business School. He was the world’s first
professor of CRM. Francis founded and serves as Principal of Francis Buttle & Associates, a
worldwide network of customer management experts established in 1979. He has published over 300
items including 8 books. His most recent books are Customer Relationship Management: Concepts
and Technologies and the eBook Social CRM: what is it and what does it mean for your business?
Downloaded by [Dalhousie University] at 01:00 08 August 2013

References
Ahuja, R. D., T. A. Michels, M. M. Walker, and M. Weissbuch. 2007. “Teen Perceptions of
Disclosure in Buzz Marketing.” Journal of Consumer Marketing 24 (3): 151– 159.
AMA. 2012. “American Marketing Association Dictionary of Marketing Terms.” Accessed August
18, 2012. http://www.marketingpower.com/_layouts/Dictionary.aspx
Armstrong, G., and P. Kotler. 2011. Marketing: An Introduction, Global Edition. 10th ed. Upper
Saddle River, NJ: Pearson Higher Education.
Berger, J., and E. M. Schwartz. 2011. “What Drives Immediate and Ongoing Word of Mouth?”
Journal of Marketing Research 48 (5): 869– 880.
Breazeale, M. 2009. “Word of Mouse– An Assessment of Electronic Word-of-Mouth Research.”
International Journal of Market Research 51 (3): 297– 318.
Carl, W. J. 2006. “What’s All the Buzz About?” Management Communication Quarterly 19 (4):
601– 634.
Carl, W. J. 2007. “Measuring the Ripple: Creating the g2x Relay Multiple and Industry Standard
Methodology to Measure the Spread of Word of Mouth Conversations and Marketing-Relevant
Outcomes.” WOMMA, Measuring Word of Mouth 3: 37 – 45.
Carl, W. J. 2011. “Annual State of WOM Research Address 2011.” Paper presented at the WOMMA
Summit 2011, Las Vegas, NV, November 16-18.
Carl, W. J., B. Libai, and A. Ding. 2008. “Measuring the Value of Word of Mouth.” WOMMA,
Measuring Word of Mouth 4: 1 – 15.
Cheong, Y., F. De Gregorio, and K. Kim. 2010. “Beyond Reach and Frequency the Power of Reach
and Frequency in the Age of Digital Advertising: Offline and Online Media Demand Different
Metrics.” Journal of Advertising Research 50 (4): 403– 415.
Chevalier, J., and D. Mayzlin. 2006. “The Effect of Word of Mouth on Sales: Online Book
Reviews.” Journal of Marketing Research 43 (3): 345– 354.
Christakis, N. A., and J. H. Fowler. 2009. Connected: The Surprising Power of Our Social Networks
and How They Shape Our Lives. New York: Little, Brown and Company.
Cuppari, S., W. Carl, S. Sheldon, and D. Hunter. 2010. “ROI Demystified.” Presentation at the
WOMMA Summit 2010, Las Vegas, NV, November 17 – 19.
Davis, J. A. 1970. “Clustering and Hierarchy in Interpersonal Relations: Testing Two Graph
Theoretical Models on 742 Sociomatrices.” American Sociological Review 35 (5): 843– 851.
Eagle, N., A. S. Pentland, and D. Lazer. 2009. “Inferring Friendship Network Structure by Using
Mobile Phone Data.” Proceedings of the National Academy of Sciences 106 (36): 15274– 15278.
Ellison, N. B., C. Steinfield, and C. Lampe. 2007. “The Benefits of Facebook ‘Friends’: Social
Capital and College Students’ Use of Online Social Network Sites.” Journal of Computer
Mediated Communication 12 (4): 1143– 1168.
Facebook. 2012. “Facebook Statistics.” Accessed December 2011. http://www.facebook.com/press/i
nfo.php?statistics
Freeman, L. C. 1979. “Centrality in Social Networks. Conceptual Clarification.” Social Networks
1 (3): 215– 239.
20 L. Groeger and F. Buttle

Godes, D., and D. Mayzlin. 2004. “Using Online Conversations to Study Word-of-Mouth
Communication.” Marketing Science 23 (4): 545– 560.
Godes, D., and D. Mayzlin. 2009. “Firm-Created Word-of-Mouth Communication: Evidence from a
Field Test.” Marketing Science 28 (4): 721– 739.
Granovetter, M. 1973. “The Strength of Weak Ties.” American Journal of Sociology 78 (6):
1360– 1380.
Granovetter, M. 1982. “The Strength of Weak Ties: A Network Theory Revisited.” In Social
Structure and Network Analysis, edited by P. V. Marsden and N. Lin, 105– 130. London: Sage.
Hinz, O., B. Skiera, C. Barrot, and J. U. Becker. 2011. “Seeding Strategies for Viral Marketing: An
Empirical Comparison.” Journal of Marketing 75 (6): 55 – 71.
Hogan, B. 2008. “A Comparison of On and Offline Networks Through the Facebook API.”
Proceedings from European Science Foundation for the Quantitative Methods in the Social
Sciences, Amsterdam, The Netherlands.
Holmes, J. H., and J. D. Lett. 1977. “Product Sampling and Word of Mouth.” Journal of Advertising
Research 17 (3): 35 – 40.
Hollebeek, L. 2011. “Exploring Customer Brand Engagement: Definition and Themes.” Journal of
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Strategic Marketing 19 (7): 555– 573.


Huang, L. 2010. “Social Contagion Effects in Experiential Information Exchange on Bulletin Board
Systems.” Journal of Marketing Management 26 (3– 4): 197– 212.
IBISWorld. 2012. “Beer and Malt Manufacturing Market Research Report: ANZSIC c2182.” http://
www.ibisworld.com.au
Jain, D., V. Mahajan, and E. Muller. 1995. “An Approach for Determining Optimal Product
Sampling for the Diffusion of a New Product.” Journal of Product Innovation Management 12
(2): 124– 135.
Kamin, H. 1978. “Advertising Reach and Frequency.” Journal of Advertising Research 18 (1):
21 – 25.
Katz, E., and P. F. Lazarsfeld. 1955. Personal Influence: The Part Played by People in the Flow of
Mass Communications. Glencoe, IL: Free Press.
Katz, H. E. 2010. The Media Handbook: A Complete Guide to Advertising Media Selection,
Planning, Research, and Buying. Abingdon: Taylor & Francis.
Kilduff, M., and D. J. Brass. 2010. “Organizational Social Network Research: Core Ideas and Key
Debates.” The Academy of Management Annals 4 (1): 317– 357.
Kotler, P., K. L. Keller, and S. Burton. 2009. Marketing Management. 1st ed. Frenchs Forest:
Pearson.
Kozinets, R. V., K. De Valck, A. C. Wojnicki, and S. J. S. Wilner. 2010. “Networked Narratives:
Understanding Word-of-Mouth Marketing in Online Communities.” Journal of Marketing
74 (2): 71 – 89.
Krugman, H. E. 1972. “Why Three Exposures May be Enough.” Journal of Advertising Research
12 (6): 11 – 14.
Lazarsfeld, P. F., and R. K. Merton. 1954. “Friendship as a Social Process: A Substantive and
Methodological Analysis.” Freedom and Control in Modern Society 18: 18 – 66.
Lewis, K., J. Kaufman, M. Gonzalez, A. Wimmer, and N. Christakis. 2008. “Tastes, Ties, and Time:
A New Social Network Dataset using Facebook.com.” Social Networks 30 (4): 330– 342.
Liu, Y. 2006. “Word of Mouth for Movies: Its Dynamics and Impact on Box Office Revenue.”
Journal of Marketing 70 (3): 74 – 89.
Mancuso, J. R. 1969. “Why Not Create Opinion Leaders for New Product Introductions?” Journal of
Marketing 33: 20 – 26.
Marks, L. J., and M. A. Kamins. 1988. “The Use of Product Sampling and Advertising: Effects of
Sequence of Exposure and Degree of Advertising Claim Exaggeration on Consumers’ Belief
Strength, Belief Confidence, and Attitudes.” Journal of Marketing Research 25 (3): 266– 281.
Marsden, P. V., and K. E. Campbell. 1984. “Measuring Tie Strength.” Social Forces 63 (2):
482– 501.
Mathwick, C., C. Wiertz, and K. De Ruyter. 2008. “Social Capital Production in a Virtual p3
Community.” Journal of Consumer Research 34 (6): 832– 849.
Mcdonald, C. 1995. Reach and Frequency: Maximizing Advertising Results Through Effective
Frequency. 2nd ed. Lincolnwood, IL: NTC Business Books.
McPherson, M., L. Smith-Lovin, and J. M. Cook. 2001. “Birds of a Feather: Homophily in Social
Networks.” Annual Review of Sociology 57: 415– 444.
Journal of Marketing Communications 21

Moreno, J. L. 1934. Who Shall Survive?: A New Approach to the Problem of Human Interrelations
Nervous and Mental Disease Monograph Series No. 58. Washington, DC: Nervous and Mental
Disease.
Nelson-Field, K., E. Riebe, and B. Sharp. 2012. “What’s Not to ‘Like?’ Can a Facebook Fan Base
Give a Brand the Advertising Reach It Needs?” Journal of Advertising Research 52 (2): 262.
Petróczi, A., T. Nepusz, and F. Bazsó. 2006. “Measuring Tie-Strength in Virtual Social Networks.”
Connections 27 (2): 39 – 52.
Reingen, P. H., and J. B. Kernan. 1986. “Analysis of Referral Networks in Marketing: Methods and
Illustration.” Journal of Marketing Research 23 (4): 370–378.
Rosen, E. 2009. The Anatomy of Buzz Revisited: Real-life Lessons in Word-of-Mouth Marketing.
New York: Crown Business.
Sernovitz, A. 2009. Word of Mouth Marketing: How Smart Companies Get People Talking. Austin:
Kaplan.
Shi, X., L. A. Adamic, and M. J. Strauss. 2007. “Networks of Strong Ties.” Physica A: Statistical
Mechanics and its Applications 378 (1): 33 – 47.
Soup. 2012. “Interview by Authors with Research Director of Agency Soup.” February 18, 2012.
Downloaded by [Dalhousie University] at 01:00 08 August 2013

Sydney, Australia.
Subrahmanyam, K., S. M. Reich, N. Waechter, and G. Espinoza. 2008. “Online and Offline Social
Networks: Use of Social Networking Sites by Emerging Adults.” Journal of Applied
Developmental Psychology 29 (6): 420–433.
Sundaram, D. S., K. Mitra, and C. Webster. 1998. “Word-of-Mouth Communications: A
Motivational Analysis.” Advances in Consumer Research 25 (1): 527– 531.
Trusov, M., R. E. Bucklin, and K. Pauwels. 2009. “Effects of Word-of-Mouth Versus Traditional
Marketing: Findings from an Internet Social Networking Site.” Journal of Marketing 73 (5):
90 – 102.
Van Den Bulte, C., and S. Wuyts. 2009. “Leveraging Customers Networks.” In The Network
Challenge, edited by P. R. Kleindorfer and Y. Wind, 243–258. Upper Saddle River, NJ:
Pearson.
Verhoef, P. C., W. J. Reinartz, and M. Krafft. 2010. “Customer Engagement as a New Perspective in
Customer Management.” Journal of Service Research 13 (3): 247– 252.
Webster, C. M., and P. D. Morrison. 2004. “Network Analysis in Marketing.” Australasian
Marketing Journal 12 (2): 8 – 18.
Zhu, F., and X. Zhang. 2010. “Impact of Online Consumer Reviews on Sales: The Moderating Role
of Product and Consumer Characteristics.” Journal of Marketing 74 (2): 133– 148.

You might also like