Professional Documents
Culture Documents
net/publication/358945387
CITATIONS READS
6 919
2 authors:
All content following this page was uploaded by Dawn Iacobucci on 07 August 2023.
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
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
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
Social Media
Marketing in SMA
Figure 1.1: The scope of this monograph at the intersection of social media, SMA,
and marketing.
219
220 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
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)
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
Acquisi�on
Feedback Analysis
Value
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):
229
230 Review of Social Media and SMA Related to Marketing
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
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)
Consumer Behavior eWOM; review posting; social Berger and Milkman (2012),
on Social Media media reviewers’ motivation Hayat et al. (2019), Kaplan and
and characteristics; Haenlein (2011), Khan (2018),
crowdsourcing; J-shaped rating Kietzmann et al. (2011), Lee
distribution; (2018), McDonald (2021), Sinha
positivity/negativity bias et al. (2012)
Social Media Impact Market structure and Chevalier and Mayzlin (2006),
on Firm product/brand position; sales Godes and Mayzlin (2009),
Performance forecasting; stock price and Gopinath et al. (2014), Kim
firm value changes by social and Ko (2012), Moe and
media activities; ecommerce Schweidel (2017), Moe and
sales conversion; web traffic Trusov (2011), Moon et al.
management; review volume (2014), Pauwels et al. (2016)
and polarity
Business/Marketing Marketing actions on social Aswani et al. (2017), Ma et al.
Strategy media; social media promotion (2015), Stephen and Galak
campaigns; managerial (2012), Tsimonis and
responses to consumers on Dimitriadis (2014), Verhoef
social media; digital marketing et al. (2015)
strategy; consumer
segmentation and profiling
Product/Brand Online brand communities; De Vries et al. (2012), Ho et al.
Management product positioning maps using (2012), Lovett et al. (2013),
product reviews; product Netzer et al. (2012), Rust et al.
innovation; raising brand (2021)
awareness; brand reputation
monitoring
Social Media Network structure (strong and Ahuja and Alavi (2018), Fan and
Network Analysis weak ties); member relationship Gordon (2015); Godes et al.
analysis (voting, tagging, (2005), Iyengar et al. (2011),
comments); social tagging; Lee (2018), Schweidel and Moe
influential reviewers; reviewers’ (2014), Wang et al. (2020)
dynamic growth and learning
Consumer Privacy Consumers’ consent to personal Michaelidou and Micevski (2019),
and Data Security data collection; consumers’ Oulasvirta et al. (2014),
on Social Media reaction to data breach; privacy Schneider and Iacobucci (2020)
laws
Fictitious/Biased Fake product reviews; content Chen et al. (2016), Dellarocas and
Content on Social manipulation by firms; bias in Wood (2008), Luca and Zervas
Media social media representation (2016), Moon et al. (2021)
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.
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
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
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
(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
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
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
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.
262
5.1. General Trends and Future Research Avenues in SMA 263
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
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
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
276
References 277
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
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