See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/358945387

Social Media Analytics and Its Applications in Marketing

Article in Foundations and Trends® in Marketing · March 2022

DOI: 10.1561/1700000073

CITATIONS READS

6 919

2 authors:

Sangkil Moon Dawn Iacobucci

University of North Carolina at Charlotte Vanderbilt University
39 PUBLICATIONS 1,166 CITATIONS 219 PUBLICATIONS 18,982 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Dawn Iacobucci on 07 August 2023.

The user has requested enhancement of the downloaded file.

 Foundations and Trends® in Marketing
Social Media Analytics and Its
Applications in Marketing
Suggested Citation: Sangkil Moon and Dawn Iacobucci (2022), “Social Media Analytics
and Its Applications in Marketing”, Foundations and Trends® in Marketing: Vol. 15, No.
4, pp 213–292. DOI: 10.1561/1700000073.

Sangkil Moon
Belk College of Business
The University of North Carolina at Charlotte
USA
smoon13@uncc.edu
Dawn Iacobucci
Owen Graduate School of Management
Vanderbilt University
USA
dawn.iacobucci@vanderbilt.edu

This article may be used only for the purpose of research, teaching,
and/or private study. Commercial use or systematic downloading
(by robots or other automatic processes) is prohibited without ex-
plicit Publisher approval.
Boston — Delft
 Contents

1 Introduction 215

2 Understanding Social Media Analytics (SMA) 219

2.1 Status of Social Media . . . . . . . . . . . . . . . . . . . 219
2.2 Status of SMA . . . . . . . . . . . . . . . . . . . . . . . . 223
2.3 SMA Frameworks and Processes . . . . . . . . . . . . . . 224

3 Review of Social Media and SMA Related

to Marketing 229
3.1 Review of Social Media Related to Marketing . . . . . . . 229
3.2 Review of SMA Related to Marketing . . . . . . . . . . . 233
3.3 Review of Social Media Networking . . . . . . . . . . . . . 237
3.4 Review of Consumer Privacy and Fake Reviews . . . . . . 242

4 Useful Methods for SMA 245

4.1 Social Media Data as Part of Big Data . . . . . . . . . . . 245
4.2 Artificial Intelligence (AI), Machine Learning (ML),
and Deep Learning (DL) . . . . . . . . . . . . . . . . . . 247
4.3 Text Analytics . . . . . . . . . . . . . . . . . . . . . . . . 252
4.4 Visual Analytics . . . . . . . . . . . . . . . . . . . . . . . 258
5 Trends, Future Research, and Education in SMA 262
5.1 General Trends and Future Research Avenues in SMA . . . 262
5.2 SMA Education in Colleges . . . . . . . . . . . . . . . . . 269

6 Conclusions 273

References 276
Social Media Analytics and Its
Applications in Marketing
Sangkil Moon1 and Dawn Iacobucci2
1 BelkCollege of Business, The University of North Carolina at
Charlotte, USA; smoon13@uncc.edu
2 Owen Graduate School of Management, Vanderbilt University, USA;

dawn.iacobucci@vanderbilt.edu

ABSTRACT
The rise of the Internet and smartphones in the 21st century
has created and developed social media as an extremely
effective means of communication in society. In life, business,
sports, and politics, social media facilitates the democrati-
zation of ideas like never before. Social media content gives
consumers different information sources that they must de-
cipher to discern its trustworthiness and influence in their
own opinions. Marketers must be savvy about using social
media in their attempts to persuade consumers and build
brand equity.
As social media has permeated our everyday lives, scholars
in various disciplines are actively conducting research into
this aspect regarding our way of life. In this scholarly en-
deavor, marketing has taken a leading role in this research
endeavor as a discipline involving human communications
and idea persuasion. Thus, rather than considering social
media broadly across multiple disciplines, in this monograph,
we concentrate on social media analytics in marketing.

Sangkil Moon and Dawn Iacobucci (2022), “Social Media Analytics and Its Appli-
cations in Marketing”, Foundations and Trends® in Marketing: Vol. 15, No. 4, pp
213–292. DOI: 10.1561/1700000073.
©2022 S. Moon and D. Iacobucci
214

This monograph comprises the following four sections:

• First, we provide an overview of social media and social

media analytics (SMA). While much has already been
said about social media generally, relatively less has
been said about social media analytics. Thus, much of
our focus is on SMA in terms of contributing to the
current understanding of SMA in the field.
• Second, we concentrate on social media analytics in
marketing. We discuss practical industry perspectives
and examples, as well as recent marketing research by
academics. Notably, we show how analytics may be
used to address concerns about social media privacy
and help detect fake reviews.
• Third, we summarize common tools for social media
analytics in marketing. These methods can be complex,
but they must be mastered for sound SMA practice.
They encompass big data, artificial intelligence, ma-
chine learning, deep learning, text analytics, and visual
analytics.
• Fourth, we discuss trends and a future research agenda.
We also discuss how SMA might be better integrated
into higher education.
 1
Introduction

The rise of the Internet and smartphones in the 21st century has
created and developed social media as an extremely effective means of
communication in society. In life, business, and politics, social media can
facilitate the democratization of ideas like never before. The COVID-19
pandemic dramatically changed every aspect of human life worldwide,
including the roles of and our increased reliance on social media (Cuello-
Garcia et al., 2020; Yu et al., 2020).
Yet, even prior to the pandemic shifting lifestyles and jobs more
online, consider the following statistics: Techjury estimates that some
3.2 billion people worldwide are active users of social media sites: on
Facebook for 2 hours and 24 minutes a day, YouTube for 40 minutes,
Snapchat for 30 minutes, Instagram for 28 minutes, and Pinterest for 14
minutes a day (Deyan, 2021). Further, the recent incident of Facebook’s
outage demonstrates the severity regarding our reliance on social media
in our everyday lives. The outage on Facebook (renamed Meta) and its
associated platforms (such as Instagram and WhatsApp) lasted only six
or seven hours on October 4, 2021. However, its impacts were significant
and memorable. During this outage, many users migrated to other social
media platforms such as Twitter, Gmail, and Snapchat, which created

215
216 Introduction

slowdowns on those platforms. During the outage, Facebook’s stock

price decreased by nearly 5%, and Facebook is estimated to have lost
at least $60 million in its advertising revenues (en.wikipedia.org/wiki/
2021_Facebook_outage).
Companies and brand teams aim to take advantage of this consumer
time and engagement. Statistica.com estimates that 92% of U.S. com-
panies (with 100 or more employees) expect to use social media for
marketing purposes. Even small neighborhood retailers heavily rely on
their business presence and promotions on social media sites (e.g., Ama-
zon, Yelp, Tripadvisor). Therefore, marketers aim to gain the benefits
of brand exposure, lead generation, fan loyalty, and online traffic on
social media.
Given that social media has permeated our everyday lives, scholars in
various disciplines are conducting research into social media phenomena.
Marketing has taken a leading role in this research endeavor, given its
strengths in understanding how human ideas and communications may
be expressed and exchanged. For example, social media content gives
consumers different information sources that they must decipher to
discern its trustworthiness and influence in their own opinions. Marketers
must be savvy about using social media in their attempts to persuade
consumers and build brand equity.
Rather than offering a view with breadth but not depth on social
media across multiple disciplines, this monograph focuses on the in-
tersection of three closely related domains: marketing, social media,
and SMA (see Figure 1.1). In other words, this monograph concerns
marketing-based decisions related to social media and social media
analytics (SMA) or social media and social media analytics from the
perspective of the marketing discipline. Even social media in marketing
has breadth and depth that are too broad and too deep for a single
monograph. In SMA, we emphasize the roles of data and tools in tack-
ling marketing problems. In marketing applications within the context
of SMA, we discuss not only current problems, but also trends and
future directions. We hope that the basic approach of this monograph
can help both marketing practitioners and academics better understand
how to use marketing analytics to identify the common patterns and
trends produced by consumers and firms (Iacobucci et al., 2019).
 217

Social Media

Social Media Analytics (SMA)

(Data and Tools)

Current Marketing Applications

Trends and Future in

Marketing in SMA

Figure 1.1: The scope of this monograph at the intersection of social media, SMA,
and marketing.

To this end, the monograph is organized as follows. First, we describe

the general state of social media and SMA. We offer both practical
and theoretical perspectives. In particular, we propose a four-step SMA
framework called AAVF (Acquisition-Analysis-Value-Feedback) based
on our examination of some existing frameworks and processes for SMA.
Second, we examine social media analytics (SMA) particularly within
marketing as a collection of tools and methods used to unearth the
practical values of consumer and brand data. Plenty has been said
about social media, but we endeavor to provide more insights into
SMA. We focus on SMA in marketing or marketing in SMA to tackle
marketing questions and obtain insights about consumers and brands.
218 Introduction

As an example, we discuss how SMA has been used to detect fictitious

reviews on products out of a large volume of such product reviews
posted every day.
Third, we show how common methods for SMA have been utilized
for common marketing problems. Some common SMA methods encom-
pass the analysis of big data, text analytics, visual analytics, artificial
intelligence (AI), machine learning (ML), and deep learning (DL).
Finally, we discuss trends and future research topics for SMA, par-
ticularly within marketing. We also explore how SMA is being taught
in colleges and universities to educate students interested in SMA. We
hope that this emphasis can help illuminate how social media may be
used for the greater good of society.
 2
Understanding Social Media Analytics (SMA)

To understand SMA, we must understand the object of the analytics:

social media. Therefore, in this section, we offer a status report of
social media, and then a status report of SMA. In doing so, we take
both practical and theoretical perspectives. Specifically, we explore how
organizations make efforts to utilize social media data; moreover, we
examine SMA from the perspective of the marketing literature. Drawing
upon these perspectives, we present our four-step SMA framework called
AAVF (Acquisition-Analysis-Value-Feedback).

2.1 Status of Social Media

Let us begin by showing how several experts define “social media.”

(While “media” is plural, “social media” is frequently referred to as
singular, and we adopt that convention.) The term “social media” was
initially used around 2004 to describe content and Internet-based appli-
cations that can be continuously modified by users through participation
and collaboration, rather than traditionally published content created,
prepared, and published by individuals or corporate entities without
opportunities for subsequent revision (Ghani et al., 2019; Kaplan and
Haenlein, 2011). Such web applications facilitate interactive information

219
220 Understanding Social Media Analytics (SMA)

sharing and interoperability, thus allowing users to edit and add to

website content in an open manner (Sinha et al., 2012).
The development of hardware and software to utilize social media
platforms produced a host of user-generated content (UGC). Ellison
(2007) listed three aspects of UGC to explain social media. First, users
possess the ability to create their public profiles. Second, users should
be able to connect with others to construct a network. Third, users are
allowed to view other users’ activities and interact with them.
Khan (2018) uses a similar philosophy to define social media as an
easy-to-use Internet-based platform that provides users with opportuni-
ties to create and exchange content (such as text, photos, and videos)
in one-to-one, one-to-many, and many-to-many contexts. In particular,
a social networking service is a type of social media platform focused on
online social relationships among users (e.g., Facebook and LinkedIn).
These allow users to build and maintain personal and professional rela-
tionships among people who share interests, activities, opinions, and
information.
These definitions converge on the common user experience. That is,
to most users, social media platforms are online communities through
which members seek and share common interests, activities, experiences,
and information (Lee, 2018). Consider the following examples:

• Content economy communities such as YouTube and Meta (the old

Facebook) are defined by multiple users interacting online around
an object of common interest (e.g., discussion topics, hobbies).

• Blogs present discussions in written, image, or audio forms, often

authored by someone with deep content knowledge. A microblog
is the type of a blog that enables smaller chunks of content to
be posted in comparison with a regular blog (Hayat et al., 2019;
Kaplan and Haenlein, 2011). Instant messages and emails are
considered as effective ways to keep such blogs active and current.

• LinkedIn is a business and employment-oriented social networking

service; its main strength lies in connecting hiring organizations
and job applicants.
2.1. Status of Social Media 221

• Wikipedia collects and organizes online collaborative encyclopedic

projects on almost any topic that people might be interested in
learning and discussing.

Marketing managers are intrigued by social media as a marketing

platform. Hardly a day goes by that traditional news media does not
mention Twitter, Instagram, or YouTube, often within the context
of some new way to reach customers and create buzz about a brand
(Ma et al., 2015). Social media marketing requires a content marketing
strategy, a way to systematically produce interesting content that people
will enjoy enough to “hang out” on social media, and therefore be
exposed to and potentially influenced by brand messaging (McDonald,
2021).
Thus, social media offers marketers an unparalleled opportunity to
promote products and enhance brand loyalty through networking with
their customers. Accordingly, social media has affected the traditional
form of marketing research in terms of how companies can better
understand their customers. Before social media, marketers used labor-
intensive and time-consuming marketing research methods such as
surveys and focus group interviews for efficient product and brand
management. Nowadays, consumers initiate their actions on social
media to post their opinions publicly (Moe et al., 2017).
For example, Ahuja and Alavi (2018) studied Jet Airways’ presence
on Facebook. Jet Airways, a popular airline company in India, is a
multinational corporation that emphasizes good customer experience
management. Ahuja and Alavi (2018) demonstrated how Jet Airways
takes advantage of Facebook to cultivate consumers’ trust. They em-
pirically demonstrated that high frequencies of likes and comments
pertaining to relational content influence customer trust development.
There are several noteworthy gurus who describe the strengths of
social media for marketers. For example, McDonald (2021) asserts that
marketers rely on social media for five main reasons: (1) social media is
big; (2) your customers are on social media; (3) social media is mostly
free; (4) social media can reach not only existing but also new cus-
tomers, and (5) social media offers promotion/advertising opportunities.
222 Understanding Social Media Analytics (SMA)

Similarly, with their analysis of social media data, Moe and Schweidel
(2017) indicated three major functions of social media: (1) social media
as a source of consumer insights; (2) SMA as a group of analytics tools
that can complement traditional tools to reveal consumer insights; and
(3) social media as a way to help marketing actions be more proactive
and timelier.
To know whether social media messages are reaching and influencing
their customers, companies should implement the basics of a marketing
strategy: (1) choose suitable media platforms according to the target
market; and (2) develop a content strategy (Ahuja and Alavi, 2018).
The content strategy variables, such as the content type, posting agility,
and content context (e.g., number of likes and comments), must be
measured by organizations on a regular basis (Chauhan and Pillai,
2013). Similarly, in understanding social media, Kietzmann et al. (2011)
focused on seven functional building blocks of social media: identity,
conversations, sharing, presence, relationships, reputation, and groups,
with each one examining a specific aspect of the social media user
experience and its implications for organizations.
All brand marketers want people to talk about their brand in a
positive way, and today much of that conversation occurs on social
media. As marketers, a common goal for social media is to grow and
nurture positive electronic word-of-mouth. On the other hand, social
media is apparently complicated and more difficult to conquer. That
is, using social media is one thing, while marketing on social media is
another. Even many experienced marketers find that their knowledge
of traditional marketing channels does not translate easily to social
media networks. Great social media marketing efforts create an illusion
because they look spontaneous and effortless. However, behind the
scenes, marketers devote enormous amounts of work to manage, promote,
and grow their brands on these social media platforms (McDonald,
2021).
Given these challenges, SMA can help marketers enhance the efficacy
of their social media usage. We turn to this topic next.
2.2. Status of SMA 223

2.2 Status of SMA

Based on Google’s trends data, the term “social media analytics” (SMA)
seems to have appeared on the Internet horizon in July 2006 (Khan,
2018). Since then, interest in the term has consistently grown. We
define SMA as the art and science of extracting valuable information
from vast amounts of structured and unstructured social media data to
enable informed and insightful decision making. SMA is a science, as
it involves systematically identifying, extracting, and analyzing social
media data using sophisticated tools and techniques. Concurrently, it is
an art with regard to interpreting results and aligning insights obtained
with business objectives and goals.
To produce value from SMA, analysts attempt to understand SMA
as both an art and a science. The science aspect of SMA requires
skilled data analysts, sophisticated tools and technologies, and reliable
data. Further, to adequately understand results and put them into
action, analysts must also master the art of interpreting analytics,
which involves representing data in meaningful ways with sufficient
domain-specific knowledge. It is essential that a company’s efforts in
SMA be strategically aligned with its business objectives.
Social media data are inherently an invaluable source of hidden
business insights (Khan, 2018). Organizations can realize the potential
in utilizing social media data by identifying which consumer behaviors
create value. To gain competitive advantage, organizations seek to
monitor and analyze not only their customer-generated content, but also
their business partners’ and competitors’ customer-generated content
on various social media sites. For example, Lee (2018) showed how SMA
can improve supply chain efficiency and effectiveness with business
partners.
Such massive amounts of unstructured data naturally give rise to
challenges and opportunities for data analysis (Chang et al., 2019).
One question is simply, how does a company begin? Like any other
business investment, analytics investment is a difficult decision to make.
To facilitate this decision, Verhoef et al. (2015) stated that the decision-
making process unfolds over three stages: (1) investment (wherein there
224 Understanding Social Media Analytics (SMA)

Business Marketing, Social media’s

strategic CRM, and brand impact on firm
planning development performance

Social media’s Social media and

customer network
engagement analytics

Figure 2.1: Organizational use of social media and SMA.

is enthusiasm for new analytics efforts); (2) frustration and disinvest-

ment (due to disappointment regarding early SMA efforts); and (3)
reinvestment. In the third phase, as companies gain knowledge from
past trial and error, they can refine their analytics expectations and
establish more realistic analytics investment strategies.
Thus, from a strategic standpoint, how to manage social media
and SMA is becoming increasingly more important. Figure 2.1 shows
a view of how companies can manage social media and its associated
SMA in an organized manner. In this view, social media’s impact on
firm performance is evaluated based on social media activities (CRM
and brand development activities), consumer engagement, and existing
social media network structures.

2.3 SMA Frameworks and Processes

Table 2.1 presents several frameworks that have been offered as guidance
for understanding SMA processes. All these frameworks are intended to
help organizations achieve their SMA objectives such as describing data,
analyzing trends, predicting future problems and opportunities, and
optimizing business processes. We draw from these frameworks to build
a parsimonious (simpler yet more comprehensive) framework of SMA to
integrate these perspectives in Figure 2.2. We provide a straightforward
framework based on practical problem solving and associated theories.
On the surface, the processes and frameworks introduced in Table 2.1
appear diverse and different. Therefore, we aimed to derive a more
2.3. SMA Frameworks and Processes 225

Table 2.1: Current frameworks delineating the social media analytics (SMA) process

Source Social Media Analytics (SMA) Process

Fan and Gordon (2015) Capture (input) stage: Acquire relevant social media data by
monitoring sources of social media platforms, store relevant
data in an organized manner, and extract necessary
information from the collection. Understand (process)
stage: Select pertinent data for analysis. Present (output)
stage: Display analysis findings in a meaningful way.
Holsapple et al. (2018) (1) Attention: Search for intelligence gathering opportunities,
insight generation, and decision making.
(2) Acquisition: Gather relevant information from both
internal and external sources.
(3) Design: Develop multiple courses of action according to
available information.
(4) Evaluation: Assess the value of alternative courses of
action; construct a set of possible actions.
(5) Choice: Resolve conflicts in the set of possible actions
and choose the best course of action.
(6) Implementation: Execute the chosen action plan.
(7) Feedback and review: Obtain information about the
outcomes and objectively evaluate these outcomes.
(8) Learning and behavior modification: Gain new
managerial and operational insights based on #7 and
reiterate more of phases #1∼#6, if necessary.
Khan (2018) (1) Identification of the data and source platforms
(2) Extraction
(3) Cleaning
(4) Analyzing
(5) Data visualization
(6) Interpretation
Lee (2018) (1) Develop key performance/evaluation metrics.
(2) Choose social media platforms that generate data and
monitor/listen to those platforms.
(3) Analyze data using a variety of tools.
(4) Build social media intelligence.

abstract, higher-level SMA framework to simplify and integrate across

these processes and concepts. Our framework is intended to guide SMA
analysts and researchers through the process of accomplishing practical
business objectives in an effectively organized fashion. We build on Fan
and Gordon’s (2015) popular CUP process of Capturing, Understanding,
and Presenting social media data, analysis, and results (Andryani et al.,
2019; Chang et al., 2019; Ribarsky et al., 2014), to which we add a
226 Understanding Social Media Analytics (SMA)

Acquisi�on

Feedback Analysis

Value

Figure 2.2: The AAVF cycle framework for SMA.

phase for feedback and learning (Einhorn and Hogarth, 1981). Therefore,
we propose this four-step framework for SMA called AAVF (refer to
Figure 2.2):

(1) Acquisition: This stage involves acquiring necessary data

and information from relevant social media platforms. It
includes web-scraping techniques by utilizing application pro-
gramming interfaces (APIs). The Big Data era has allowed
analysts to access various kinds of information efficiently
and oftentimes freely. It is widely known that analyses and
the value created by these analyses cannot surpass the qual-
ity or relevance of the data. Therefore, it is important to
know what relevant and proper data are acquired and how.
By improving the “how” part, the data collection step can
be more efficient. Thus, data acquisition should be part of
the cyclical learning process. The importance of the data
relevance is emphasized in all the frameworks in Table 2.1.
In particular, Fan and Gordon (2015) emphasize the im-
portance of monitoring data sources for better and more
relevant data acquisition. Proper data acquisition methods
2.3. SMA Frameworks and Processes 227

can facilitate data reorganization for analysis in the next

step.
(2) Analysis: Using various techniques such as machine learn-
ing methods, text analytics, and visualization techniques,
marketers can develop and implement analysis plans. In the
process, analysts should be able to select the proper data
and reorganize them for analysis with managerial input. The
analysis requires proper intuition and domain knowledge
from both managers and analysts.
(3) Value (through interpretation): The ultimate purpose
of SMA is to create value for decision makers in the social
media field (e.g., developing effective social media promo-
tion campaigns and new products that reflect ever-changing
consumer preferences). This process may include how best
to present the analysis objectives and results to decision
makers by utilizing visualization techniques. When the data
are presented effectively, decision makers are more likely to
adopt the analysis results. This effort will make the entire
analytics process valuable to the organization using this
process.
(4) Feedback (and learning): One project can be concluded,
but the process may never end because organizations typ-
ically intend to last forever. In the process, analysts and
decision makers will naturally learn and improve. On the
other hand, because consumers and the market environment
are everchanging, this analysis process should continue re-
flecting those market changes not only to stay with the
market, but also to succeed and prosper. Therefore, learning
during the process is crucial to subsequent projects tackling
similar business problems. Because analysts and decision
makers cannot focus on how to improve the process during a
single project, this step is important as a separate learning
step in this cyclical analysis process. The importance of this
step is particularly emphasized by Holsapple et al. (2018).
228 Understanding Social Media Analytics (SMA)

In using our proposed process as an SMA framework, we empha-

size flexibility (Simon, 1977). In this particular application, flexibility
implies the following: (1) while we depict these phases as occurring
sequentially in a cycle, multiple phases may be undertaken in a parallel
and interactive manner; (2) analysts may also move back and forth
between phases in a flexible manner; and (3) the operation of any stage
will likely entail decision making.
Next, we drill down from an abstract discussion of social media and
analytics to examine more specific applications in the marketing field.
 3
Review of Social Media and SMA
Related to Marketing

In this section, we discuss marketing problems that may be addressed

using SMA in the space of social media. For example, marketers often
wish to identify influential opinion leaders on social media, which can
be done with an understanding of networking patterns among social
media users. As we discuss social media and SMA in this section, we
introduce a variety of common marketing problems such as consumer
behavior on social media and social media’s impact on firm performance,
business/marketing strategy, and product/brand management. In par-
ticular, we further highlight social media network analysis, consumer
privacy and data security on social media, and fictitious/biased content
on social media.

3.1 Review of Social Media Related to Marketing

Financial Performance: A common and fundamental business question

involves the profitability of a business action. On social media, we might
posit that if Chipotle Mexican Grill offers a $2 coupon on its website,
it would likely get a good response from its customers. However, how
valuable would that be to the company in the long term? Would such
a campaign drive long-term sales by inducing new customers? ROI is

229
230 Review of Social Media and SMA Related to Marketing

the gain from an investment minus the cost of an investment. When

companies must evaluate their return-on-investment (ROI) effectiveness,
marketing professionals must determine the duration of a social media
marketing campaign and what milestones they need to attain. Similarly,
when an online influencer makes a recommendation, marketers need a
benchmark to measure the influence. Although social media platforms
are expressions of data collections from various angles, no industry-wide
accepted standards and practices have emerged (Sponder, 2012).
In response to this criticism, and with an eye on financial growth,
Rust et al. (2021) developed a new social media-based brand reputation
tracker by mining Twitter comments for the world’s top 100 brands
using Rust–Zeithaml–Lemon’s value-brand-relationship framework, on
a weekly, monthly, and quarterly basis. Their study demonstrated that
brand reputation can be monitored in real time and longitudinally;
moreover, it can be connected to firm financial performance.
Brand Equity: Social media value to firms includes all forms of
financial and nonfinancial assets that firms produce from social media
investment and engagement (Khan, 2018). Financial value to firms from
social media can come in a variety of forms such as sales growth and
cost reduction. Further, financial value to firms from engaging via social
media is not immediately clear. For example, a customer reading a
product review on Amazon or clicking on a social advertisement may
not purchase the product right away. However, reviews, views, likes,
and clicks can raise brand awareness and create more interest in the
product, which may ultimately lead to increased sales. Thus, the value
creation from social media is complicated and may usually take a while
to materialize.
In addition to financial gains, engaging via social media creates intan-
gible value for brands. Research suggests that social media engagement
has a significant positive impact on brand equity—a brand’s reputation,
trust, and loyalty. Intangible value to firms, which comes from social
media, includes (Kim and Ko, 2012): (1) gaining insights into customer
values and behaviors; (2) tracking the impact and effectiveness of social
media campaigns; and (3) understanding brand reputation and engage-
ment purposes, along with customers’ values, relationships, and brand
equities. Advertisements on Facebook can produce instant intangible
3.1. Review of Social Media Related to Marketing 231

value in the form of brand awareness through views and shares, which
can increase website traffic. In this case, the effect on brand awareness
will clearly depend on the content quality of the advertisement.
Consumers expect to gain specific values and benefits from their
social media engagement. Social media values to consumers cover various
kinds of financial and non-financial benefits that social media users
desire to obtain from engaging with their favorite brands. For example,
in focus group interviews at The University of Waikato in 2017, Khan
(2018) asked a group of 56 participants to name possible financial
benefits they expect to gain from their brand engagement on social
media. In their responses, 90% said that they expect to get some
types of discounts, better deals, promotions, and competitive prices.
During the same focus group interview, participants were also asked
to identify the nonfinancial value that they expect to receive from
social media engagement. This time, their answers included brand
awareness, belonging to the community, obtaining the latest information
and news, faster responses to customer complaints and questions, and
easier comparisons of product qualities and features with those of
competitors.
Social Media Measures: Marketing problems have been identified
by marketers as important and common tasks in the space of social
media, giving them valuable information in terms of consumers’ impres-
sions, sentiments, and satisfaction toward their brands. Furthermore,
the activation of customers’ influences based on the referrals and ad-
vocacy activities for products and services constitutes the marketing
objective of a positive online word-of-mouth promotion policy (Misirlis
and Vlachopoulou, 2018).
Businesses and their brands are increasingly measuring social media
metrics such as engagement and word-of-mouth (Lipsman et al., 2012),
which in turn has resulted in marketing strategies that incorporate
social media (Aswani et al., 2017). At the same time, many brands are
seeing a declining ROI in traditional media, along with demographic
shifts toward young target customers and low-cost marketing campaigns
(Gillin and Moore, 2007). As a result, brand strategies are increasingly
conceived and executed for social media (Tsimonis and Dimitriadis,
2014).
232 Review of Social Media and SMA Related to Marketing

Promotions: Moe and Schweidel (2017) characterized the academic

research on social media as focusing primarily on its role as a promotional
tool (Fossen and Schweidel, 2017; Stephen and Galak, 2012; Trusov
et al., 2009) or as a source of marketing research. Evidently, social media
marketing is changing the way organizations promote their products
(De Vries et al., 2012). Many organizations see social media as a new
and effective channel to interact with their stakeholders. In this sense,
they accept social media as a critical strategy tool to deliver their brand
messaging to their target consumers.
As a negative consequence of social media growth, some organiza-
tions engage in undesirable behaviors such as spamming social media
platforms, without the message recipients’ consent. Often, intention-
ally manipulated content misleads honest consumers, which results in
strengthening negative public opinion of social media (Aswani et al.,
2017). There are other issues such as unethical outsourcing to generate
user-generated content, spamming, and promoting fake content in the
name of online consumer reviews (Luca and Zervas, 2016). These unde-
sirable trends make the detection of spam and manipulated or fictitious
reviews even more important and worthy of investigation.
Marketing Research: A unique aspect of social media, unlike tra-
ditional promotional media, is that it allows for and facilitates very
efficient two-way communications between companies and customers, as
well as peer-to-peer communications (Lee and Bradlow, 2011; Schweidel
and Moe, 2014; Tirunillai and Tellis, 2014). This aspect has led organi-
zations to collect, analyze, monitor, and act upon social media data in
the name of marketing and social media research.
Thus, it is clear that many marketing activities and functions may
be facilitated through social media and SMA. In a literature review,
Misirlis and Vlachopoulou (2018) found that scholars referred most
frequently to customer engagement, consumer behavior research, and
relationship marketing as important marketing research objectives that
social media enhances (see Table 3.1). They also classified the types of
media platforms that scholars have most frequently examined: social
networking sites and microblogs.
3.2. Review of SMA Related to Marketing 233

Table 3.1: Marketing objectives and social media platform types

Item References
Marketing Objectives
Awareness and branding Kim and Ko (2012)
Engagement Malthouse et al. (2013)
eWOM advertising and promotion Stephen (2016)
Predictive marketing research Qiu et al. (2014)
Consumer behavior research Mostafa (2013)
Social capital value, business equity, and ROI Fan and Gordon (2015)
Relationship marketing, CRM, and social CRM Ma et al. (2015); Osborne and
Ballantyne (2012)
Types of Social Media Platforms
Social networking sites Kim and Ko (2012)
Blogs Paek et al. (2013)
Microblogs Mostafa (2013)
Content communities and video-sharing sites Carim and Warwick (2013)

3.2 Review of SMA Related to Marketing

In this section, we focus on SMA in marketing. We have been discussing

the great potential for firms to extract business intelligence from social
media data bearing on important business applications. In particular,
social media platforms provide an enormous public repository of textual
data from which valuable information can be extracted. For example,
Hu et al. (2019) developed a text analytics framework that integrates
different sources of social media data to measure brand personality.
Their social media data included self-descriptions of approximately
2 million consumers who followed some brands on social media with
more than 300,000 employee reviews and approximately 700,000 brand
official tweets. They found that the profiles of consumers who followed
the brands were an important predictor of brand personality. They also
showed a link between brand personality as judged by employees and
consumers.
Further, research has examined how marketing strategy can inte-
grate SMA (Stephen and Galak, 2012). Moreover, given the rapidly
changing technological advances of SMA, organizations are required to
adopt a more flexible marketing strategy to respond and adapt to the
234 Review of Social Media and SMA Related to Marketing

developing environment. For instance, real-time analytics implementa-

tion to social media data will likely be more accurate than historical
data for prediction, and linguistic analytics may help reveal hidden,
underlying mechanisms, all to help strategic decision making.
Wang et al. (2020) delved into the output of SMA used in marketing
research. They found that firms’ tracking of consumer behavior was the
most researched topic, e.g., analyzing consumers’ purchase decisions,
experiences, and satisfaction. Another common theme was tracking
e-WOM, its antecedents, and identifying key influencers. SMA is also
utilized for product management (Toubia and Netzer, 2017), branding
(Kumar et al., 2016), personalized marketing (Ghose et al., 2012),
pricing (Chevalier and Goolsbee, 2003) and sales (Moon et al., 2014).
We summarize various applications of SMA in marketing problems in
Table 3.2.
Consumer-generated textual data and platform-based data from
third-party reviews are a major source for understanding consumer atti-
tudes and opinions (Ordenes et al., 2017). Beyond business-to-consumer
research, some studies have used consumer-generated textual data to
extract useful information about the entire industry, including com-
petitors (Netzer et al., 2012) and to understand brand positioning and
create product/brand maps (Moon and Kamakura, 2017). Further, for
business-to-business contexts, research has focused on improvements in
digital marketing and online communications (Järvinen and Karjaluoto,
2015). This research demonstrates the ability of big data and social
media analytics to enable B2B organizations to become profitable and
remain sustainable (Sivarajah et al., 2020).
Social media data are a naturally rich source of information for
research on Customer Relationship Management (CRM). Ma et al.
(2015) demonstrated how customers’ past social media behavior can be
used to target prospective customers. Coupled with customers’ other
touchpoints regarding the brand (including transactions and exposure to
other marketing activities), SMA for CRM can help take a more holistic
view of a customer’s relationship with a brand (Moe and Schweidel,
2017).
3.2. Review of SMA Related to Marketing 235

Table 3.2: Summary of SMA applications in marketing

Application Area Subareas Selected Studies

Consumer Behavior eWOM; review posting; social
on Social Media media reviewers’ motivation
and characteristics;
crowdsourcing; J-shaped rating
distribution;
positivity/negativity bias
Social Media Impact Market structure and
on Firm product/brand position; sales
Performance forecasting; stock price and
firm value changes by social
media activities; ecommerce
sales conversion; web traffic
management; review volume
and polarity
Business/Marketing Marketing actions on social
Strategy media; social media promotion
campaigns; managerial
responses to consumers on
social media; digital marketing
strategy; consumer
segmentation and profiling
Product/Brand Online brand communities;
Management product positioning maps using
product reviews; product
innovation; raising brand
awareness; brand reputation
monitoring
Social Media Network structure (strong and
Network Analysis weak ties); member relationship
analysis (voting, tagging,
comments); social tagging;
influential reviewers; reviewers’
dynamic growth and learning
Consumer Privacy Consumers’ consent to personal
and Data Security data collection; consumers’
on Social Media reaction to data breach; privacy
laws
Fictitious/Biased Fake product reviews; content
Content on Social manipulation by firms; bias in
Media social media representation
Berger and Milkman (2012),
Hayat et al. (2019), Kaplan and
Haenlein (2011), Khan (2018),
Kietzmann et al. (2011), Lee
(2018), McDonald (2021), Sinha
et al. (2012)
Chevalier and Mayzlin (2006),
Godes and Mayzlin (2009),
Gopinath et al. (2014), Kim
and Ko (2012), Moe and
Schweidel (2017), Moe and
Trusov (2011), Moon et al.
(2014), Pauwels et al. (2016)
Aswani et al. (2017), Ma et al.
(2015), Stephen and Galak
(2012), Tsimonis and
Dimitriadis (2014), Verhoef
et al. (2015)
De Vries et al. (2012), Ho et al.
(2012), Lovett et al. (2013),
Netzer et al. (2012), Rust et al.
(2021)
Ahuja and Alavi (2018), Fan and
Gordon (2015); Godes et al.
(2005), Iyengar et al. (2011),
Lee (2018), Schweidel and Moe
(2014), Wang et al. (2020)
Michaelidou and Micevski (2019),
Oulasvirta et al. (2014),
Schneider and Iacobucci (2020)
Chen et al. (2016), Dellarocas and
Wood (2008), Luca and Zervas
(2016), Moon et al. (2021)

Although social media data present methodological challenges, such

data can be particularly useful in assisting product development deci-
sions. For example, product attributes and their relative importance
to consumers can be identified from product ratings and review data
(Lee and Bradlow, 2011). They can also enable marketers to target
236 Review of Social Media and SMA Related to Marketing

suitable market segments with which different attribute configurations

may resonate. The competitive landscape can be analyzed to identify
potential new product development opportunities (Netzer et al., 2012).
While firms may not have direct control over the social media conver-
sations among consumers, their marketing actions can have an impact
on these conversations. One way that marketing actions can wield influ-
ence is to stimulate social media activities (Fossen and Schweidel, 2017;
Pauwels et al., 2016). This way, social media may serve to strengthen
other marketing activities, as well as to act as a persuasive channel on
its own.
Lovett et al. (2013) compared online and offline word-of-mouth. They
found that social and functional drivers are more prominent in online
word-of-mouth, whereas emotional drivers were important in offline
word-of-mouth. Anderson (1998) found that offline word-of-mouth was
predominantly negative, as dissatisfied customers were more likely to
engage in word-of-mouth activities; yet, online word-of-mouth tends to
be predominantly positive. This positivity has been reported in multiple
studies (Chevalier and Mayzlin, 2006; Godes and Mayzlin, 2004). One
visible finding in product reviews (a form of online word-of-mouth) is
the J-shaped relationship between satisfaction with the product and
the frequency of posts (Moon et al., 2014). While the positivity and
negativity of online product ratings have been well documented, no
explanation has yet been provided to explain the differences between
online and offline word-of-mouth behaviors (Moe et al., 2017).
Marketing researchers have shown that online product ratings and
reviews tend to show a downward trend over time. Li and Hitt (2008)
suggested that early and later buyers look for different product features
(Li and Hitt, 2008). As the number of ratings increases, later buyers
may have more difficulty sifting through more posted reviews (Godes
and Silva, 2012).
Social media has created a major power shift between consumers
and firms. It provides consumers with a means by which to easily
and publicly express their liking and disliking for products. Consumers
frequently consult product reviews posted on social media before making
a purchase. Accordingly, the content of social media data can be used for
sales prediction (Chevalier and Mayzlin, 2006; Godes and Mayzlin, 2004;
3.3. Review of Social Media Networking 237

Moe and Trusov, 2011; Moe et al., 2017; Moon et al., 2014). However,
demonstrating causality can be challenging. To overcome this challenge,
researchers have used vector autoregressive models (Stephen and Galak,
2012), econometric models (Mayzlin et al., 2014), and field experiments
(Godes and Mayzlin, 2009).
In the space of SMA, social media network analysis has gained
sufficient attention. Because this topic is key to understanding the
functions and structures of social media related to marketing, we allocate
the following subsection to this networking issue. Finally, we complete
this section reviewing social media and SMA related to marketing by
discussing two topics with great societal implications: consumer privacy
and fake reviews.

3.3 Review of Social Media Networking

An old Fabergé television commercial featured actress Heather Locklear,

who was so pleased with her shampoo that she was going to tell her
friends about it, and they would tell their friends, “and so on, and so
on.” The advertising cleverly expresses how today’s word-of-mouth and
social media work through the mechanism of social networks. Infor-
mation propagation and dissemination can be a two-edged sword in
today’s social media-reliant society: not only correct information, but
also misinformation can spread quickly through social media networks.
Given the importance of social media networking in the SMA domain,
we now discuss social network structures and how information and mis-
information are conveyed. Further, we focus on social media networking
related to marketing because we consider the area to be an essential
and primary function of social media.
Iyengar et al. (2011) stated that when examining social influence
with regard to message diffusion, one must consider the social network’s
structure—who tells whom what? On social media, one’s social network
connections will dictate the content to which individuals are exposed.
When analyzing social media activity, users’ positions within the social
network influence the outcomes of activities such as the volume of their
messaging and the extent to which they are perceived as influential.
238 Review of Social Media and SMA Related to Marketing

A social networking site is an online domain where a user can create

a profile and build a personal network that connects him or her to other
users (Ahuja and Alavi, 2018). Social networking sites are web-based
services that allow individuals to: (1) construct a public profile to share;
(2) deliver their messages to other users connected with them; and (3)
view and traverse their list of connections and those made by others in
the system. The nature of these connections may vary from site to site,
but commonly in a social network, nodes are individuals, and ties are
the relationships between these individuals (Scott, 2017). Social network
analysis provides visual and mathematical analyses of user relationships
in a network by modeling social network dynamics and growth (e.g.,
network density, network centrality, network flows) (Lee, 2018).
The venue itself matters, of course. Schweidel and Moe (2014) used
data from various social media sites and found that the topics discussed,
sentiments expressed, and dynamics shown varied greatly across sites.
Social media users naturally have different motivations to use social
media platforms that are more customized to themselves (Karahanna
et al., 2015).
Humans are social beings and have the need to belong to groups
with shared norms, values, and interests (Gangadharbatla, 2008; Inkpen
and Tsang, 2005). Online interactions offer benefits for participants such
as an increased sense of community and social support (Oh et al., 2014).
Analogously, online social networks can be used by organizations to
generate trust among participants and build social relationships (Brown
et al., 2007; Lin et al., 2014). For example, a brand can create an online
social network to elicit participation from customers who can benefit
from online interactions; likewise, a brand can translate into long-term
gains for an organization in terms of greater customer loyalty. The
marketing effectiveness of social media provides immense opportunities
to organizations (Fossen and Schweidel, 2017; Leung et al., 2015).
Wang et al. (2020) highlighted two different types of network an-
alytics: semantic network analytics (which focus on product/brand
relationships; Toubia and Netzer, 2017) and social network analytics
(which focus on user relationships; Ahuja and Alavi, 2018). To marketers,
the two are interconnected. Relationships connecting two people may
be direct, or they may be revealed from indirect behaviors that create
3.3. Review of Social Media Networking 239

relationships such as commenting and tagging. Research has shown that

a “nudge” in the form of recommendations from friends appears to be
influential in changing online shopping behavior (Harris and Dennis,
2011). Social media brand communities offer brand owners the ability
to enhance brand value, trust, loyalty, and feelings of community among
members (Laroche et al., 2012).
Social network analysis is a large discipline unto itself, with numerous
texts (e.g., Borgatti et al., 2018; Iacobucci, 1996; Scott, 2017; Wasserman
and Faust, 1994), journals (e.g., Social Networks), and even software
packages (e.g., UCINET, analytictech.com) dedicated to developing
analytical tools to help identify patterns of interconnections within
social network data. Social networks are modeled at the micro level of
the actors or nodes in the network, at the sub-group level such as cliques
of actors who are interconnected, and at the most macro level when
comparing the density of connections in one network to those in another
(Iacobucci and Hopkins, 1992). Actor-level analyses (e.g., measures of
centrality) are especially popular, as marketing seeks to locate influential
actors or opinion leaders whose connections may suggest an important
role in diffusing attitudes or actions, including purchasing. Other tools
within social network analysis can be used to detect sub-communities
within a larger online community, allowing for greater precision in
customizing products and marketing materials. Social network analysis
is also useful in predictive modeling, as in marketing campaigns aimed
at consumers most likely to buy a particular product (Fan and Gordon,
2015). Some SMA marketers use these tools to locate network structures
and embedded actors, whereas other SMA marketers are more focused
on whether the number of posted likes or comments provides greater
utility in forecasting subsequent attitude or behavioral change.
Social media is capable of rapidly disseminating not only good-
quality information beneficial to society and individuals’ lives, but also
unsubstantiated rumors and conspiracy theories that often result in
significantly negative consequences. Recently, a number of social science
studies have illuminated such information and misinformation prop-
agation on social media. For example, based on how Facebook users
exchange information pertaining to two distinct types of narratives –
scientific and conspiracy news – Del Vicario et al. (2016) showed that
240 Review of Social Media and SMA Related to Marketing

information related to these two types of narratives generates homoge-

neous and polarized communities (that is, echo chambers) that reveal
similar information exchange patterns. In the context of information
dissemination, social media research has attempted to explain why some
organizations succeed in producing social media messages that inspire
far-reaching conversation among social media users. Bail (2016), who
focused on advocacy organizations, found that these organizations are
more likely to produce a large volume of comments on social media if
they create “cultural bridges” or generate messages that combine rarely
discussed conversational themes. Furthermore, using a varied sample of
social media communications with regard to polarizing issues in public
policy debates (such as climate change and same-sex marriage), Brady
et al. (2017) found that moral-emotional political messages substantially
increase in propagation within and beyond ideological group boundaries.
This finding can potentially offer insights into how a variety of marketing
messages can disperse within and beyond subgroups (such as online
brand communities).
Social media has various degrees of content in terms of information
accuracy quality. Given the propagation efficiency of social media, reduc-
ing users’ exposure to low-quality content and misinformation is impera-
tive. One way of doing so is to use rating algorithms to expose relatively
low-quality content from untrustworthy sources. Using news sources
on social media, Pennycook and Rand (2019) found that laypeople are
quite good at distinguishing between lower- and higher-quality sources.
They concluded that incorporating such quality ratings of laypeople into
social media rating algorithms may be effective against misinformation
and news content with extreme political bias. In fact, Facebook CEO
Zuckerberg has repeatedly called on the U.S. government to regulate
election manipulation through social media because democracies remain
vulnerable to foreign and domestic attacks on social media. In the same
vein, Aral and Eckles (2019) advocated a research agenda to measure
election manipulation on social media and discussed the political, legal,
and ethical aspects of conducting such a measurement-related analysis.
However, they indicated that such research faces a lack of access to
proper data and a lack of cooperation from these platforms, which are
constrained by user privacy and public policy. Similarly, Smith and
3.3. Review of Social Media Networking 241

Seitz (2019) noted that scientific misinformation can lead to ill-founded

educational practices, health trends, and public policies. Using an online
survey-based experiment in the form of a mock Facebook newsfeed, they
corrected neuroscience myths. They reduced people’s belief in these
myths by providing them with corrective related articles immediately
following the myths.
Social media research on societal health is much broader than curbing
the rapid spread of misinformation. Garcia et al. (2018) presented
the Facebook Gender Divide, which captures the standard indicators
of Internet penetration and gender equality indices in education and
economic opportunities. They showed that the tendency of countries
to approach economic gender equality has a negative association with
a high Facebook Gender Divide. These results suggest that, despite
their gender imbalance, social networks may lower information access
hurdles for women and close the economic gender gap. In the same
vein of contributing to societal health, social networking platforms can
provide efficient ways for nonprofits to reach out to communities for
fundraising. Saxton and Wang (2014) used data from Facebook Causes
to understand the nature of charitable giving in social networking. Their
results suggest that donations on social media platforms are driven
by different factors from traditional offline settings. The donations on
Facebook were typically small, and fundraising success was related to its
web capacity rather than the organization’s financial capacity. Moreover,
online donors are more responsive to certain causes than offline donors,
especially those related to health.
Apparently, understanding who spread what to whom within the
context of social media networking warrants additional intensive scrutiny
from researchers, given the increasing importance of social media to
humanity. As part of social media networking, we also emphasize the
importance of dissemination patterns involving information and misin-
formation. The next subsection highlights the misinformation effects of
consumers’ product reviews as a dominantly important type of social
media content from a marketing perspective.
242 Review of Social Media and SMA Related to Marketing

3.4 Review of Consumer Privacy and Fake Reviews

As societally influential marketing topics related to SMA, we turn to

two essential ones—the topics of consumer privacy and data security,
and detecting and managing fake reviews. First, consumers’ engagement
with other consumers and products/brands brings about new challenges
pertaining to privacy, security, information accessibility, governance,
and other information security issues, including data breaches. Khan
(2018) identified three types of social media risks: (1) legal risks; (2)
privacy risks; and (3) security risks. Privacy within the context of social
media involves: the user’s ability to decide what information he or she
discloses or withholds about himself or herself on social media; who has
access to such information; and for what purposes one’s information
may be used or shared. In fact, some users voluntarily yield their privacy
in exchange for using social media platforms and services. Research
into privacy issues suggests that people are more likely to voluntarily
sacrifice their privacy if the data collector is perceived to be transparent
about the nature of the information being collected and how it will be
used (Oulasvirta et al., 2014).
On social media, public data involve information that can be seen
by anyone connected to the Internet, which includes the user’s country,
the public groups of which a user is a member, and personal photos. In
contrast, private data might include the user’s address, gender, work
address, email address, and phone number. Normally, social media
applications allow users to control what information they want to make
public or private. However, it is very difficult or almost impossible for
users to foresee how their specific information can be potentially used in
the future. To make it more challenging, regardless of a user’s selected
privacy settings, all information is always available to the social media
platforms themselves (Khan, 2018).
Nowadays, it has become common for big data sectors, such as
healthcare systems and financial systems, to be involved in multiple in-
stitutions that might have different privacy policies. The problem is that,
while they may not explicitly share their data publicly, joint data pro-
cessing may be conducted (Brei, 2020). In this circumstance, there are
different challenges and perspectives involving data privacy and security
3.4. Review of Consumer Privacy and Fake Reviews 243

such as social network mining, the safety of outsourced databases, data

exchange procedures, and querying cloud-enabled database management
systems. Most of these issues became even more relevant after the imple-
mentation date (May 2018) of the European General Data Protection
Regulation (GDPR), the first extensive privacy law in Europe.
The proliferation of social media and the tendency of consumers
to share their brand experiences on these platforms have presented
companies with opportunities to collect such information (Punj, 2017).
More broadly speaking, data privacy and security are gaining momen-
tum in the research community, partly due to emerging technologies
such as cloud computing and artificial intelligence analytics within the
context of social networking. Therefore, scholars have naturally raised
privacy and ethical concerns for consumers, social media users, and
other stakeholders (e.g., governments and policymakers) because such
data are often collected unobtrusively, without the knowledge or consent
of the user (DiStaso and Bortree, 2014; Hajli and Lin, 2016). Consumers
typically view the collection and use of their information for corporate
purposes as an invasion of their privacy, and as an unethical or unfair
practice (Hajli and Lin, 2016; Le and Liaw, 2017); as a result, they may
refuse to provide information (Cohn, 2010). For this reason, such ethical
concerns are frequently examined in business and marketing studies
(Michaelidou and Micevski, 2019; Schneider and Iacobucci, 2020).
As social media information has become ubiquitous and prevalent, so
have concerns that the quality of information can be compromised when
businesses attempt to game the system. Consumers’ online product
reviews are now a part of everyday decision making for both consumers
and firms. However, the credibility of these reviews is fundamentally
compromised when businesses attempt to commit review fraud by
creating fake reviews for themselves or against their competitors. Luca
and Zervas (2016) investigated economic incentives to commit review
fraud on the popular review platform, Yelp. They reported that roughly
16% of restaurant reviews on Yelp appear to be filtered because they
were suspected of being fictitious. These reviews tend to be more extreme
than other reviews, both favorably or unfavorably. To make matters
worse, the prevalence of suspicious reviews has significantly grown over
time. Luca and Zervas (2016) also reported that firms are more likely
244 Review of Social Media and SMA Related to Marketing

to game the system when facing increased competition and when they
have a poor or less established reputation.
The main challenge in empirically identifying review fraud is that we
cannot directly observe this phenomenon; thus, we must infer whether a
review is fake (Moon et al., 2019). The situation is further complicated
by the lack of a single standard for what makes a review “fictitious”
or “genuine.” To tackle this issue, Moon et al. (2021) used a survey
approach, where they asked different survey participants to write genuine
or fictitious hotel reviews. The findings of this research suggest that
there is an extensive amount of systematic review fraud in online review
platforms. Given this reality, it is imperative for platform managers
to develop mechanisms that can dramatically reduce the number of
fake reviews. Although there is no perfect mechanism to eliminate all
review fraud, it is hoped that future work will develop more of such
mechanisms. One can conceive of leveraging behavioral economics to
reduce such fraud. For example, a platform can highlight the ethical
elements of decision making every time someone wishes to post a review
(Luca and Zervas, 2016; Riquelme et al., 2019).
Further, user-generated contents in social media (such as online
reviews) are inherently incomplete in the sense that we do not capture
the opinions of users who do not write reviews. These silent users may
be systematically different from those who are active on social media.
Such differences can be driven by users’ varying sentiments toward
their shopping experiences, as well as their disposition to post reviews.
For example, Dellarocas and Wood (2008) examined eBay’s feedback
mechanisms for possible reporting bias and found that eBay traders
are more likely to post feedback when satisfied than otherwise. Because
not counting silent users’ opinions can result in reporting bias, Chen
et al. (2016) developed a method to model users’ user-generated content-
generating process and then control for this bias through an inverse
probability weighting approach.
 4
Useful Methods for SMA

In this section, we transition from discussing the research and theories

about social media and SMA to focus more on the practical question of
“how to do SMA.” We discuss this “how” question within the context
of: (1) big data, (2) artificial intelligence (AI), machine learning (ML),
and deep learning (DL), (3) text analytics, and (4) visual analytics.
Text and visual analytics can be seen as a part of ML. However, we
highlight these two analytics categories separately from ML because
they are distinctively unique in terms of the data format (textual and
visual information) and corresponding analysis techniques. Apparently,
each of these topics is technically complicated. Nevertheless, we attempt
to present an accessible introduction of each issue by highlighting
conceptual applications.

4.1 Social Media Data as Part of Big Data

Traditionally, marketing researchers and practitioners have analyzed

numeric, quantitative, and structured data. The growth of Big Data
has ushered in large volumes of data with a host of new practical
problems stemming from advanced and developing technologies such as
the Internet, smartphones, fintech, and cryptocurrencies. Big data has

245
246 Useful Methods for SMA

also produced various types of new data in the form of texts, pictures,
videos, and audios. In the marketing field, this big data phenomenon has
become prominent in marketing problems related to e-commerce and
social media. Further, big data are increasingly driving firm innovation.
In particular, social media is turning consumers into a non-step generator
of not only traditional, structured, demographic, and transactional data,
but also unstructured and visual data (Dong and Yang, 2020). Therefore,
to efficiently utilize big data from various sources for decision making
and innovation, firms must use big data analytics to handle information
with unprecedented volume, velocity, variety, veracity, and value (Chen
et al., 2012).
SMA has become a major research activity, resulting in a large
number of practical analytics services involving web scraping software
and online application programming interfaces (APIs). It is also a re-
search domain undergoing dramatic changes driven by practical interests
(Batrinca and Treleaven, 2014).
Analytics clients in industries (e.g., marketing, PR and communi-
cations, enterprise reporting) have constituencies that require diverse
reporting requirements. Due to developing and diversifying big data
trends, clients are asking more difficult and diverse questions, realizing
that the data summaries they observe are dependent on the platforms
and tools used. This finding suggests the need for a more enlightened
approach to data quality and tool selection (Sponder, 2012).
Common applications of big data analytics for social media include
social media analytics, sentiment analysis, and trend discovery (Ghani
et al., 2019). For instance, social media helps firms obtain customer
feedback regarding their products, which can be used to generate greater
value from their business (Wu et al., 2014). Most existing approaches to
social media big data analysis rely on machine learning techniques (to be
discussed shortly; Cambria et al., 2013). In this task, a notable challenge
involves how to analyze the peculiarities of social media data such as
slang, sarcastic expressions, and deeply technical terms (Gandomi and
Haider, 2015).
Social media monitoring tools are sentiment analysis tools used to
track and measure what people are saying with regard to a company
or its products, or any topic across the web’s social media landscape
4.2. AI, ML and DL 247

(Batrinca and Treleaven, 2014). For example, Google provides several

useful free tools for social media monitoring. Google Trends shows the
frequency of a particular search term over time. Google Alerts is a
content change detection tool that provides automatic notifications.
Researchers have also examined sentiment trends on social media.
Chae et al. (2012) presented an approach that provides users with
interactive social media data analysis and visualization, including the
exploration and examination of abnormal topics and events captured
in various social media data sources. For example, Twitter data can be
viewed as a temporally unfolding narrative. Using the data, researchers
may see sentiments evolving positively or negatively over the course of
a conversation (Brooker et al., 2016).
In summary, the big data phenomenon has advanced various big data
analytics tools, which are also applied for SMA. Many of such tools take
machine learning (ML) approaches. These broad ML approaches include
text analytics (to analyze textual information) and visual analytics (to
analyze visual information) (refer to Table 4.1). Text analytics are
useful for analyzing textual data such as online consumer reviews on
products. Visual analytics are also developing consistently to assist
analysts in examining images, pictures, and videos. ML is a collection of
tools developed by computer scientists as a part of artificial intelligence
(AI), which typically includes neutral networks, random forests, and
support vector machines. Further, ML approaches are an integral part
of conducting text and visual analytics. In this sense, we classify the
two analytics categories as parts of ML. However, for the purpose of
exposition, we introduce ML, Text Analytics, and Visual Analytics as
three major tool categories in the following subsections.

4.2 Artificial Intelligence (AI), Machine Learning (ML),

and Deep Learning (DL)

The widespread impact of artificial intelligence (AI) and machine learn-

ing (ML) on all segments of society have driven researchers to call the
present day as the AI Age. This AI revolution has greatly increased mul-
tidisciplinary research. In the business world, such processes have been
impactful as a significant source of innovation (Huang and Rust, 2018).
248 Useful Methods for SMA

Table 4.1: Summary of common tools for SMA

SMA Area Methods Examples

Machine Learning Neural network (NN) recurrent Predicting product sales;
(ML) and Deep neural network (RNN) variable (product feature)
Learning (DL) convoluted neural network selection; combining ML and
(CNN) decision tree and econometric modeling for
random forest regression both practicality and rigor;
clustering [also includes Text object identification
and Visual Analytics]
Text Analytics Natural language processing Extracting product features
(NLP) sentiment analysis from product reviews (Yelp,
(opinion mining) topic Amazon, Twitter) using topic
modeling (text categorization) modeling; feature-specific
latent Dirichlet allocation sentiment analysis; detecting
(LDA) social media fake reviews based on
monitoring netnography language usage; developing
product-specific taxonomies;
developing sentiment lexicons
Visual Analytics Computer vision interactive How features of product
visualization multilevel pictures on social media
visualization convolutional influence consumer
neural network (CNN) RGB preferences and choices;
color histogram dashboard visualizing brand locations
using consumer reviews;
analyzing product pictures;
location-based visualization

Brei (2020) discussed the central role that AI and, more specifically, ML
can play as research methods in marketing (Hagen et al., 2020). Thus,
we can easily find references using AI, ML, and deep learning (DL) as
synonyms. Although all three are closely related, in fact, they are not
synonyms. First, AI refers to a system’s ability to correctly interpret
external data, learn from such data, and use this learning to achieve
specific goals through flexible adaptation (Kaplan and Haenlein, 2011).
ML is a subset of AI that addresses ways to build algorithms that
improve automatically through data (Jordan and Mitchell, 2015). There-
fore, ML is considered as the opposite concept of human learning. In
other words, ML refers to the study of methods or algorithms designed
to learn underlying patterns in data and make predictions based on
its learning of these patterns (Dzyabura and Yoganarasimhan, 2018).
Common ML models include neural networks, support vector machines,
4.2. AI, ML and DL 249

decision trees, and random forests. Lastly, DL is an ML approach that

involves gathering knowledge from data, thus forming a hierarchy of con-
cepts that enables a computer to learn complex concepts from simpler
ones. DL deals with a collection of machine learning methods that train
multiple layers of data representation. It uses a multilayered cascade of
nonlinear processing units (that is, neurons) for feature extraction and
transformation. In this process, the output from each layer can be used
as input for the succeeding layer, all together comprising a number of
layers (Goodfellow et al., 2016).
ML (and DL) are implemented by algorithms, which are unambigu-
ous specifications of ways to solve problems through learning from data.
An ML algorithm is said to learn from experience (E) with respect to
some class of tasks (T) and performance measure (P), if its performance
on T as measured by P improves with E (Mitchell, 1997). There are
three standard features of a learning algorithm: representation, evalu-
ation, and optimization. Representation indicates the presentation of
a sequence of tasks in a way a computer can manage. Evaluation (i.e.,
the objective function) involves distinguishing between good actions
from bad ones in terms of consequences. Optimization is the process
of searching for and selecting the best solutions, given the specified
objective function criteria (Domingos, 2012).
The learning in ML and DL can be supervised (i.e., trained on an
initial data set with outcomes controlled by the user-programmer) or
unsupervised (Kaplan and Haenlein, 2011). Supervised learning applies
when the training data are provided with known outcome values, whereas
unsupervised learning is applied when the training data have no known
outcome values (Hayat et al., 2019). Supervised learning is also called
predictive learning because the outcome of the objective function can
be predicted using data training, and therefore is commonly applied
(Jordan and Mitchell, 2015). Its purpose is to map from inputs to
outputs based on a labeled set of input-output pairs called the training
set. In supervised learning, a model is developed and specified based
on previous observations as a training set (Khan, 2018). Inputs can be
as simple as a vector of numbers but can also include more complex
and unstructured objects such as sentences, texts, times-series data,
and images. When an output is categorical or nominal, the problem
250 Useful Methods for SMA

refers to classification or pattern recognition. When outputs are real

values, the problem can be estimated using various regression models.
In supervised learning, a comparative metric is used to evaluate how
accurate the algorithm is in its output prediction compared to other
competing algorithms. Marketing scholars have used supervised learning
algorithms frequently in the form of linear and logistic regressions, naïve
Bayes, and latent semantic analysis (Batrinca and Treleaven, 2014).
Almost all applications of DL belong to the boundary of ML, including
those of object detection, image classification, speech recognition, and
language translation (Chollet and Allaire, 2018).
The unsupervised machine learning approach seeks to reveal patterns
in the data instead of attempting to fit the data into a predefined
structure. Thus, unsupervised learning involves finding interesting and
informative patterns in the data with the use of any outputs and,
accordingly, is called descriptive learning. In this approach, the problem
is much less specifically defined, as one does not know in advance
which patterns to search for; moreover, there is no straightforward
metric for determining the effectiveness of a learning task (Murphy,
2012). Typically, the unsupervised learning algorithm is designed to
find similarities between objects so that objects with commonalities are
categorized together. Some widely used unsupervised learning algorithms
in marketing include principal components analysis (for dimensionality
reduction) and k-means clustering (for segmentation) (Hagen et al.,
2020). It is not uncommon to use unsupervised learning to understand
data patterns before attempting to tackle a supervised learning problem.
Reinforcement learning (RL) involves mapping situations to actions
to maximize a numerical reward signal (Brei, 2020). In RL, the algorithm
must identify the actions that can maximize rewards through trial and
error. It is designed to capture the most critical aspects of the problem
faced by the learning algorithm, which aims to accomplish a goal by
repeatedly interacting with the environment over time. RL possesses
four primary elements: (1) a policy (the agent’s way of behaving); (2) a
reward signal (the goal of a reinforcement problem); (3) a value function
(what is effective in the long run); and (4) a model of the environment.
RL differs from supervised learning because it has no predefined correct
answer. On the other hand, it differs from unsupervised learning because
4.2. AI, ML and DL 251

it does not involve unearthing hidden relationships. In ML, the algorithm

must attempt a variety of actions and progressively choose those that
appear to improve the current outcome (Sutton and Barto, 2018). In
this sense, RL is a form of ML that most approximates AI. Some
applications of RL include robots, games, and self-driving cars (Brei,
2020).
As a widely used ML and DL method, the Neural Network (NN) is
a paradigm used to process information with terminology drawn from
an analogy to biological neurons (Briesch and Iacobucci, 1995; Hayat
et al., 2019). NN attempts to find relationships in datasets by mimicking
the way the human brain works. NN is composed of numerous tier-wise
arranged processors operating in parallel. Raw input is given to the
first tier, and the output from each preceding tier serves as input to the
succeeding tier. The last tier finally gives the final output (van Gerven
and Bohte, 2018). In contrast to DL, in basic NN, training in neural
networks happens simultaneously for all layers.
The Recurrent Neural Network (RNN) is a class of NN, where
the links between components construct a directed graph, to model
progressive dynamic behavior. RNN can be used to process seemingly
random input sequences, and has been used for tasks such as converting
handwriting and speech to text (Deng, 2014). Next, the Convolutional
Neural Network (CNN) is a somewhat simplified neural network that
starts with convolutional and sub-sampling layers. These layers filter
and organize the network system and do not depend on being hand-
engineered. Also, these layers filter and organize the network system by
“convoluting,” which inputs direct information into subsequent network
nodes rather than allowing all nodes to be interconnected (Chen and
Lin, 2014). Thus, training CNN can be easier than more fully linked
networks, thus reducing calculation time.
ML offers several advantages and new perspectives for knowledge
generation in marketing, which has intensively been using ML meth-
ods such as regression, classification, visualization, and dimensionality
reduction. Its applications are increasing to new heights with recent de-
velopments in Big Data and machine learning techniques (Hagen et al.,
2020). Despite these manifold applications, in the marketing discipline,
ML methods are usually compared to traditional econometric methods
252 Useful Methods for SMA

(Brei, 2020). Thus, Dzyabura and Yoganarasimhan (2018) noted some

differences between machine learning approaches and traditional econo-
metric approaches in marketing. First, ML methods are focused on
obtaining the best predictions, whereas causal econometric methods are
focused on deriving the best unbiased estimators. Thus, techniques for
causal inference often do not perform well in out-of-sample predictions,
as the different approaches optimize different objective functions. In
the ML discipline, this problem is known as the bias-variance tradeoff.
Second, many ML approaches are developed in a situation with no
a priori theory about the process through which the outcomes observed
in the data are generated. Third, the ML approach can handle an ex-
tremely large number of variables better by selecting variables to retain.
Thus, the ML approach employs feature selection and optimization to
achieve scale and efficiency, an increasingly important goal for real-time
marketing problems in the current Big Data era (Brei, 2020). On the
other hand, most ML applications in marketing adopt a descriptive
or predictive perspective without rigorous causal modeling. Therefore,
when we suitably combine ML applications (for practicality and scale)
and econometric modeling (for theoretically rigorous modeling), we can
offer a higher level of approach for marketing problems and theories.

4.3 Text Analytics

Much of the data available on social media platforms are unstructured,

qualitative text data such as online consumer reviews of various prod-
ucts. Researchers have used text analytics for important marketing
problems by eliciting relevant information from the textual information,
which was not feasible from traditional marketing research approaches
that primarily relied on structured numerical data. Thus, social media
has been an enriched source of consumer purchase and consumption
experiences for marketing researchers and practitioners. Common ap-
plications of text analytics to social media problems include extract-
ing product attributes (Ghose et al., 2012; Lee and Bradlow, 2011),
inferring brand perceptions and preferences (Aggarwal et al., 2009;
Moon and Kamakura, 2017), identifying factors that influence sales
4.3. Text Analytics 253

(Liu et al., 2019; Moon et al., 2014), and detecting fake reviews (Luca
and Zervas, 2016; Moon et al., 2021).
Text analytics requires dissecting the target textual document to
turn it into structured information. One common way of imposing
structure to an unstructured document is to generate a term-and-
frequency matrix. Terms are expressions with a single meaning, including
single words. Commonly used multi-word expressions are an important
part of terms (e.g., low price, high speed, brand loyalty). When we
generate such a matrix, we can more readily use the textual information
for further analysis. Natural language processing (NLP) is an essential
and fundamental tool for such a task. NLP is a field of computer science,
artificial intelligence, and linguistics concerned with the interactions
between computers and human (natural) languages. It is the process of
a computer extracting meaningful information from natural language
input and producing natural language output (Batrinca and Treleaven,
2014). Text analytics involves information retrieval and lexical analysis
to analyze word frequency distribution, pattern recognition, tagging,
and annotation supported by association analysis, visualization, and
predictive analytics.
A primary challenge in utilizing consumer-generated content for
analysis is that the data are mostly qualitative and unstructured in
nature (Godes et al., 2005). Before text analytics tools became common
and popular, many researchers resorted to the quantitative character-
istics of consumer-generated data such as product ratings, including
their volume, valence, and variance (Moe et al., 2017). These measures
have been used to investigate the relationship between ratings and sales
(Dellarocas et al., 2007). Similarly, Godes and Mayzlin (2004) demon-
strated the relationship between the variance in online conversations
and television ratings.
In many online product reviews, consumers typically provide a
measure of valence in the form of a 5-point star rating, as found on
Amazon and Yelp. However, there are various types of product reviews
that do not include a quantitative rating, as in blogs and tweets. In
this case, the direction and degree of valence can be estimated with
sentiment analysis (Pang and Lee, 2008), which extracts expressions
associated with positive, neutral, and negative emotions to capture
254 Useful Methods for SMA

the overall sentiment in a message. However, more detailed content

analyses beyond valence measures must usually employ sophisticated
and time-consuming manual coding.
Academics have used text mining to analyze the knowledge base
on a particular topic in an automatic fashion. Computer scientists and
information system scholars have made the greatest leap in developing
advanced text mining techniques. Importantly, collaborations between
computer scientists or information systems scholars and business scholars
helped facilitate the dissemination of these tools in business research
(Das and Chen, 2007; Lee and Bradlow, 2011). In marketing, the first
attempts to text analyze user-generated content conducted manual
text mining, including humans reading messages and judging their
content (Godes and Mayzlin, 2004). Because this manual text mining
is considered to be too time-consuming and too laborious, computer
scientists and marketing scholars have developed efficient software and
algorithms to automatize the process in various contexts (Rust et al.,
2021; Toubia and Netzer, 2017).
Text mining has been increasingly used in marketing to extract
brands, product attributes, product-describing adjectives, and consumer
sentiments (Lee and Bradlow, 2011; Ludwig et al., 2013). Numerous
commercial companies provide similar services. One example would be
measuring how a new product is received by the target market using
social media. Applications of text mining user-generated content in
marketing can be classified into two main streams, according to the
goals of the analysis: (1) to describe and monitor markets; and (2)
to predict important market outcomes (Moe et al., 2017). In the first
stream, by text mining the textual information, companies can listen to
and monitor consumer discussions and opinions about their products
and competitor products, possibly in real time. For example, Tirunillai
and Tellis (2014) used linguistic data analysis to describe the dimensions
of product quality elicited from computer reviews over time. They found
that comments about Dell’s perception of “ease of use” were highly
volatile over time, whereas those of Hewlett Packard were steadier.
Regarding the second research stream, researchers have investigated
the relationships between text-mined information and market outcomes
such as consumer choice, aggregate sales, and firm value (Berger and
4.3. Text Analytics 255

Milkman, 2012; Decker and Trusov, 2010; Eliashberg et al., 2007). For
example, Archak et al. (2011) demonstrated that the textual information
in product reviews can help inform consumer preferences for product
attributes, which subsequently improves accuracy in sales prediction.
Some studies have shown that the textual analysis of user-generated
content helps predict stock prices (Tirunillai and Tellis, 2012; Yu et al.,
2013).
Netnography is a qualitative, interpretive research methodology
that uses Internet-optimized ethnographic research techniques to study
virtual communities and networks. With the help of netnography, re-
searchers can conduct Internet-related research by either actively en-
gaging the members of a community or by passively monitoring the
community (Kozinets, 2002). Netnography can be faster, simpler, and
less expensive than ethnography, and more naturalistic and unobtru-
sive than focus group interviews. It can provide information on the
symbolism, meanings, and consumption patterns of online consumer
groups (Kozinets et al., 2010). Netnographic analysis can be aimed at
exploring what companies are posting, what consumers are saying, and
how companies are responding to consumers. It can help companies
monitor their digital presence constantly to grasp the type of content
that resonates better with their customer bases (Ahuja and Alavi, 2018).
SMA aims to elicit information and knowledge about individuals
and organizations, as it emerges from the communication dynamics in
social networks. Topic modeling (also called text categorization) is used
to sift through large bodies of text to detect useful and relevant topics.
In analyzing a large body of product reviews on Amazon, topics can be
product features (such as price, quality, service, and durability). Recent
advances in topic modeling allow numerous algorithms to be used with
streaming data from Twitter and other continuous data feeds, making
the methodology an increasingly important analytics tool in SMA. Topic
modeling uses a variety of advanced statistics and machine learning
techniques. For instance, some topic modeling techniques identify “latent”
topics through the co-occurrence of words in communications among
users (Fan and Gordon, 2015). Latent Dirichlet Allocation (LDA) has
been used as a popular tool to identify unobserved topics in SMA (Blei
et al., 2003; Zhong and Schweidel, 2020).
256 Useful Methods for SMA

As topic modeling is efficient at organizing and analyzing large

volumes of textual data, researchers and practitioners are increasingly
choosing it over conventional, human-based coding methods. However,
challenges exist when using automated methods. For example, LDA and
most other text mining algorithms require that the number of topics
to be extracted from a sample be predetermined when its true value
is unknown. Further, human interpretation of the identified topics is
necessary. Accordingly, analysts must choose the technique that works
best for them based on their resource availability, the volume of data,
and the intended analysis objectives (Kordzadeh and Young, 2020).
As a developed topic modeling software program, Linguistic Inquiry
and Word Count (LIWC; Pennebaker et al., 2015) is efficient at conduct-
ing general concept-based text mining analysis (Moon et al., 2021). This
dictionary has been used in academic research for content analyses of
textual documents (Humphreys, 2010). This approach of using a refined
and detailed dictionary for topic modeling was recently adopted by an-
other text-analytics program called The Evaluative Lexicon (Rocklage
and Fazio, 2020; www.evaluativelexicon.com) as a free software. LIWC
topics cover various grammar and linguistic styles (e.g., word count,
question marks, comparatives), as well as various human psychological
and social processes (e.g., negative emotion, affiliation, religion). LIWC
can be an effective tool to analyze consumers’ generalized consumption
and purchase experiences across various products in part due to its
breadth, covering general human psychological and social processes.
Analyzing user sentiments can be applied to recognizing subjective
opinions and preferences in texts produced in online social contexts,
gathering collective evidence across communities, and tracing under-
lying social phenomena (Basili et al., 2017). The process is complex
and often enabled by manually coded dictionaries, capturing peculiar
human emptions expressed in textual comments. Sentiment analysis is
attracting increasing attention in academia. It involves the process of
extracting and classifying opinions expressed in online documents; then,
the resulting data are analyzed against rigorously developed dictionaries,
that is, a list of words or phrases classified into positive, negative, and
neutral sentiment. Companies can analyze online comments that reflect
people’s sentiments or perceptions (Chang et al., 2019).
4.3. Text Analytics 257

Sentiment analysis mainly uses two approaches: (1) a machine learn-

ing approach with which the analytics learns to evaluate sentiments
by acquiring and integrating knowledge gained from a large amount of
sentiment examples and (2) a lexical-based approach with which the an-
alytics evaluates sentiments by utilizing a dictionary of prescored words
and phrases (Lee, 2018). This lexical-based approach uses a predefined
set of expressions that carries a specific type of sentiment. These terms
can include simple word or phrase counts, emoticons, sentiment lexi-
cons (based on words in the lexicon that have received specific features
connoting positive or negative terms in a message), and psychometric
scales (to identify mood-based sentiments).
Various degrees of sentiment analysis can be implemented. In other
words, the unit of sentiment analysis can be at various levels such as
a collection of documents, a single whole document, a sentence, or a
particular aspect (Liu, 2012). Capturing the general sentiment of a
collection of documents may not turn out to be very difficult, given
the amount of information for the analysis purpose. However, precisely
capturing a sentiment pertaining to a single sentence can be more
challenging because there is much less information. Trying to capture
the sentiment of a particular aspect, such as a particular product feature
(e.g., service satisfaction at a restaurant), can be even more difficult
because a reviewer typically mentions multiple product features in the
review. Specifically, determining which specific terms and expressions
are related to the target feature in the review is complicated and can
be made possible only with very conscientious coding to capture the
related terms precisely and reflect the sentence structures accurately.
For example, the reviewer may praise a restaurant for good-tasting
food but may also criticize it for poor service. Because these terms
can be dispersed and mixed, automatically capturing these emotional
expressions and matching them to a particular feature does not come
easy. In this case, both aspect extraction and sentiment analysis are
required to improve the analysis. In responding to these challenges,
various methods of aspect-based sentiment analysis have been proposed,
including rule-based, topic model-based, and supervised learning (Chang
et al., 2019). In recent years, some research has been conducted based
on LDA to develop topic modeling methods for aspect-based sentiment
258 Useful Methods for SMA

analysis tasks. These approaches utilize variations of LDA to uncover

latent topics in a dataset, expecting that these topics will correspond
to ratable aspects of the entity under review.

4.4 Visual Analytics

Data visualization involves the representation of data through graph-

ical means to communicate information clearly and effectively to the
intended audience (Batrinca and Treleaven, 2014). Visual analytics
integrates techniques from two fields: information visualization and
computational modelling. It has already contributed significantly to
SMA (Hassan et al., 2014). Interactive visualization is a strong feature of
this methodology, as opposed to static outputs displaying the outcomes
of an analysis (Thomas and Cook, 2005). Central to the methodology
is the notion that visualizations do not replace the work of analysts,
but rather amplify their inherent capabilities by capitalizing on the
high-bandwidth processing of visual perception. The process usually
begins with an overview of high-level abstractions that guide analysts
toward potentially interesting and relevant aspects of the data (Brooker
et al., 2016). Then, researchers can either transform the view or create
new visualizations that enable them to conduct finer analyses of their
data.
Visual analytics processes data to reveal the underlying relationships
in the structure of the model. Computing variable relationships for data
reduction, displaying correlations across multiple data sources, and
enabling analysts to manipulate data displays are the major strengths
of visual analytics. The dashboard is a commonly used interface, where
multiple metrics and key performance indicators are portrayed in a
way that mimics an automobile’s dashboard. The displays typically
allow users to interactively investigate the underlying data structure
and perform data transformations by using sliders or other types of
controls. In the SMA area, analysts can employ visual analytics to
detect major changes using social media chatter. A major challenge
for visual analytics is to remain responsive to, and create better visual
representations of increasingly massive and complex data requiring
real-time interpretation (Fan and Gordon, 2015).
4.4. Visual Analytics 259

Spurred initially by U.S. defense needs, visualization works across

a variety of application areas to support the synthesis, exploration,
discovery, and confirmation of insights from data that are typically
voluminous and come from multiple sources. Many statistical analyses
rely on visual analytics and the human ability to capture underlying
patterns from visual graphs and pictures. The beauty of visual analytics
comes from a computer’s capability to store and analyze big data
efficiently, along with humans’ perception skills, cognitive reasoning, and
domain knowledge. Chang et al. (2019) used visual analytics techniques
to explore extracted data (hotel ratings, aspect-sentiment, and types of
travelers) using Tableau, which is a popular and representative software
for visualization. It can instantly provide a timeline analysis, location-
based analysis, and various dashboard metrics.
Given the magnitude of the data involved in the big data era,
the importance of visualizing key results and insights from analyses
has been increasing. In an analysis, visualization can reveal hidden
relationships that lead to novel insights (Fan and Gordon, 2015). It
effectively combines machine and human strengths to process and
explore big data and produces decision-supporting information (Sacha
et al., 2014). It also helps organize and interpret textual information
within the context of SMA.
Recently, the marketing field has increased its usage of visual analyt-
ics to marketing problems. One example of visualization in marketing is
product/brand preference maps combined with psychometrics (Netzer
et al., 2012). For example, Moon and Kamakura (2017) text analyzed
wine reviews to visualize wine preferences and perceptions in wine
product maps. It is an efficient way to visually summarize 29 wine
features (e.g., floral, dried fruit, chocolate, smoked, earthy, medium
body) elicited from a corpus of wine reviews.
Another application example of visualization in marketing involves
examining the impacts of product features elicited from product pictures
on consumer preferences and choices. Liu et al. (2020) maintained that
images are close to overtaking text as the medium of choice on social
media because images contain significant information about users’ con-
sumption experiences. Given the importance of images, they proposed
260 Useful Methods for SMA

a “visual listening in” approach, which mines the visual content of user-
generated content to measure how brands are presented to consumers
on social media. Specifically, they developed and applied BrandIma-
geNet, a multi-label deep convolutional neural network model, to brand
pictures to extract the descriptions of 56 national brands in apparel
and beverage products. They found a strong association between brand
descriptions in the product images and consumers’ brand perceptions
collected by the survey method. Similarly, Li and Xie (2020) found that
professional, high-quality photos generate stronger consumer engage-
ment on social media using airline and sport utility vehicle brand data
and pictures on Twitter and Instagram. In addition, Jalali and Papatla
(2016) examined the roles of color composition involving visual content
in consumer responses. They found that users’ click-rates are higher for
photos with greater proportions of green and lower proportions of red
and cyan. On the other hand, Hartmann et al. (2021) classified brand
images on social media into three categories: packshots (standalone
product images) and two types of brand-related selfie images (consumer
selfies featuring brands and consumers’ faces and brand selfies showing
invisible consumers holding a branded product). The authors used con-
volutional neural network models to identify these types and trained
language models to infer user responses to images of 185 brands on
Twitter and Instagram. They found that consumer-selfies resulted in
more comments about the senders, whereas brand selfies received more
brand-related likes and comments.
Nanne et al. (2020) demonstrated how computer vision can be ap-
plied for marketing practitioners and academics alike by investigating
the usage of three computer vision models – YOLOV2, Google Cloud
Vision, and Clarifai. They used more than 20,000 Instagram pictures
pertaining to 24 brands using the three computer vision models. Com-
puter vision is considered as an effective tool to automatically analyze
the content of visual images, particularly for object identification (such
as brand identification in social media pictures). Their analysis showed
that the three models have differential applicability. Specifically, Google
Cloud Vision provided more accurate results in object detection, while
Clarifai produced more useful labels for the interpretation of brand
portrayal. However, YOLOV2 turned out not to be useful in analyzing
4.4. Visual Analytics 261

brand-related images. Although the influences of product pictures on

consumer preferences and decisions are widely believed, we have a very
limited number of studies that have examined this topic. Along with the
advancing visual analytics techniques, this stream of research appears
very promising.
 5
Trends, Future Research, and Education in SMA

In this section, we discuss the current trends in SMA and future research
avenues based on the current trends we identify. After we present general
future research avenues, we then discuss SMA education in higher
education.

5.1 General Trends and Future Research Avenues in SMA

Social media presents a host of opportunities with various challenges.

Social media technology may be called disruptive because it often
shifts brand control away from brand managers to end consumers.
Without clearly established business practices, most current social
media marketing initiatives still rely on traditional marketing measures
such as transactional sales instead of building long-term relationships
through customer relationship management (Oh et al., 2015). Research
on the “voice of the customer” has shown that organizations have
difficulty in collecting, managing, and analyzing social media data
(Fowler and Pitta, 2013). Such challenges emphasize the imperative
need for SMA. Particularly due to the unstructured and qualitative
data formats commonly found on social media, researchers have pointed
out that SMA, as an essential tool, has transformed marketing from

262
5.1. General Trends and Future Research Avenues in SMA 263

an art to a science (Davenport and Harris, 2007). If organizations can

learn how to better implement it, SMA can give them a competitive
advantage. Given the demand and need for SMA, there are ample
opportunities for more solid outcome-oriented research and practices.
With the rise in social media data, such as product reviews and
brand communities, marketing scholars have created domain taxonomies
to facilitate analyses in specific domains in the name of ontology learn-
ing. Ontology is defined as an explicit, formal specification of a shared
conceptualization within a domain of interest, where “formal” suggests
that the ontology should be machine-readable and shared with the
general domain community (Buitelaar et al., 2005; Liao et al., 2009;
Missikoff and Navigli, 2002). For example, Moon and Kamakura (2017)
used ontology learning to expand, deepen, and revise an original taxon-
omy called the Wine Aroma Wheel, sourced from a single expert, by
collecting actual wine reviews, identifying relevant terms, and classifying
these terms around the basic taxonomy embedded in the Wine Aroma
Wheel (www.winearomawheel.com).
An important practical issue in producing a taxonomy from product
reviews is scalability, namely, how to process large amounts of textual
information (Soucek, 1991). When marketers collect online product
reviews, they must organize and analyze large amounts of unstructured
textual reviews. Thus, any ontology learning should be conducted with
an automatic text mining approach, as opposed to a time-consuming and
tedious manual approach. As with many other AI methods, there are two
main approaches to this taxonomy building (Feldman and Sanger, 2007).
First, the knowledge engineering approach encodes experts’ knowledge
about the categories into the system either declaratively or in the
form of procedural classification rules. The other one is the machine
learning approach, where a general inductive process builds a classifier
by learning from a set of pre-classified examples. It is known that
the knowledge engineering approach usually outperforms the machine
learning approach in the document management domain, but the former
approach requires deep-level domain knowledge and a significant amount
of manual labor. The approach Moon and Kamakura (2017) utilized
in mining text from wine reviews falls in the knowledge engineering
approach: they started with an expert’s taxonomy (the Wine Aroma
264 Trends, Future Research, and Education in SMA

Wheel) and expanded the taxonomy, combining machine learning with

human judgment through an iterative process. The primary challenge
of the knowledge engineering approach is the knowledge acquisition
bottleneck. In other words, the approach requires a large amount of
expert knowledge to create and operate knowledge-encoding rules. In
their research, this burden was reduced by using an existing wine
taxonomy (i.e., the Wine Aroma Wheel) and experts’ detailed wine
reviews.
Most studies have been limited to analyzing each domain (e.g.,
industry) separately without combining the mechanisms of multiple
related domains in text analytics due to the unique and subjective
aspects of each domain (Abbasi et al., 2018). For example, when we
travel, we use multiple services, including flights, hotels, and restaurants.
Some platforms provide a view of consumers’ comprehensive activities
(e.g., Tripadvisor). Similarly, on Amazon, we can see a consumer make
numerous purchases across various product categories over time. This
presents cross-category dynamic modeling opportunities, which can be
technically very challenging. Given the number of product categories
(e.g., electronic goods) and products (e.g., laptops, TVs, refrigerators)
available on social media, building product-specific taxonomies seems
a big challenge. In particular, this task becomes more challenging
when we tackle cross-product problems such as recommending products
based on cross-product analysis (as Amazon does with their customers)
and cross-product pricing effects. Generally speaking, in social media
research, scholars have typically focused on a single product but have
not sufficiently examined how cross-product effects unfold in terms of
marketing actions. Yet, of course, that is not necessarily how customers
think and behave. Thus, if we can figure out how to examine a portfolio
or basket of products, we will have a more comprehensive picture of
consumers’ general purchase and consumption activities. In particular,
building proper domain taxonomies for related products can be useful
in such cross-product research within the social media space.
This type of research requires multi-domain text analytics, which
means that instead of considering each domain separately, we need
to find a more general technique that covers both the commonalities
of the multiple domains and the peculiarities of each domain. Being
5.1. General Trends and Future Research Avenues in SMA 265

more comprehensive may require high-level sophisticated AI techniques

using machine learning methods. Generally speaking, the widespread
impacts of artificial intelligence (AI) and machine learning (ML) in many
segments of society have not yet been felt strongly in the marketing
field.
Many studies have focused on understanding reviews, but only a few
studies have focused on reviewers. To understand the roles of reviewers,
some studies have examined how reviewer characteristics (e.g., reviewer
rank, number of past reviews) influence their review helpfulness and
product sales (Ghose and Ipeirotis, 2011). Further, individual reviewers’
product preferences and writing styles revealed in their product reviews
may have differential impacts on other consumers (Naylor et al., 2011).
Other studies have focused on reviewers’ motivations in posting their
reviews, despite a lack of direct compensation (McIntyre et al., 2016).
Based on the motives of consumer online articulation, Hennig-Thurau
et al. (2004) classified online consumers into four groups: self-interested
helpers, multiple-motive consumers, consumer advocates, and true altru-
ists. Similarly, based on reviewers’ review engagement, Mathwick and
Mosteller (2017) identified three types of reviewer segments: indifferent
independents, challenge seekers, and community collaborators.
Online product reviews contain information not only about the
products under review, but also about the consumers writing them due
to reviewers’ differences in perceptual acuity, writing styles, and product
responses. For example, in hotel reviews, leisure travelers may focus on
proximity to touristic attractions, whereas business travelers may be
interested in proximity to business centers. SMA needs more research
on reviewers, their thinking, behavior, and dynamic development. In
particular, there is insufficient research on how reviewers differ from
one another, reflecting inherent differential product preferences revealed
on social media.
Visual information in social media (such as product pictures) is
following textual information (such as online consumer reviews on
products) and, accordingly, is emerging as the next frontier in the
Big Data revolution. For example, we have a very limited number
of studies that have examined the influences of product pictures on
consumer preferences (Hartmann et al., 2021; Jalali and Papatla, 2016;
266 Trends, Future Research, and Education in SMA

Li and Xie, 2020; Liu et al., 2020; Nanne et al., 2020). Along with the
advancing visual analytics techniques, this stream of research appears
very promising. Computer vision (Nanne et al., 2020) can be a very
useful tool in this endeavor.
From the perspective of social dynamics on social media, some
studies have analyzed the magnitude of reviewer influence on other
consumers (Moe and Trusov, 2011; Mu et al., 2018; Trusov et al., 2009).
Generally speaking, social influence is delivered in complex social set-
tings that include interaction among participants. Such social influence
is known to be positively associated with reviewers who have high
product expertise and strong social connections within the context of
social media (Agnihotri and Bhattacharya, 2016). These studies, based
on the social dynamics of online reviewers, are primarily focused on in-
teractions among social media users. By contrast, there is little research
on individual reviewers’ developmental changes over time. Generally,
longitudinal studies are more challenging to execute, and this case is no
exception.
In a general dynamic view, trend analysis on social media can forecast
the growth of customers or sales, or predict ad campaign effectiveness.
Further, combining transactional and social media activity (including
how other consumers react to social media messages) may pave the
way to revisit the marketing view of customer value, including both the
transaction value and social influence (Ho et al., 2012).
New methodological approaches can widely open SMA applications
to marketing practitioners. As a promising SMA tool, visual analytics
have only begun to be utilized in deriving novel insights for marketers.
Visual analytics enables analysts to probe their own assumptions and
perspectives to understand developing phenomena (Brooker et al., 2016).
Images and videos are essential elements of social media data due
to rapidly growing content-sharing platforms such as Instagram and
YouTube (Lee, 2018). Numerous images and videos are uploaded every
day, and mining such data can generate additional novel insights beyond
what can be captured from textual information. While various methods
have been developed primarily for text mining, analytics for images
and videos have thus far been marginally utilized. Advanced image
analysis utilizes image processing techniques and image recognition.
5.1. General Trends and Future Research Avenues in SMA 267

Video analysis involves metrics such as the number of users, response

rate, subject, and location. More advanced techniques include accessing
video clips posted to social media sites and analyzing voice to determine
the user’s emotional state (Lee, 2018). Because social media data include
a variety of images and videos (e.g., product pictures, product review
videos), the potential for visual analytics in SMA and marketing seems
enormous.
Another potentially promising SMA tool is automated machine
learning (AutoML), which offers methods and processes for making
ML available to nonexperts by automating the end-to-end processes of
applying ML to real-world problems. In ML applications, researchers
usually perform a variety of tasks to develop efficient and precise models
such as data cleaning, feature extraction and selection, and algorithm
selection. Developments in ML have generated the demand for easy-
to-implement ML methods for practitioners (Brei, 2020). For example,
customer relationship management (CRM) can benefit from AutoML, as
AutoML is introducing new automated CRM approaches to marketers.
On the other hand, social media analytics tools possess two notable
limitations that reduce sensemaking on social media (Abbasi et al., 2018).
First, communication modes such as chat rooms and forums are highly
susceptible to intertwined conversations and inexplicable incoherence
(Fu et al., 2008). Second, current text analytics tools mostly focus on
the semantic dimension of language (what people are saying). However,
while using such tools, organizations have difficulty in understanding
how consumers are saying something and what their motives are (Abbasi
et al., 2018).
When addressing these two limitations, there are two notable re-
search gaps. First, most current text analytics research has adopted a
semantic view (Abbasi and Chen, 2008; Lau et al., 2012), with thou-
sands of studies focusing on sentiment analysis for topic identification
in textual information. The body of literature on the pragmatic view,
highlighting the communication context, actions, and interactions, has
received relatively less attention. Second, text analytics-based research
using the pragmatic view is still sparse in the sense that there is no
comprehensive framework encompassing developments in this research
area. In an effort to close these two gaps, Abbasi et al. (2018) adopted
268 Trends, Future Research, and Education in SMA

the design science paradigm (Hevner et al., 2004). As a language action

perspective-based text analytics framework, this paradigm is intended
to produce insights for the design of information systems. Marketing
scholars can contribute to this effort in facilitating effective and efficient
communications on social media, particularly from the social networking
perspective. This perspective can focus on effective communications
among social media users, which can contribute to consumer welfare.
To promote effective communications on social media, we revisit
applying LIWC to social media data as a useful text analytics tool to
analyze product reviews from a theoretical perspective (Moon et al.,
2021). This notion of using existing topic-modeling-based software has
been applied to another text-analytics program called the Evaluative
Lexicon (Rocklage and Fazio, 2020; www.evaluativelexicon.com). LIWC
covers a wide variety of topics pertaining to human psychological,
social, and biological processes. Instead, the Evaluative Lexicon focuses
on measuring people’s emotionality. This approach can contribute to
future research in two ways. First, more developments in such software
can facilitate marketing research that examines the covered topics.
For example, software that measures various types of social network
relationships can be effectively used for social media networking research.
Second, research using such software is encouraged to glean insights
from the hidden relationships among key social media concepts (e.g.,
relationships between influential reviewers and their followers). Because
product reviews consist of huge amounts of unstructured information,
many key relationships may not be easily revealed. Using the key topics
contained in such software tools can be an effective way to conduct
research on social media.
Thus, academics might wish to use the general LIWC dictionary,
whereas managers may wish to expend initial efforts to tailor the
dictionary to be more specific to their industry so that their searches
and results will be more precise. For example, a healthcare marketing
manager may wish to extend the LIWC dictionary for their specific
industry keywords to enhance the relevance of the text analyses. The
outlay of effort in creating an industry-custom dictionary is primarily
prior to the first use, though managers may wish to update their tailored
dictionaries as new events and terminology arise.
5.2. SMA Education in Colleges 269

5.2 SMA Education in Colleges

It is common for academic articles to conclude with “future directions”

for academic research. In this subsection, we take the unusual step of
going back further to consider possible interventions for young people
who will work with and create future social media and SMA systems.
Thus, we provide our view of how SMA is currently taught in higher
education. In particular, we offer some guidance to instructors teaching
SMA for college students.
Big Data (such as textual and visual information) has become a
newly impactful domain for everyday business operations. All kinds
of organizations are trying to utilize this new type of data to improve
their businesses. To respond to this phenomenon properly, colleges are
developing courses focused on applying Big Data Analytics to business
and marketing-related problems. Among a variety of SMA tools, the
primary focus of such courses has been applications of text mining
as an innovative analytical tool to examine unstructured qualitative
information, coupled with structured quantitative information. Thus,
SMA courses are expected to train students with real-world business
examples.
A pedagogical philosophy we would like espouse involves embracing
the principle of learning by doing. According to this principle, we wish
to delve into concepts and theories using associated hands-on exercises
with real-world data. Students are expected to struggle while attempting
to learn SMA tools, which is the learning-by-doing process. In this way,
courses can offer quantitative skills that can lead to specific business
operation plans from a conceptual understanding. In executing this
pedagogical philosophy, we present a three-step approach to tackle
each topic (e.g., product review analysis): (1) problem detection and
formulation → (2) data analysis → (3) interpretation for value creation.
This approach emphasizes each of the three steps equally to prepare
students for the emerging AI era, where mechanical, standardized data
analysis will be easily replaced by AI. We observe that some students
concentrate on the data analysis stage too much and take the first
and third steps less seriously in their learning of SMA. This learning
270 Trends, Future Research, and Education in SMA

approach will grow more vulnerable with the growth of AI, as AI will
be more autonomous in conducting standardized analysis.
In many statistics and analytics classes, students are given cleaned-
up data with prespecified objectives related to the task. For example,
students might be asked to identify the primary factors determining
restaurant preferences (using the 5-point rating system for individual
restaurants) on Yelp using a specified technique such as text analyt-
ics (to analyze and structure textual restaurant reviews from actual
consumers) and regression (to determine which restaurant features sig-
nificantly influence consumers’ restaurant preferences). That may be a
fine start to learn the models and tools, but in real-world analytics, the
research problems will be unclear, the data will be dirty and cluttered,
and the appropriate tools to use will be confusing to the researcher.
Therefore, students must be trained to strategically identify valuable,
proper, and implementable problems connected with the data. We have
seen many students who come up with big problems without properly
understanding the data or context. They expect sufficient data to be
available, but it is helpful for them to first learn about potential data
availability and acquire domain-specific knowledge (e.g., coffee shops)
with some broadly defined objectives (learning about local people’s
preferences for coffee shops and coffee) as a way to get started. One
helpful data source for students is Kaggle.com, which contains various
kinds of secondary data.
As students learn more about the data, they can specify what spe-
cific objectives are achievable. This conceptual understanding will also
dictate their analysis implementation efficiently. We see students who
are predetermined to use certain tools (e.g., neural networks, decision
trees) because they have learned about these programs that can im-
plement such tools. Then, they try to apply these tools without fully
considering how such tools can contribute to their practical problems
at hand. If students make sufficient progress in their analysis, they
usually want to go deeper to create practical value from their empirical
results. This step requires a higher level of understanding of the industry
and business, which is also a primary challenge in generating valuable
information.
5.2. SMA Education in Colleges 271

Students also want to learn how to present their project within the
given amount of time (e.g., 15 minutes). One common mistake in such
a presentation is spending too much time on the technical details (e.g.,
theoretical basics of neural network models) and nonessential informa-
tion (e.g., companies’ basic historical facts instead of how competitors
are progressing). Another common mistake during a presentation is not
allocating time wisely to deliver the main takeaways, while spending
most of the time covering basic statistical results and less on primary
actionable recommendations. Eventually, students need to deliver a
practically relevant business story.
Various courses on social media, social media marketing, and SMA
are offered in various formats that include both online and in-person
modes. Many colleges and universities cover SMA-related courses in
their undergraduate and graduate (MBA and data science) programs.
Although most of these courses attempt to emphasize both the practical
and technical aspects in a balanced manner, there is great variation in
how the two aspects should be balanced, according to the program and
the instructor. Business programs tend to cover the practical aspect
better than non-business programs from a strategic perspective, whereas
non-business programs (computer science and data science) tend to
train the technical aspect more deeply than understanding the practical
implications. There are online courses on SMA, including massive open
online courses (MOCCs) such as Coursera and edX. Most SMA courses
are offered as a part of institutions’ MBA and data science programs;
programs exclusively devoted to SMA are still rare.
As an example, Northwestern launched a social media marketing
specialization composed of five courses on Coursera in 2015. The five
courses are: (1) What is Social; (2) The Importance of Listening; (3)
Engagement and Nurture Marketing Strategies; (4) Content, Advertising
and Social IMC; and (5) The Business of Social. The program states
that the specialization has two objectives: (1) providing social analytics
tools and training to help students become social media influencers
and (2) providing the knowledge and resources needed to construct a
complete social media marketing strategy.
Similarly, the University of California, Berkeley offers a certificate
program called the MicroMasters® Program in Marketing Analytics on
272 Trends, Future Research, and Education in SMA

the edX MOOC platform. This program also focuses on marketing per-
spectives and strategies using marketing campaigns, product promotion,
and pricing. Students complete four graduate-level courses to obtain the
certificate: (1) Marketing Analytics: Marketing Measurement Strategy;
(2) Marketing Analytics: Price and Promotion Analytics; (3) Marketing
Analytics: Competitive Analysis and Market Segmentation; and (4) Mar-
keting Analytics: Products, Distribution and Sales. Other universities
that provide SMA-related courses include the University of Washington
(which offers Social Media Data Analytics using Python focused on data
collection, analysis, and visualization) and Emory University (which
offers Introduction to Social Media Analytics).
Social media platforms also offer a variety of courses on social media
and SMA. For example, Facebook offers Measure and Optimize Social
Media Marketing Campaigns focused on advertising on social media.
LinkedIn Learning, an online education platform offering video courses
in three categories (Business, Creative, and Technology), offers many
SMA-related courses, including (1) Social Media Marketing: ROI; (2)
Social Media Marketing Foundations; (3) Social Media Monitoring:
Strategies and Skills; (4) Social Media Marketing: Social CRM; (5) SEO
(Search Engine Optimization) for Social Media; (6) Introduction to
Social Media Strategy; (7) Social Media for Government; and (8) Social
Media Marketing for Small Business. Other institutions offering SMA
courses include: Udemy (Social Media Monitoring), HubSpot Academy
(Social Media Marketing Course: Get Certified in Social Media Strategy),
Quintly (Social Media Analytics), Alison.com (Social Media Strategist),
and Eduonix (Complete Guide to Social Media Marketing). The variety
of these courses and host platforms indicate the strong need for such
education and training. Educators who do not want to create course
content from scratch can access these extant resources to deliver their
educational purposes.
 6
Conclusions

In the past decade, social media has steadily and rapidly grown to
become a quintessential way of social communication. This phenomenon
is particularly salient in marketing, where communications between con-
sumers and firms and among consumers are essential. This monograph
attempted to provide not only a concise summary of the developments
in social media and SMA, but also future research guidance on the
subject. Further, as a general, practical, and flexible way to execute
SMA, we proposed the AAVF framework comprising four stages: (1)
Acquisition; (2) Analysis; (3) Value; and (4) Feedback.
Focusing on the roles of social media (as a substantive matter) and
SMA (as a method) from the marketing perspectives, we summarized
research advances pertinent to SMA that have arisen predominantly
over the past two decades by considering both consumers and firms’
standpoints, while trying to reflect what is happening in the real-world
social media field. Doing so also led us to examine the tools that have
been used to analyze social media data (such as text analytics, visual
analytics, and ML).
Given the current status of SMA related to Marketing, this mono-
graph offers desirable future directions within this research domain.

273
274 Conclusions

Research on the voice of the customer expressed on social media has

shown that firms are facing challenging times in managing their social
media data in their operations. SMA has great potential to present a
competitive advantage to those who have the knowledge and capac-
ity to implement it adequately. Therefore, we believe there are ample
opportunities for more solid outcome-oriented research and practices.
We highlighted the increasing roles of textual and visual information
in the social media space. For the past two decades, marketing schol-
ars have yielded significant results by examining the roles of textual
information, such as online product reviews and online fundraising com-
ments, with the advancing technologies pertaining to text analytics. On
the other hand, their contribution to understanding social media-based
visual information (such as brand pictures) has been relatively new.
Recently, studies on such visual information and its roles in consumer
decisions are on the increase, owing to the advancing AI techniques
capable of analyzing such information. We would like to emphasize the
bright prospects of this research stream. Further, most visual infor-
mation research is currently limited to static images because our AI
techniques for analyzing moving images is still primitive. Therefore, we
call for forward-looking pioneering research that investigates the roles of
moving images (such as the specific content of YouTube videos) related
to marketing problems.
From a substantive perspective, we maintain that SMA has not
accomplished its potential in contributing to societal health and welfare.
One area that requires our immediate attention is understanding how to
curb the spread of misinformation because such events have continually
undermined the foundations of our society. With the rising influence
of social media, these incidents will be more frequent and even more
harmful if we do not adequately address the problem. In regard to mar-
keting, such a problem is strongly pronounced in fake online consumer
reviews. There have been great efforts against such negative review
manipulation by practitioners and academics, which have produced only
limited success until now. Our increased efforts against this problem
will make social media a more trustworthy space.
Like many phenomena, social media has a dark side (such as the
rapid propagation of misinformation, fake reviews, firms’ manipulation,
 275

privacy invasion), yet we are optimistic that its positive powers can
and will outweigh its negative powers. Given the potential impacts of
social media on humanity, we maintain that these advances have only
just begun. With the growth and development of social media activities,
we expect to see more sophisticated and detailed research forthcoming.
In addition, we emphasize that research should maintain relevance to
the roles of social media in regard to humans and organizations, and
marketing scholars can play a key role in these endeavors.
 References

Abbasi, A. and H. Chen (2008). “CyberGate: A design framework and

system for text analysis of computer-mediated communication”. MIS
Quarterly. 32(4): 811–837.
Abbasi, A., Y. Zhou, S. Deng, and P. Zhang (2018). “Text analytics to
support sense-making in social media: A language-action perspec-
tive”. MIS Quarterly. 42(2): 427–464.
Aggarwal, P., R. Vaidyanathan, and A. Venkatesh (2009). “Using lexical
semantic analysis to derive online brand positions: An application
to retail marketing research”. Journal of Retailing. 85(2): 145–158.
Agnihotri, A. and S. Bhattacharya (2016). “Online review helpfulness:
Role of qualitative factors”. Psychology and Marketing. 33(11): 1006–
1017.
Ahuja, V. and S. Alavi (2018). “Using Facebook as a digital tool for
developing trust amongst consumers using netnography and social
media analytics: A study of Jet Airlines”. Journal of Relationship
Marketing. 17(3): 171–187.
Anderson, E. W. (1998). “Customer satisfaction and word of mouth”.
Journal of Service Research. 1(1): 5–17.
Andryani, R., E. S. Negara, and D. Triadi (2019). “Social media analyt-
ics: Utilization of social media for research”. Journal of Information
Systems and Informatics. 1(2): 193–205.

276
References 277

Aral, S. and D. Eckles (2019). “Protecting elections from social media

manipulation”. Science. 365(6456): 858–861.
Archak, N., A. Ghose, and P. G. Ipeirotis (2011). “Deriving the pricing
power of product features by mining consumer reviews”. Manage-
ment Science. 57(8): 1485–1509.
Aswani, R., A. K. Kar, and P. V. Ilavarasan (2017). “Detection of
spammers in Twitter marketing: A hybrid approach using social
media analytics and bio inspired computing”. Information Systems
Frontiers. 20(November): 515–530.
Bail, C. A. (2016). “Combining natural language processing and net-
work analysis to examine how advocacy organizations stimulate
conversation on social media”. Proceedings of the National Academy
of Sciences of the United States of America. 113(42): 11823–11828.
Basili, R., D. Croce, and G. Castellucci (2017). “Dynamic polarity lexi-
con acquisition for advanced social media analytics”. International
Journal of Engineering Business Management. 9(December): 1–18.
Batrinca, B. and P. C. Treleaven (2014). “Social media analytics: A sur-
vey of techniques, tools, and platforms”. AI and Society. 30(1): 89–
116.
Berger, J. and K. L. Milkman (2012). “What makes online content
viral?” Journal of Marketing Research. 49(2): 192–205.
Blei, D. M., A. Y. Ng, and M. I. Jordan (2003). “Latent dirichlet
allocation”. The Journal of Machine Learning Research. 3: 993–
1022.
Borgatti, S. P., M. G. Everett, and J. C. Johnson (2018). Analyzing
social networks. 2nd edition. Los Angeles, CA: Sage.
Brady, W. J., J. A. Wills, J. T. Jost, J. A. Tucker, and J. J. van Bavel
(2017). “Emotion shapes the diffusion of moralized content in social
networks”. Proceedings of the National Academy of Sciences of the
United States of America. 114(28): 7313–7318.
Brei, V. A. (2020). “Machine learning in marketing: Overview, learning
strategies, applications, and future developments”. Foundations and
Trends® in Marketing. 14(3): 173–236.
278 References

Briesch, R. and D. Iacobucci (1995). “Using neural networks to compare

theoretical models: An application to modeling persuasive commu-
nications”. In: Marketing Theory and Applications. Ed. by D. W.
Stewart and N. J. Vilcassim. Vol. 6. Chicago: AMA. 177–184.
Brooker, P., J. Barnett, and T. Cribbin (2016). “Doing social media
analytics”. Big Data and Society. 3(2): 1–12.
Brown, J., A. J. Broderick, and N. Lee (2007). “Word of mouth com-
munication within online communities: Conceptualizing the online
social network”. Journal of Interactive Marketing. 21(3): 2–20.
Buitelaar, P., P. Cimiano, and B. Magnini (2005). Ontology Learning
From Text: Methods, Evaluation and Applications. IOS Press.
Cambria, E., D. Rajagopal, D. Olsher, and D. Das (2013). “Big social
data analysis”. Big Data Computing. 13: 401–414.
Carim, L. and C. Warwick (2013). “Use of social media for corpo-
rate communications by research-funding organisations in the UK”.
Public Relations Review. 39(5): 521–525.
Chae, J., D. Thom, H. Bosch, Y. Jang, R. Maciejewski, D. S. Ebert, and
T. Ertl (2012). “Spatiotemporal social media analytics for abnormal
event detection and examination using season-trend decomposi-
tion”. IEEE Conference on Visual Analytics Science and Technology.
(October): 143–152.
Chang, Y., C. Ku, and C. Chen (2019). “Social media analytics:
Extracting and visualizing Hilton hotel ratings and reviews from
TripAdvisor”. International Journal of Information Management.
48(October): 263–279.
Chauhan, K. and A. Pillai (2013). “Role of content strategy in social
media brand communities: A case of higher education institutes in
India”. Journal of Product and Brand Management. 22(1): 40–51.
Chen, H., R. Chiang, and V. Storey (2012). “Business intelligence and
analytics: From big data to big impact”. MIS Quarterly. 36(4): 1165–
1188.
Chen, H., Z. Zheng, and Y. Ceran (2016). “De-biasing the reporting bias
in social media analytics”. Production and Operations Management.
25(5): 849–865.
Chen, X.-W. and X. Lin (2014). “Big data deep learning: Challenges
and perspectives”. IEEE Access. 2: 514–525.
References 279

Chevalier, J. and A. Goolsbee (2003). “Measuring prices and price

competition online: Amazon.com and BarnesandNoble.com”. Quan-
titative Marketing and Economics. 1(2): 203–222.
Chevalier, J. A. and D. Mayzlin (2006). “The effect of word of mouth on
sales: Online book reviews”. Journal of Marketing Research. 43(3):
345–354.
Chollet, F. and J. J. Allaire (2018). Deep Learning with R. Manning
Publications Company.
Cohn, D. Y. (2010). “Commentary essay on exploring origins of ethical
company/brand perceptions—A consumer perspective of corporate
ethics”. Journal of Business Research. 63(12): 1267–1268.
Cuello-Garcia, C., G. Pérez-Gaxiola, and L. van Amelsvoort (2020).
“Social media can have an impact on how we manage and investigate
the COVID-19 pandemic”. Journal of Clinical Epidemiology. 127:
198–201.
Das, S. R. and M. Y. Chen (2007). “Yahoo! for Amazon: Sentiment
extraction from small talk on the web”. Management Science. 53(9):
1375–1388.
Davenport, T. and J. Harris (2007). “Competing with multichannel
marketing analytics”. Advertising Age. 78(14): 16–17.
De Vries, L., S. Gensler, and P. S. Leeflang (2012). “Popularity of brand
posts on brand fan pages: An investigation of the effects of social
media marketing”. Journal of Interactive Marketing. 26(2): 83–91.
Decker, R. and M. Trusov (2010). “Estimating aggregate consumer
preferences from online product reviews”. International Journal of
Research in Marketing. 27(4): 293–307.
Del Vicario, M., A. Bessi, F. Zollo, F. Petroni, A. Scala, G. Caldarelli,
H. E. Stanley, and W. Quattrociocchi (2016). “The spreading of
misinformation online”. Proceedings of the National Academy of
Sciences of the United States of America. 113(3): 554–559.
Dellarocas, C. and C. A. Wood (2008). “The sound of silence in online
feedback: Estimating trading risks in the presence of reporting bias”.
Management Science. 54(3): 460–476.
Dellarocas, C., X. M. Zhang, and N. F. Awad (2007). “Exploring the
value of online product reviews in forecasting sales: The case of
motion pictures”. Journal of Interactive Marketing. 21(4): 23–45.
280 References

Deng, L. (2014). “A tutorial survey of architectures, algorithms, and

applications for deep learning”. APSIPA Transactions on Signal
and Information Processing. 3(April): 1–29.
Deyan, G. (2021). “How much time do people spend on social media in
2021?” techjury.net, url: https://techjury.net/blog/time-spent-on
-social-media/#gref.
DiStaso, M. W. and D. Bortree (2014). Ethical Practice of Social Media
in Public Relations. Routledge.
Domingos, P. (2012). “A few useful things to know about machine
learning”. Communications of the ACM. 55(10): 78–87.
Dong, J. Q. and C. Yang (2020). “Business value of big data analytics:
A systems-theoretic approach and empirical test”. Information and
Management. 57(1): 1–9.
Dzyabura, D. and H. Yoganarasimhan (2018). “Machine learning and
marketing”. In: Handbook of Marketing Analytics: Methods and Ap-
plications in Marketing Management, Public Policy, and Litigation
Support. Ed. by N. Mizik and D. M. Hanssens. Research Handbooks
in Business and Management Series. 255–279.
Einhorn, H. J. and R. M. Hogarth (1981). “Behavioral decision theory:
Processes of judgment and choice”. Journal of Accounting Research.
19: 1–31.
Eliashberg, J., S. K. Hui, and Z. J. Zhang (2007). “From story line
to box office: A new approach for green-lighting movie scripts”.
Management Science. 53(6): 881–893.
Ellison, N. B. (2007). “Social network sites: Definition, history, and
scholarship”. Journal of Computer-Mediated Communication. 13(1):
210–230.
Fan, W. and M. D. Gordon (2015). “The power of social media analytics”.
Communications of the ACM. 57(6): 74–81.
Feldman, R. and J. Sanger (2007). The Text Mining Handbook: Advanced
Approaches in Analyzing Unstructured Data. Cambridge University
Press.
Fossen, B. L. and D. A. Schweidel (2017). “Television advertising and on-
line word-of-mouth: An empirical investigation of social TV activity”.
Marketing Science. 36(1): 105–123.
References 281

Fowler, D. and D. Pitta (2013). “Advances in mining social media:

Implications for marketers”. Journal of Information Technology
Management. 24(4): 25–33.
Fu, T., A. Abbasi, and H. Chen (2008). “A hybrid approach to web
forum interactional coherence analysis”. Journal of the American
Society for Information Science and Technology. 59(8): 1195–1209.
Gandomi, A. and M. Haider (2015). “Beyond the hype: Big data con-
cepts, methods, and analytics”. International Journal of Information
Management. 35(2): 137–144.
Gangadharbatla, H. (2008). “Facebook me: Collective self-esteem, need
to belong, and Internet self-efficacy as predictors of the iGeneration’s
attitudes toward social networking sites”. Journal of Interactive
Advertising. 8(2): 5–15.
Garcia, D., Y. M. Kassa, A. Cuevas, M. Cebrian, E. Moro, I. Rah-
wan, and R. Cuevas (2018). “Analyzing gender inequality through
large-scale Facebook advertising data”. Proceedings of the National
Academy of Sciences of the United States of America. 115(27): 6958–
6963.
Ghani, N. A., S. Hamid, I. A. T. Hashem, and E. Ahmed (2019). “Social
media big data analytics: A survey”. Computers in Human Behavior.
101(December): 417–428.
Ghose, A. and P. G. Ipeirotis (2011). “Estimating the helpfulness and
economic impact of product reviews: Mining text and reviewer char-
acteristics”. IEEE Transactions on Knowledge and Data Engineering.
23(10): 1498–1512.
Ghose, A., P. G. Ipeirotis, and B. Li (2012). “Designing ranking systems
for hotels on travel search engines by mining user-generated and
crowdsourced content”. Marketing Science. 31(3): 493–520.
Gillin, P. and G. A. Moore (2007). The New Influencers: A Marketer’s
Guide to the New Social Media. Quill Driver Books.
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 com-
munication: Evidence from a field test”. Marketing Science. 28(4):
721–739.
282 References

Godes, D., D. Mayzlin, Y. Chen, S. Das, C. Dellarocas, B. Pfeiffer,

B. Libai, S. Sen, M. Shi, and P. Verlegh (2005). “The firm’s man-
agement of social interactions”. Marketing Letters. 16(3–4): 415–
428.
Godes, D. and J. C. Silva (2012). “Sequential and temporal dynamics
of online opinion”. Marketing Science. 31(3): 448–473.
Goodfellow, I., Y. Bengio, and A. Courville (2016). Deep Learning.
Cambridge: MIT Press.
Gopinath, S., J. S. Thomas, and L. Krishnamurthi (2014). “Investigat-
ing the relationship between the content of online word of mouth,
advertising, and brand performance”. Marketing Science. 33(2): 241–
258.
Hagen, L., K. Uetake, N. Yang, B. Bollinger, A. J. B. Chaney, D. Dzya-
bura, J. Etkin, A. Goldfarb, L. Liu, K. Sudhir, Y. Wang, J. R. Wright,
and Y. Zhu (2020). “How can machine learning aid behavioral
marketing research?” Marketing Letters. 31: 361–370.
Hajli, N. and X. Lin (2016). “Exploring the security of information
sharing on social networking sites: The role of perceived control of
information”. Journal of Business Ethics. 133(1): 111–123.
Harris, L. and C. Dennis (2011). “Engaging customers on Facebook:
Challenges for e-retailers”. Journal of Consumer Behaviour. 10(6):
338–346.
Hartmann, J., M. Heitmann, C. Schamp, and O. Netzer (2021). “The
power of brand selfies”. Journal of Marketing Research. (Forthcom-
ing): 1–19.
Hassan, S., J. Sanger, and G. Pernul (2014). “SoDA: Dynamic visual
analytics of big social data”. In: IEEE International Conference on
Big Data and Smart Computing (BIGCOMP). January. 183–188.
Hayat, M. K., A. Daud, A. A. Alshdadi, A. Banjar, R. A. Abbasi, Y. Bao,
and H. Dawood (2019). “Towards deep learning prospects: Insights
for social media analytics”. IEEE Access. 7(March): 36958–36979.
Hennig-Thurau, T., K. P. Gwinner, G. Walsh, and D. D. Gremler
(2004). “Electronic word-of-mouth via consumer-opinion platforms:
What motivates consumers to articulate themselves on the Internet?”
Journal of Interactive Marketing. 18(1): 38–52.
References 283

Hevner, A. R., S. T. March, J. Park, and S. Ram (2004). “Design science

in information systems research”. MIS Quarterly. 28(1): 75–105.
Ho, T.-H., S. Li, S.-E. Park, and Z.-J. M. Shen (2012). “Customer
influence value and purchase acceleration in new product diffusion”.
Marketing Science. 31(2): 236–256.
Holsapple, C. W., S. Hsiao, and R. Pakath (2018). “Business social media
analytics: Characterization and conceptual framework”. Decision
Support Systems. 110(June): 32–45.
Hu, Y., A. Xu, Y. Hong, D. Gal, V. Sinha, and R. Akkiraju (2019).
“Generating business intelligence through social media analytics:
Measuring brand personality with consumer-, employee-, and firm-
generated content”. Journal of Management Information Systems.
36(3): 893–930.
Huang, M.-H. and R. T. Rust (2018). “Artificial intelligence in service”.
Journal of Service Research. 21(2): 155–172.
Humphreys, A. (2010). “Megamarketing: The creation of markets as a
social process”. Journal of Marketing. 74(2): 1–19.
Iacobucci, D. (1996). Networks in Marketing. Thousand Oaks, CA: Sage.
Iacobucci, D. and N. Hopkins (1992). “Modeling dyadic interactions
and networks in marketing”. Journal of Marketing Research. 29(1):
5–17.
Iacobucci, D., M. Petrescu, A. Krishen, and M. Bendixen (2019). “The
state of marketing analytics in research and practice”. Journal of
Marketing Analytics. 7: 152–181.
Inkpen, A. C. and E. W. Tsang (2005). “Social capital, networks, and
knowledge transfer”. Academy of Management Review. 30(1): 146–
165.
Iyengar, R., C. van den Bulte, and T. W. Valente (2011). “Opinion
leadership and social contagion in new product diffusion”. Marketing
Science. 30(2): 195–212.
Jalali, N. Y. and P. Papatla (2016). “The palette that stands out: Color
compositions of online curated visual UGC that attracts higher con-
sumer interaction”. Quantitative Marketing and Economics. 14(De-
cember): 353–384.
284 References

Järvinen, J. and H. Karjaluoto (2015). “The use of web analytics for

digital marketing performance measurement”. Industrial Marketing
Management. 50: 117–127.
Jordan, M. I. and T. M. Mitchell (2015). “Machine learning: Trends,
perspectives, and prospects”. Science. 349(6245): 255–260.
Kaplan, A. M. and M. Haenlein (2011). “The early bird catches the
news: Nine things you should know about micro-blogging”. Business
Horizons. 54(2): 105–113.
Karahanna, E., S. X. Xu, and N. A. Zhang (2015). “Psychological
ownership motivation and use of social media”. Journal of Marketing
Theory and Practice. 23(2): 185–207.
Khan, G. F. (2018). Creating Value with Social Media Analytics. Cre-
ateSpace, Washington, USA: Kindle Edition.
Kietzmann, J. H., K. Hermkens, I. P. McCarthy, and B. S. Silvestre
(2011). “Social media? Get serious! Understanding the functional
building blocks of social media”. Business Horizons. 54(3): 241–251.
Kim, A. J. and E. Ko (2012). “Do social media marketing activities
enhance customer equity? An empirical study of a luxury fashion
brand”. Journal of Business Research. 65(10): 1480–1486.
Kordzadeh, N. and D. K. Young (2020). “How social media analytics
can inform content strategies”. Journal of Computer Information
Systems. (April): 1–13.
Kozinets, R. V. (2002). “The field behind the screen: Using netnography
for marketing research in online communities”. Journal of Marketing
Research. 39(1): 61–72.
Kozinets, R. V., K. De Valck, A. C. Wojnicki, and S. J. Wilner (2010).
“Networked narratives: Understanding word-of-mouth marketing in
online communities”. Journal of Marketing. 74(2): 71–89.
Kumar, A., R. Bezawada, R. Rishika, R. Janakiraman, and P. K. Kannan
(2016). “From social to sale: The effects of firm-generated content
in social media on customer behavior”. Journal of Marketing. 80(1):
7–25.
Laroche, M., M. R. Habibi, M. O. Richard, and R. Sankaranarayanan
(2012). “The effects of social media based brand communities on
brand community markers, value creation practices, brand trust and
brand loyalty”. Computers in Human Behavior. 28(5): 1755–1767.
References 285

Lau, R., S. Liao, K. F. Wong, and K. Dickson (2012). “Web 2.0 environ-
mental scanning and adaptive decision support for business mergers
and acquisitions”. MIS Quarterly. 36(4): 1239–1268.
Le, T. M. and S.-Y. Liaw (2017). “Effects of pros and cons of applying big
data analytics to consumers’ responses in an e-commerce context”.
Sustainability. 9(5): 798.
Lee, I. (2018). “Social media analytics for enterprises: Typology, meth-
ods, and processes”. Business Horizons. 61(2): 199–210.
Lee, T. Y. and E. T. Bradlow (2011). “Automated marketing research
using online customer reviews”. Journal of Marketing Research.
48(5): 881–994.
Leung, X. Y., B. Bai, and K. A. Stahura (2015). “The marketing
effectiveness of social media in the hotel industry: A comparison of
Facebook and Twitter”. Journal of Hospitality and Tourism Research.
39(2): 147–169.
Li, X. and L. M. Hitt (2008). “Self-selection and information role of
online product reviews”. Information Systems Research. 19(4): 456–
474.
Li, Y. and Y. Xie (2020). “Is a picture worth a thousand words? An
empirical study of image content and social media engagement”.
Journal of Marketing Research. 57(1): 1–19.
Liao, S.-H., J.-L. Chen, and T.-Y. Hsu (2009). “Ontology-based data
mining approach implemented for sport marketing”. Expert Systems
with Applications. 36: 11045–11056.
Lin, H., W. Fan, and P. Y. Chau (2014). “Determinants of users’ con-
tinuance of social networking sites: A self-regulation perspective”.
Information and Management. 51(5): 595–603.
Lipsman, A., G. Mudd, M. Rich, and S. Bruich (2012). “The power of
like”. Journal of Advertising Research. 52(1): 40–52.
Liu, B. (2012). Sentiment Analysis and Opinion Mining. San Rafael:
Morgan and Claypool Publishers.
Liu, L., D. Dzyabura, and N. Mizik (2020). “Visual listening in: Ex-
tracting brand image portrayed on social media”. Marketing Science.
39(4): 669–686.
286 References

Liu, X., D. Lee, and K. Srinivasan (2019). “Large-scale cross-category

analysis of consumer review content on sales conversion leveraging
deep learning”. Journal of Marketing Research. 56(6): 918–943.
Lovett, M. J., R. Peres, and R. Shachar (2013). “On brands and word
of mouth”. Journal of Marketing Research. 50(4): 427–444.
Luca, M. and G. Zervas (2016). “Fake it till you make it: Reputation,
competition, and yelp review fraud”. Management Science. 62(12):
3412–3427.
Ludwig, S., K. De Ruyter, M. Friedman, E. C. Brüggen, M. Wetzels,
and G. Pfann (2013). “More than words: The influence of affective
content and linguistic style matches in online reviews on conversion
rates”. Journal of Marketing. 77(1): 87–103.
Ma, L., B. Sun, and S. Kekre (2015). “The squeaky wheel gets the grease—
An empirical analysis of customer voice and firm intervention on
Twitter”. Marketing Science. 34(5): 627–645.
Malthouse, E. C., M. Haenlein, B. Skiera, E. Wege, and M. Zhang
(2013). “Managing customer relationships in the social media era:
Introducing the social CRM house”. Journal of Interactive Marketing.
27(4): 270–280.
Mathwick, C. and J. Mosteller (2017). “Online reviewer engagement:
A typology based on reviewer motivations”. Journal of Service
Research. 20(2): 204–218.
Mayzlin, D., Y. Dover, and J. Chevalier (2014). “Promotional reviews:
An empirical investigation of online review manipulation”. American
Economic Review. 104(8): 2421–2455.
McDonald, J. (2021). Social Media Marketing Workbook 2021. Kindle
Edition, JM Internet Group.
McIntyre, S. H., E. F. McQuarrie, and R. Shanmugam (2016). “How
online reviews create social network value: The role of feedback
versus individual motivation”. Journal of Strategic Marketing. 24(3–
4): 295–310.
Michaelidou, N. and M. Micevski (2019). “Consumers’ ethical percep-
tions of social media analytics practices: Risks, benefits and potential
outcomes”. Journal of Business Research. 104(November): 576–586.
References 287

Misirlis, N. and M. Vlachopoulou (2018). “Social media metrics and

analytics in marketing—S3M: A mapping literature review”. Inter-
national Journal of Information Management. 38(1): 270–276.
Missikoff, M. and R. Navigli (2002). Integrated Approach to Web Ontol-
ogy Learning and Engineering. IEEE Computer. 54–57.
Mitchell, T. M. (1997). Machine Learning. McGraw-Hill.
Moe, W. W., O. Netzer, and D. A. Schweidel (2017). “Social media
analytics”. In: Handbook of Marketing Decision Models. Chapter 16,
483–504.
Moe, W. W. and D. A. Schweidel (2017). “Opportunities for inno-
vation in social media analytics”. Journal of Product Innovation
Management. 34(5): 697–702.
Moe, W. W. and M. Trusov (2011). “The value of social dynamics
in online product ratings forums”. Journal of Marketing Research.
48(3): 444–456.
Moon, S. and W. A. Kamakura (2017). “A picture is worth a thousand
words: Translating product reviews into a product positioning map”.
International Journal of Research in Marketing. 34(1): 265–285.
Moon, S., M.-Y. Kim, and P. K. Bergey (2019). “Estimating deception
in consumer reviews based on extreme terms: Comparison analysis
of open vs. closed hotel reservation platforms”. Journal of Business
Research. 102(September): 83–96.
Moon, S., M.-Y. Kim, and D. Iacobucci (2021). “Content analysis of fake
consumer reviews by survey-based text categorization”. International
Journal of Research in Marketing. 38(2): 343–364.
Moon, S., Y. Park, and Y. S. Kim (2014). “The impact of text product
reviews on sales”. European Journal of Marketing. 48(11/12): 2176–
2197.
Mostafa, M. M. (2013). “More than words: Social networks’ text mining
for consumer brand sentiments”. Expert Systems with Applications.
40(10): 4241–4251.
Mu, J., E. Thomas, J. Qi, and Y. Tan (2018). “Online group influ-
ence and digital product consumption”. Journal of the Academy of
Marketing Science. 46(5): 921–947.
Murphy, K. P. (2012). Machine Learning: A Probabilistic Perspective.
MIT Press.
288 References

Nanne, A. J., M. L. Antheunis, C. G. van der Lee, E. O. Postma, S.

Wubben, and G. van Noort (2020). “The use of computer vision to
analyze brand-related user generated image content”. Journal of
Interactive Marketing. 50: 156–167.
Naylor, R. W., C. P. Lamberton, and D. A. Norton (2011). “Seeing
ourselves in others: Reviewer ambiguity, egocentric anchoring, and
persuasion”. Journal of Marketing Research. 48(3): 617–631.
Netzer, O., R. Feldman, J. Goldenberg, and M. Fresko (2012). “Mine
your own business: Market structure surveillance through text min-
ing”. Marketing Science. 31(3): 521–543.
Oh, C., S. Sasser, and S. Almahmoud (2015). “Social media analytics
framework: The case of Twitter and super bowl ads”. Journal of
Information Technology Management. 26(1): 1–18.
Oh, H. J., E. Ozkaya, and R. LaRose (2014). “How does online social
networking enhance life satisfaction? The relationships among online
supportive interaction, affect, perceived social support, sense of
community, and life satisfaction”. Computers in Human Behavior.
30(January): 69–78.
Ordenes, F. V., S. Ludwig, K. De Ruyter, D. Grewal, and M. Wetzels
(2017). “Unveiling what is written in the stars: Analyzing explicit, im-
plicit, and discourse patterns of sentiment in social media”. Journal
of Consumer Research. 43(6): 875–894.
Osborne, P. and D. Ballantyne (2012). “The paradigmatic pitfalls of
customer-centric marketing”. Marketing Theory. 12(2): 155–172.
Oulasvirta, A., T. Suomalainen, J. Hamari, A. Lampinen, and K. Karvo-
nen (2014). “Transparency of intentions decreases privacy concerns
in ubiquitous surveillance”. Cyberpsychology, Behavior, and Social
Networking. 17(10): 633–638.
Paek, H. J., T. Hove, Y. Jung, and R. T. Cole (2013). “Engagement
across three social media platforms: An exploratory study of a
cause-related PR campaign”. Public Relations Review. 39(5): 526–
533.
Pang, B. and L. Lee (2008). “Opinion mining and sentiment analysis”.
Foundations and Trends® in Information Retrieval. 2(1–2): 1–135.
References 289

Pauwels, K., Z. Aksehirli, and A. Lackman (2016). “Like the ad or the

brand? Marketing stimulates different electronic word-of-mouth con-
tent to drive online and offline performance”. International Journal
of Research in Marketing. 33(3): 639–655.
Pennebaker, J. W., R. L. Boyd, K. Jordan, and K. Blackburn (2015). The
Development and Psychometric Properties of LIWC2015. University
of Texas at Austin.
Pennycook, G. and D. G. Rand (2019). “Fighting misinformation on
social media using crowdsourced judgments of news source quality”.
Proceedings of the National Academy of Sciences of the United States
of America. 116(7): 2521–2526.
Punj, G. (2017). “Consumer intentions to falsify personal information
online: Unethical or justifiable?” Journal of Marketing Management.
33(15–16): 1402–1412.
Qiu, L., H. Rui, and A. B. Whinston (2014). “Effects of social networks
on prediction markets: Examination in a controlled experiment”.
Journal of Management Information Systems. 30(4): 235–268.
Ribarsky, W., D. X. Wang, and W. Dou (2014). “Social media analytics
for competitive advantage”. Computer and Graphics. 38(February):
328–331.
Riquelme, I. P., S. Román, and D. Iacobucci (2019). “Perceived decep-
tion in online consumer reviews: Antecedents, consequences, and
moderators”. In: Review of Marketing Research: Marketing in a
Digital World. Ed. by A. Rindfleisch and A. Mater. Vol. 16. Emerald.
141–166.
Rocklage, M. D. and R. H. Fazio (2020). “The enhancing versus back-
firing effects of positive emotion in consumer reviews”. Journal of
Marketing Research. 57(2): 332–352.
Rust, R. T., W. Rand, M.-H. Huang, A. T. Stephen, G. Brooks, and
T. Chabuk (2021). “Real-time brand reputation tracking using social
media”. Journal of Marketing. 85(4): 21–43.
Sacha, D., A. Stoffel, F. Stoffel, B. C. Kwon, G. Ellis, and D. A. Keim
(2014). “Knowledge generation model for visual analytics”. IEEE
Transactions on Visualization and Computer Graphics. 20: 1604–
1613.
290 References

Saxton, G. D. and L. Wang (2014). “The social network effect: The de-
terminants of giving through social media”. Nonprofit and Voluntary
Sector Quarterly. 43(5): 850–868.
Schneider, M. J. and D. Iacobucci (2020). “Protecting survey data on a
consumer level”. Journal of Marketing Analytics. 8: 3–17.
Schweidel, D. A. and W. W. Moe (2014). “Listening in on social media:
A joint model of sentiment and venue format choice”. Journal of
Marketing Research. 51(4): 387–402.
Scott, J. (2017). Social Network Analysis. 4th edition. Thousand Oaks,
CA: Sage Publications.
Simon, H. A. (1977). The New Science of Management Decision. Revised
Edition, Prentice-Hall, NJ.
Sinha, V., K. S. Subramanian, S. Bhattacharya, and K. Chaudhuri
(2012). “The contemporary framework on social media analytics as
an emerging tool for behavior informatics, HR analytics and busi-
ness process”. Management: Journal of Contemporary Management
Issues. 17(2): 65–84.
Sivarajah, U., Z. Irani, S. Gupta, and K. Mahroof (2020). “Role of big
data and social media analytics for business to business sustainabil-
ity: A participatory web context”. Industrial Marketing Management.
86(April): 163–179.
Smith, C. N. and H. H. Seitz (2019). “Correcting misinformation about
neuroscience via social media”. Science Communication. 41(6): 790–
819.
Soucek, B. (1991). Neural and Intelligent Systems Integration. New
York: John Wiley and Sons.
Sponder, M. (2012). Social Media Analytics: Effective Tools for Building,
Interpreting, and Using Metrics. Kindle Edition, McGraw Hill.
Stephen, A. T. (2016). “The role of digital and social media marketing
in consumer behavior”. Current Opinion in Psychology. 10: 17–21.
Stephen, A. T. and J. Galak (2012). “The effects of traditional and
social earned media on sales: A study of a microlending marketplace”.
Journal of Marketing Research. 49(5): 624–639.
Sutton, R. S. and A. G. Barto (2018). Reinforcement Learning: An
Introduction. MIT Press.
References 291

Thomas, J. and K. Cook (2005). Illuminating the Path: The Research

and Development Agenda for Visual Analytics. Richland, WA, USA:
Pacific Northwest National Laboratory.
Tirunillai, S. and G. J. Tellis (2012). “Does chatter really matter? Dy-
namics of user-generated content and stock performance”. Marketing
Science. 31(2): 198–215.
Tirunillai, S. and G. J. Tellis (2014). “Mining marketing meaning from
online chatter: Strategic brand analysis of big data using latent
dirichlet allocation”. Journal of Marketing Research. 51(4): 463–479.
Toubia, O. and O. Netzer (2017). “Idea generation, creativity, and
prototypicality”. Marketing Science. 36(1): 1–20.
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.
Tsimonis, G. and S. Dimitriadis (2014). “Brand strategies in social
media”. Marketing Intelligence and Planning. 32(3): 328–344.
van Gerven, M. and S. Bohte (2018). Artificial Neural Networks as
Models of Neural Information Processing. Lausanne, Switzerland:
Frontiers Media.
Verhoef, P., E. Kooge, and N. Walk (2015). Creating Value with Big Data
Analytics: Making Smarter Marketing Decisions. London: Routledge.
Wang, Y., Q. Deng, M. Rod, and S. Ji (2020). “A thematic exploration
of social medial analytics in marketing research and an agenda for
future inquiry”. Journal of Strategic Marketing. (April): 1–21.
Wasserman, S. and K. Faust (1994). Social Network Analysis: Methods
and Applications. Cambridge, UK: Cambridge University Press.
Wu, X., X. Zhu, G.-Q. Wu, and W. Ding (2014). “Data mining with
big data”. IEEE Transactions on Knowledge and Data Engineering.
26(1): 97–107.
Yu, M., Z. Li, Z. Yu, J. He, and J. Zhou (2020). “Communication
related health crisis on social media: A case of COVID-19 outbreak”.
Current Issues in Tourism, ahead-of-print.
Yu, Y., W. Duan, and Q. Cao (2013). “The impact of social and conven-
tional media on firm equity value: A sentiment analysis approach”.
Decision Support Systems. 55(4): 919–926.
 292 References

Zhong, N. and D. A. Schweidel (2020). “Capturing changes in social me-

dia content: A multiple latent changepoint topic model”. Marketing
Science. 39(4): 827–846.

View publication stats