How does personalization affect brand

relationship in social commerce? A mediation

perspective
Trang P. Tran
Department of Marketing and Supply Chain Management, College of Business, East Carolina University, Greenville,
North Carolina, USA
Michelle van Solt
College of Business Administration, Valparaiso University, Valparaiso, Indiana, USA, and
James E. Zemanek, Jr
Department of Marketing and Supply Chain Management, College of Business, East Carolina University, Greenville,
North Carolina, USA

Abstract
Purpose – This paper aims to tests a conceptual model capturing the influence of personalized advertising on customer perceptions of brands in
social media and identifies three market segments based on customers’ reactions to personalized ads.
Design/methodology/approach – Two studies are developed to test the model using partial least squares structural equation modeling (PLS-SEM).
Additionally, cluster analysis, multi-group analysis (MGA) and serial mediation tests are also conducted to provide better insights into the results.
Findings – The results of the two studies show that all nine hypotheses are supported except for H4 in Study 1. Three market segments (ad lovers,
ad adjusters and ad haters) are identified. Each segment has a typical attitude toward personalized advertisements.
Research limitations/implications – Built on self-congruence literature, the current research posits that consumer-brand self-congruence can be
established when a customer sees a brand advertised on Facebook after searching for that brand online. Consistently, this paper finds that through
self-congruence, personalized advertising has a positive impact on brand-related outcomes.
Practical implications – Three segments identified – “ad lovers,” “ad adjusters” and “ad haters” are important for marketers. Companies should
develop an appropriate advertising campaign for each segment, especially once the general data protection regulation is in place. Companies will be
subject to a noncompliance penalty if an advertisement is posted on a user s account without approval. Identifying this segment promptly will not
only enable companies to save resources but also help avoid legal complications associated with privacy concerns.
Originality/value – This research sheds light on the effects of personalized advertising on customer perceptions of brands in social commerce.
Keywords Cluster analysis, Mediation, PLS-SEM, Personalized advertising, Facebook
Paper type Research paper

Introduction in electronic commerce (e-commerce) (Liang et al., 2011).

Capitalizing on these content creation tools, social networking
“Even with all the consumer data available, companies today sites such as Facebook, YouTube and Instagram have laid the
are challenged by the task of identifying ways in which brands foundation for users to buy or sell products through
can optimize personalization efforts that can benefit interactions and contributions (Liang and Turban, 2011). In
consumers” (Growth from Knowledge, 2019, p. 1). early 2018, one-quarter of US internet users actively
Social commerce (s-commerce), defined as “businesses built participated in s-commerce by purchasing a product they had
on internet-based social media that are open for users to be previously seen on a personalized social media ad (Garcia,
actively involved in buying and selling products or merchandise 2018). Applications of s-commerce increased by more than 500
in online communities and markets” (Shanahan et al., 2019, per cent from 2007 to 2008, leading s-commerce companies to
p. 63), has grown drastically over the past few years s- increase investments in venture capital finance (Stephen and
commerce is based on a key feature of Web 2.0 technology, Toubia, 2010). Facebook currently has over 2.41 billion active
allowing for content creation and promoting user interactions

The author thanks two students, Gretchen Gartman and Chelsea Brandt,
The current issue and full text archive of this journal is available on
who help with literature review and hypotheses development for this paper.
Emerald Insight at: https://www.emerald.com/insight/0736-3761.htm
Received 31 December 2017
Revised 30 August 2018
Journal of Consumer Marketing
29 April 2019
37/5 (2020) 473–486 28 October 2019
© Emerald Publishing Limited [ISSN 0736-3761] 14 February 2020
[DOI 10.1108/JCM-12-2017-2499] Accepted 21 February 2020

473
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

monthly users, and each user spends around 20 min per visit, also be defined as a firm’s opportunity to understand customer
making Facebook one of the best choices for s-commerce needs and provide proper care through customized advertising
(Zephoria Digital Marketing, 2019). messages (Imhoff et al., 2001) or as a company’s approach to
S-commerce (e.g. Starbucks’ Facebook and Toms’ Twitter) creating personal communications following customers’ stated
differs from e-commerce (e.g. eBay and Amazon) in that or implied interests (Roberts and Zanhay, 2012).
transactions are made within communities in s-commerce but Personalization has been applied to the online environment
implemented individually in e-commerce (Huang and because of recent advances in technology and has become a
Benyoucef, 2013). Customers in s-commerce are no longer solution for online marketers (Tran, 2017). It is integrated into
passive as with traditional media (i.e. television (TV) or websites to make them more relevant (Pierrakos et al., 2003).
newspapers); instead, they play an active role in creating and This research focuses on personalized ads posted by
sharing brand messages. This content sharing leads to viral s-commerce firms on Facebook and defines such ads as “the
effects among social media users and motivates customers to process of advertising in which a retailer develops a
share information online and spread word-of-mouth messages personalized ad of a product or service on Facebook based on
(Kozinets et al., 2010). The opportunity to interact with brands prior customer activities on the internet”.
(Cova and Dalli, 2009) through social media enhances brand Personalized ads have both positive and negative
image and significantly influences consumer behavior implications. On the positive side, they are a cost-saving
(Christodoulides, 2008). solution compared to mass advertising. Through
Advertising strategies became significantly more efficient personalization, only relevant messages are sent to targeted
after Facebook had partnered with IBM in 2015 by using groups of customers (Kim et al., 2001). Personalized
IBM’s targeting technology to develop more effective advertising based on customers’ interests or demographic data
advertising solutions on Facebook (Rudolph, 2015). This helps customers find products or services they may potentially
collaboration allowed s-commerce companies to display ads on want. Therefore, personalized ads are seen as more informative
Facebook users’ accounts based on their previous online and trustworthy than traditional advertising (Kim and Han,
activities. With support from this technology, s-commerce 2014), and thus, positively influence consumers’ attitudes
companies can send personalized ads to target users. Research toward the brand (De Keyzer et al., 2015; Shimp, 1981).
finds that personalized advertising helps reduce consumers’ Additionally, customers find personalized advertisements more
resistance to ads, increases ad credibility and raises brand appealing because they provide information that helps them
awareness (Baek and Morimoto, 2012; Xu, 2006). fulfill their needs (Boerman et al., 2017). Although
Although growing research examines the effects of personalization is effective in capturing consumers’ attention,
personalized ads in traditional media, research exploring the too much personalization can be risky, potentially leading to
effects of personalization on s-commerce remains scarce (Awad increased discomfort (Malheiros et al., 2012). Customers may
and Krishnan, 2006; Tucker, 2014). This paper aims to resolve associate personalization with privacy intrusion or data
this issue through three primary goals: management failure. Privacy concerns with personalized ads
1 to propose and test a conceptual model capturing the could result in reactance toward the ads (Tucker, 2014). If
influence of personalized advertising on customer consumers’ privacy is violated, unfavorable feelings, such as
perceptions of brands; avoidance and skepticism, may emerge (Tran, 2017).
2 to examine the mediation effects of two mediators:
consumer–brand engagement and consumer–brand Self-congruence with brands
identification; and According to the self-concept theory, individuals can maintain
3 to develop market segments based on the degree to which their self-concept through the brands and products they own
customers react to personalized ads. (Graeff, 1996). When consumers believe their ideal or actual
The rest of the paper is as follows: first, the literature on self-concept fits a brand’s personality, they develop more self-
perceived personalization and self-congruence is discussed, congruence with the brand (Wallace et al., 2017), resulting in
followed by hypotheses development. Next, a methodology desirable outcomes such as greater brand engagement (Coelho
including an assessment of the measurement model, structural et al., 2018), stronger brand loyalty (Sirgy and Danes, 1982),
model, serial mediation test and cluster analysis is discussed. higher purchase intention and improved self-esteem (Aguirre-
The paper ends with a discussion and theoretical and Rodriguez et al., 2012).
managerial implications. The growth of Web 2.0 has paved the way for the self-
concept to expand to the digital environment. Technology
gives customers an outlet to easily update their online persona
Theoretical background via a simple blog post, tweet, like or comment. For this reason,
Perceived personalization s-commerce sites like Facebook are considered an essential
The term “personalized marketing” can be traced back to the platform for psychological development and “a key part of self-
1870s (Ross, 1992). Personalization can mean tailoring a presentation for one-sixth of humanity” (Belk, 2013, p. 484).
product or service, customizing message content or matching Furthermore, personal possessions in the online setting have
geographical location with a customer’s personal information become less tangible and more digital (Belk, 2013). Instead of
(Wind and Rangaswamy, 2001). However, its definition has having physical picture albums, consumers store pictures on
changed over time. Personalization is defined as a company’s the cloud and have replaced iPods with Spotify lists. Social
practice of converting customer data to meaningful information media users focus less on actual ownership and more on
for developing a solution (Peppers and Rogers, 1998). It can publicly sharing digital possessions through the brands they

474
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

use, like or follow. Because brands can signal personality traits other positive outcomes such as brand pride, advocacy and
of the self, the brand’s customers interact with online are loyalty (Coelho et al., 2018; Stokburger-Sauer et al., 2012)
signals to others of their self-brand connections and posited that customer–brand identification becomes strong
preferences. Any activities consumers engage in on social media when a consumer’s perceived state of oneness with a brand is
platforms, such as likes, comments or brand interactions, established. Personalized advertising on Facebook allows users
provide clues to their online extended self to others. Online to associate themselves with a brand whose personality reflects
marketers also use this information to create personalized ads their own. This association makes consumers feel unique and
to draw more customers’ attention. prestigious. Facebook also gives users the ability to “share”
advertisements of brands with which they identify. This means
Hypotheses development the more personalized (i.e. relatable) a brand advertisement is,
Antecedents of customer brand engagement the more strongly a customer identifies him or herself with the
Customer brand engagement is a consumer’s overall brand. Therefore,
involvement and emotional investment with a brand
H2. The more personalized a brand advertisement is on
(Hollebeek et al., 2014) or the investment of cognitive,
Facebook, the more strongly a customer identifies him
emotional and behavioral resources into interactions with a
brand (Eigenraam et al., 2018). S-commerce is an interactive or herself with the brand.
platform that enables brands to create real-time customer Customer–brand engagement involves many different
engagement and strengthens consumer–brand relationships behaviors such as purchase, loyalty and becoming part of a
(Pillai and Chauhan, 2015). Digital customer engagement is a brand community. Social media platforms are a great
customer’s online interactions with a brand such as talking
medium that connects consumers and organizations,
about, learning about or working for a brand (Eigenraam et al.,
creating immediate consumer engagement. Engaged
2018). Customer–brand engagement results in numerous
consumers who are willing to purchase a brand advertised
positive outcomes, including increased sales growth and
on social media should feel some level of self-identification
profitability, low operational costs, positive word-of-mouth,
with the brand. For instance, if a customer clicks on a
enhanced co-creative experiences (Hollebeek et al., 2014) and
more competitive advantages (Eigenraam et al., 2018). Facebook ad with the intent to buy or share, then the
Personalized advertising, which uses consumer data (“e.g. customer is considered engaged with the brand and
names, past buying history, demographics, psychographics, identifies him or herself with it Hollebeek et al. (2014).
locations and lifestyle interests”), is used by companies to Similarly, when consumers interact with brands on
maintain engagement with their customers (Baek and Facebook through likes and shares, they are engaging with
Morimoto, 2012, p. 59). Because of social media, online that brand, becoming more familiar with it and increasing
marketers can more efficiently reach their target audience using the predictability of the brand’s behavior for those
personalized mechanisms. In a recent study, Dehghani and customers (Coelho et al., 2018). Highly engaged customers
Tumer (2015) suggested that Facebook advertisements offer usually show that they self-identify with a brand by posting
users the opportunity to interact with ads and allow them to about it on their Facebook page, providing their friends
like, share and view, which of their friends liked or shared the and acquaintances on this platform with identity clues.
same ads. The more personalized (i.e. the more relatable) Hence,
consumers perceive an ad to be, the more likely they are to
engage with it through shares or likes. Therefore, H3. The more engaged a customer is with a brand advertised
on Facebook, the more strongly the customer identifies
H1. The more personalized a brand advertisement is on him or herself with the brand.
Facebook, the more a customer is likely to engage with
the brand.
Antecedents of brand relationship
Marketing has recently evolved from traditional marketing to
Antecedents of customer–brand identification
relationship marketing. The branding literature confirms that
A person’s actual need for identification can be divided into
establishing consumer–brand relationship is crucial to brand
three parts, namely, a need to know herself, to feel relatively
success. All major findings stress the importance of forming
unique about herself and to feel good about herself. Once brand
identity – a set of special characteristics, such as those that are high-quality relationships between brands and potential
unique, distinctive or prestigious, that distinguishes a brand customers (Tho et al., 2016). Consumer–brand identification
from others – is consistent with a person’s need for allows consumers to establish a relationship based on shared
identification, customer–brand identification is established personal values and personality traits. Consistently, customer
(Stokburger-Sauer et al., 2012). Customer–brand engagement with brands advertised on Facebook is directly
identification is “a consumer’s perceived state of oneness with a related to the customer–brand relationship. According to Baek
brand, [and] is a valid and potent expression of our quest for and Morimoto (2012), when consumers deem ads useful and
identity-fulfilling meaning in the marketplace of brands” valuable, they are more likely to be interested in knowing more
(Stokburger-Sauer et al., 2012, p. 407). about the brand. As a result, customer–brand engagement is
The development of customer–brand identification not only improved, ultimately resulting in stronger customer–brand
results in a customer–brand relationship but also generates relationship. Hence,

475
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

H4. The more engaged a customer is with a brand advertised customer has developed a sense of loyalty, he or she will be more
on Facebook, the stronger the relationship between the inclined to purchase from that brand after seeing it advertised on
customer and the brand. Facebook:

In a marketplace filled with similar products, a key question for H8. The more loyal a customer is to a brand advertised on
marketers is, “how do brands distinguish themselves from their Facebook, the more likely it is that the customer
competitors?” brands can embody, inform and communicate purchases the brand.
desirable identities (Stokburger-Sauer et al., 2012). Moreover,
customers tend to form attachments to these brand values and
other intangible qualities aligned with their self-concept. When Consumer–brand engagement and identification as
brands effectively communicate their personality that is mediators
consistent with that of customers, the relationship between As previously mentioned, perceived personalization has a
customers and brands develops. This relationship can be positive impact on consumer–brand engagement and
extended to social media platforms. If a brand advertised on subsequently on consumer–brand identification, ultimately
Facebook is recognizable to a customer, and the brand’s values strengthening the brand relationship. Therefore, we anticipate
promote the customer’s positive feelings, the relationship that perceived personalization affects brand relationships
between the brand and customer strengthens. Formally, through consumer–brand engagement and identification.
Formally (Figure 1),
H5. The more strongly a customer identifies him or herself
with a brand advertised on Facebook, the stronger the
relationship between the customer and brand. H9. Customer–brand engagement and identification
mediate the relationship between perceived
personalization and brand relationship.
Antecedents of brand loyalty and brand usage intent
Brand loyalty is an action or mentality arising after a customer
identifies strongly with a given brand based on shared values Methodology
and similar personalities. These commonalities enhance the
Pretest
customer’s positive attitudes toward the brand and result in
This research centers on online activities in which customers
trust and loyalty (Coelho et al., 2018). Stokburger-Sauer et al.
have previously engaged. These activities, which could be the
(2012) argued that for a brand to be considered strong in the
search, purchase or view of a product or brand online, are
eyes of customers, there should be a special connection
recorded in the form of cookies. Then, when consumers open
between the two. The connection is derived from the matching
their Facebook accounts, those data are used by Facebook as a
between a brand’s and a customer’s identity. The more a
source of information for online marketers to target appropriate
customer identifies with a brand, the stronger the relationship.
customers through personalized advertisements. In this pretest,
This is a key factor that keeps the customer on the side and
participants were asked to think of ads they had previously
motivates him or her to be involved in pro-brand activities such
as supporting the company’s goals and decisions (Coelho et al., encountered on Facebook. To determine whether the ads to
2018). Hence, which participants were exposed were personalized, 288
responses were collected. A simple t-test designed to compare
H6. The stronger the relationship between a customer and a the middle point (4) with the mean scores confirmed that the ad
brand advertised on Facebook, the more likely it is that was viewed as personalized (M = 3.12, SD = 2.17, t (df =
the customer will become loyal to the brand. 279) =
6.81 and p < 0.001), lending support to the
manipulation check being successful.
A brand that is effective in articulating “brand personality”
plays a role in enhancing relationships with existing and Study 1
potential customers. Customers are more likely to purchase
brands that match their self-concept and with which they Sampling and procedure
experience greater self-congruity (Liu and Mattila, 2017). After the pretest indicated that the ad was perceived as
Therefore, we expect that when customers have a strong personalized, another data set was collected to test the
relationship with a brand, they will be more inclined to conceptual model. Out of the sample of 422 responses
purchase the brand when possible. Formally, collected from Amazon Mechanical Turk, 309 completed

H7. The stronger the relationship between a customer and a

Figure 1 Conceptual model
brand advertised on Facebook, the more likely the
customer purchases the brand.

An increased interest in studying brand loyalty is apparent in

recent marketing research. Studies have found that customer
loyalty leads to repeat purchases (Hasnin, 2018). Consumers who
feel highly loyal toward a given brand will generally voice their
opinions about it, have a positive brand image, and likely purchase
products from that brand (Dehghani and Tumer, 2015). Once a

476
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

observations met requirements. Demographic information “no,” they were directed to the end of the survey. If they said
presented in Table I showed that respondents were relatively “yes,” they were directed to the next section of the study.
young (52.1 per cent between 20 and 30 years old), well Participants who passed the filter questions were asked
educated (50.81 per cent having bachelor degrees) and active to provide the name of a brand that they had seen in a
on Facebook (49.19 per cent spending one to three hours a personalized ad, explain the situation in which they had
day); the nonstudent population accounted for 71.2 per cent, previously been interacting with the brand online, and share
women 53.72 per cent, medium-income earners ($40,000- their thoughts about the ad. Then, they continued to the
79,999) 30.42 per cent and US residents 65.05 per cent. questionnaire. The survey concluded with demographics, and
Before respondents took the survey, a few questions were respondents were thanked for their participation.
asked to determine their eligibility, such as “are you older than The results revealed that the brand’s respondents saw on
eighteen?” or “are you on Facebook?” Then, participants read Facebook ranged from more popular brands like Amazon, Dell,
the definition of personalized advertising: Samsung, eBay, Walmart, Lenovo, Macy’s and Groupon.com
Personalized advertising on Facebook is defined as the process of advertising to less popular brands like godaddy.com, Angry Birds, Burt’s
in which a retailer develops a personalized ad of a product or service on Bees, Flipkart, Hoover and La Quinta (Table II).
Facebook based on prior customer activities on the internet.
Respondents’ feelings about the ads were mixed. Some were
Next, participants were asked whether they had seen positive, saying “exciting, interesting, and informative” or “I
personalized ads on their Facebook page. If they answered think the ad was nice and subtle. It was for clothes, so it was not
flashy, bright, or too ‘in your face.’ It made the user feel
Table I Demographic characteristics (Study 1) comfortable,” and some were neutral, as in “I am generally not
interested in seeing ads on Facebook.” A few respondents felt
Demographic characteristics n (%)
irritated, as in “I thought it was annoying. Why do I need
Age advertisings on Facebook? [. . .] Yes, I was irritated.” Some ads
Less than 20 11 3.56 may not have been related to: “headphones don’t have anything
20-30 161 52.10 to do with TVs. I wasn’t looking to buy a TV”.
31-40 87 28.16
41-50 33 10.68
More than 50 17 5.50 Measurements
The measures were adapted from related literature, including
Sex
perceived personalization (Srinivasan et al., 2002), consumer–
Male 143 46.28
brand engagement (Hollebeek et al., 2014), consumer–brand
Female 166 53.72
identification (Stokburger-Sauer et al., 2012), brand
Student or non-student relationship quality (Aaker et al., 2004), brand loyalty
Student 89 28.80 (Stokburger-Sauer et al., 2012) and purchase intention (Yoo
Non-student 220 71.20 and Donthu, 2001). The scale items were measured using a
Education seven-point Likert scale (1 = “strongly disagree” and 7 =
High school or lower 74 23.95 “strongly agree”).
Bachelor 157 50.81 All latent variables were the first-order constructs except for
Master 74 23.95 consumer–brand engagement and brand relationship quality,
PhD 3 0.97 which were high order constructs. For the higher-order
Postdoc 1 0.32 constructs, all items of the first-order variable were averaged,
and this average score was used as observed variables
Time used on Facebook per day
(indicators) in the structural equation modeling analysis
Less than 1 h 79 25.57
(Zeugner-Roth et al., 2015). Consumer–brand engagement
1-3 h 152 49.19
was measured by three indicators that were the average score of
3-5 h 41 13.27
cognitive process, affection and activation. Brand relationship
5-7 h 24 7.77
quality was measured by six indicators that had an average
More than 7 h 13 4.21
score of commitment, self-connection, interdependence,
Family income intimacy, passion and trust.
Less than $20,000 86 27.83
$20,000-$39,999 77 24.92
$40,000-$59,999 58 18.77 Analytical results
$60,000-$79,999 36 11.65 The partial least squares (PLS) approach
$80,000-$99,999 22 7.12 PLS-SEM has been extensively applied in strategic
More than $100,000 30 9.71 management, marketing (Hair et al., 2012) and management
information systems (Ringle et al., 2012). The model in this
Location
research explored the effects of personalization on other brand-
The USA 201 65.05
related constructs; therefore, the prediction-oriented variance-
India 100 32.36
Other 8 2.59 based approach with PLS-SEM was appropriate. The model
was estimated using SmartPLS 3 software (Ringle et al., 2014).

477
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

Table II Examples of respondents’ responses to personalized ads

Brand Situation Thought of ads
Amazon I noticed an ad for Amazon with a product I had previously put on my I thought it was creepy. I do not like that it is monitoring the websites I
wishlist. It reminded me to go back and buy the item am going to. It sounds harmless because it is just advertising a shirt
from an Amazon wishlist I like but what else are they monitoring?
Apple Browsing on a specific vendor’s website and later refreshed my I think the ad was nice and subtle. It was for clothes, so it was not
Facebook page, and the advertisements for this vendor now appeared flashy, bright, or too “in your face.” It made the user feel comfortable
on my Facebook page
Bluestone I was searching for a gift for my mate through Google. After some Exciting, interesting and informative
time, when I logged in to Facebook, I found many jewelers’
advertisements on my page. It was quite helpful to know about new
brands and company
Broadway I was doing research about some things I watched on television and I am generally not interested in seeing ads on Facebook. I would never
then went to my Facebook page click one for more information
Groupon. I used Groupon for Christmas presents and I have noticed that While I appreciate the websites that I frequent and purchase items
com Groupon.com is now advertised on the sidebar of my Facebook page from, it is a little creepy that everything I look for seems to pop on my
sidebar at some time or another
Hoover I was actually in search of something specific to my classroom when I I think Facebook reads too much into a lot of things I post. I know that
saw an advertisement on Facebook for a percentage of an online it is to make things relevant to me, but it is kind of creepy to have all
teacher’s site for resources that knowledge about me
Lego My family loves anything to do with Legos. We often look up Legos on Although I do not like ads popping up in general, it is not as annoying
different websites. After researching Lego sets, our ads on Facebook when it is something I am interested in buying
started revolving around Legos
Nike I did a search for shoes on JCPenney.com. Right after I clicked off of I thought it was a pain in my ass. Why do I need advertising on
their page, the advertisement came up Facebook? Do not I get enough of this everywhere else? Yes I was
irritated
Samsung I was looking for new headphones then I started getting all kinds of Headphones do not have anything to do with TVs, I was not looking to
TV electronic ads, especially TVs buy a TV

Assessment of measurement model engagement ( b = 0.876 and p < 0.01), and perceived
In assessing internal consistency reliability, three criteria were personalization and consumer–brand identification ( b = 0.180
evaluated: Cronbach’s alpha, composite reliability and factor and p < 0.05). However, consumer–brand engagement had a
loadings. Compared with cut-off values of 0.7, all criteria met positive impact only on consumer–brand identification ( b =
or exceeded requirements. The ranges of Cronbach’s alpha, 0.697 and p < 0.01) but not on consumer–brand relationship
composite reliability and factor loadings were 0.858-0.969, ( b = 0.078 and p > 0.05). A positive relationship existed
0.914-0.976 and 0.819-0.965, respectively, satisfying reliability between consumer–brand identification and consumer–brand
requirements. Additionally, construct validity was assessed relationship ( b = 0.759 and p < 0.01), between consumer–
through two criteria, namely, convergent and discriminant brand relationship and brand usage intention ( b = 0.370 and
validity. All values of averaged variance explained (AVE) was p < 0.01) and between brand loyalty and brand usage intention
greater than 0.5, confirming convergent validity. Following ( b = 0.458 and p < 0.01). In other words, eight hypotheses
Fornell and Larcker (1981), discriminant validity was tested capturing the main effects were supported except for H4, which
through the comparison of squared correlations of the posited that consumer–brand engagement had an impact on
constructs and AVE. Because those correlations were smaller consumer–brand relationship (Table IV and Figure 2).
than AVE as seen in Table III, discriminant validity was
Serial mediation test
confirmed. In sum, based on these findings, the requirements
A serial multiple mediator model was used to test the mediation
for validity and reliability were satisfied.
effect of both consumer–brand engagement and consumer–
Assessment of structural model brand identification on the relationship between perceived
The structural model was assessed through two criteria personalization and brand relationship (H9). This test was
suggested by Hair et al. (2013): coefficient of determination conducted based on a combination of the serial multiple
(R2) and predictive relevance (Q2). Particularly, R2 for brand mediator technique (Hayes, 2013) and the partial least squares
usage intention (0.639), consumer–brand engagement (0.767), method (Hair et al., 2013). Calculating the mediation effect
consumer–brand identification (0.738), brand loyalty (0.741) required the calculation of both the total indirect effects and the
and brand relationship (0.774) indicated medium-to-high direct effect as follows (Figure 3).
predictive power of corresponding constructs. This was also There were three indirect effects in this model. The first
supported by the positive values of Q2. (PER -> CBI -> REL) was 0.068, the second (PER -> CBE
Out of eight hypotheses testing main effects, seven were -> CBI -> REL) was 0.463 and the third (PER -> CBI ->
supported. More specifically, a positive relationship existed REL) was 0.137. The sum of three indirect effects constituted
between perceived personalization and consumer–brand the total indirect effect of 0.668. Because the bootstrap

478
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

Table III Construct loadings, validity and reliability (Study 1)

Scale items CR AVE a AVE > Con2 Loadings Mean SD t-value
Brand usage intent 0.954 0.837 0.935 0.837 > 0.604
1. It makes sense to use this brand instead of any other 0.924 0.924 0.010 95.219
brand, even if they are the same
2. Even if another brand has the same features as this 0.918 0.918 0.015 62.335
brand, I would prefer to use this brand
3. If there is another brand as good as this brand, I prefer to 0.931 0.931 0.011 82.341
use this brand
4. If another brand is not different from this brand in any 0.885 0.885 0.022 40.804
way, it seems smarter to use this brand
Consumer-brand engagement 0.975 0.928 0.961 0.928 > 0.767
1. Average score of all items of “cognitive processing” 0.965 0.965 0.005 198.823
factor
2. Average score of all items of “affection” factor 0.965 0.965 0.005 191.276
3. Average score of all items of “activation” factor 0.960 0.960 0.008 125.978
Consumer-brand identification 0.976 0.891 0.969 0.891 > 0.769
1. I feel a strong sense of belonging to this brand 0.959 0.959 0.005 175.132
2. I identify strongly with this brand 0.936 0.935 0.010 93.080
3. This brand embodies what I believe in 0.936 0.935 0.009 99.039
4. This brand is like a part of me 0.943 0.943 0.008 113.451
5. This brand has a great deal of personal meaning for me 0.946 0.946 0.007 128.723
Brand loyalty 0.914 0.779 0.858 0.779 > 0.741
1. I consider myself to be loyal to this brand 0.922 0.922 0.009 97.183
2. This brand would be my first choice 0.904 0.904 0.012 76.902
3. I will not buy other brands if this brand is available at the 0.819 0.817 0.034 24.180
store
Perceived personalization 0.962 0.837 0.951 0.837 > 0.767
1. Personalized advertising of this brand on Facebook 0.926 0.925 0.011 87.193
makes purchase recommendations that match my needs
2. I think that personalized advertising of this brand on 0.923 0.923 0.012 79.776
Facebook enables me to order products that are tailor-
made for me
3. Overall, personalized advertising of this brand on 0.890 0.890 0.018 50.103
Facebook is tailored to my situation
4. Personalized advertising of this brand on Facebook 0.896 0.896 0.012 72.985
makes me feel that I am a unique customer
5. I believe that personalized advertising of this brand on 0.938 0.938 0.008 118.459
Facebook is customized to my needs
Brand relationship quality 0.974 0.862 0.968 0.862 > 0.769
1. Average score of all items of “commitment” factor 0.943 0.943 0.007 130.976
2. Average score of all items of “self-connection” factor 0.950 0.950 0.006 146.307
3. Average score of all items of “interdependence” factor 0.880 0.880 0.014 62.265
4. Average score of all items of “intimacy” factor 0.945 0.945 0.008 125.553
5. Average score of all items of “passion” factor 0.961 0.961 0.005 184.716
6. Average score of all items of “trust” factor 0.888 0.888 0.013 66.295
Note: The seven-point response cues for each item are 1 = “strongly disagree” to 7 = “strongly agree”

confidence interval was entirely above 0 (0.561-0.775), this mediators of the link between perceived personalization and
effect was different from 0. brand relationship (Table V).
The direct effect of perceived personalization on a brand
relationship without mediators (consumer–brand engagement Cluster analysis
and consumer–brand identification) was 0.734, which was Following the guidelines from established literature (Punj and
significant (CI: 0.675-0.793). However, when mediators were Stewart, 1983), a cluster analysis was carried out, identifying
included as serial mediators, this effect decreased to 0.064, three clusters that represented three different segments: ad
which was not significant (CI:
0.061-0.189). With the haters, ad lovers and ad adjusters. Each was composed of those
variance accounting for 91.26 per cent, consumer–brand who consistently reacted to personalized advertising
engagement and consumer–brand identification were full (Table VI). Statistical chi-square ( x 2) tests showed that the

479
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

Table IV Results of hypotheses testing (Study 1)

Hypotheses Coefficients Sample mean Standard error t-statistics p-values
PER fi CBE H1: supported 0.876 0.876 0.024 37.135 0.000
PER fi CBI H2: supported 0.180 0.172 0.082 2.193 0.028
CBE fi CBI H3: supported 0.697 0.706 0.079 8.831 0.000
CBE fi REL H4: not supported 0.078 0.078 0.073 1.078 0.281
CBI fi REL H5: supported 0.759 0.760 0.053 14.220 0.000
REL fi LOY H6: supported 0.861 0.861 0.016 52.811 0.000
REL fi BUI H7: supported 0.370 0.371 0.074 5.013 0.000
LOY fi BUI H8: supported 0.458 0.457 0.079 5.775 0.000
PER fi CBE and CBI fi REL H9: supported Mediation test (VAF = 91.26%)
R2 (Q2) for BUI 0.639 (0.531)
R2 (Q2) for CBE 0.767 (0.711)
R2 (Q2) for CBI 0.738 (0.657)
R2 (Q2) for LOY 0.741 (0.574)
R2 (Q2) for REL 0.774 (0.665)
Notes: PER: perceived personalization, CBE: consumer brand engagement, CBI: consumer brand identification, REL: brand relationship, LOY: brand loyalty
and BUI: brand usage intent

Figure 2 Testing model using PLS (Study 1)

Figure 3 Mediation test (Study 1) three segments significantly differed in terms of related latent
variables.
Ad haters (n = 46). Consumers in this segment demonstrated
that they perceived the ad to be less personalized and had a
lower level of consumer–brand engagement and identification
and a less favorable attitude toward the brand (including brand
relationship quality, loyalty and usage intent).
Ad lovers (n = 98). Unlike ad haters, consumers in this
segment perceived the ad to be more personalized and had a
higher level of consumer–brand engagement and identification
and more favorable attitudes toward the brand.
Ad adjusters (n = 165). Characteristics of ad adjusters were
between those of the other two segments.
The findings confirmed that compared to ad haters, ad lovers
felt a brand’s ad on Facebook was more personalized and had
higher consumer–brand engagement and identification and a
more favorable attitude (exemplified by brand relationship
quality, loyalty and usage intent). Customers in the three
segments also differed in terms of Facebook time, income and
location but not in student status, sex, age and education.

480
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

Table V Mediation test results

Symbol Effect SD t LLCI ULCI Note
Path estimates c 0.734 0.030 24.467 0.675 0.793 (no mediators)
a1 0.876 0.024 36.500 0.829 0.923
a2 0.180 0.082 2.195 0.019 0.341
b1 0.078 0.073 1.068
0.065 0.221
b2 0.759 0.053 14.321 0.655 0.863
d21 0.697 0.079 8.823 0.542 0.852
Indirect effects Ind1 a1b1 0.068 0.064 1.065
0.057 0.194 PER ! CBE ! REL
Ind2 a1d21b2 0.463 0.071 6.565 0.325 0.602 PER ! CBE ! CBI ! REL
Ind3 a2b2 0.137 0.062 2.189 0.014 0.259 PER ! CBI ! REL
Ind Total a1b1 1 a1d21b2 1 a2b2 0.668 0.055 12.233 0.561 0.775
Direct effects c’ 0.064 0.064 1.000
0.061 0.189
Total effects c’ 1 a1b1 1 a1d21b2 1 a2b2 0.732 0.030 24.425 0.674 0.791
VAF 91.26%

Table VI Cluster profiling (Study 1)

Ad haters (n = 46) Ad adjusters (n = 165) Ad lovers (n = 98) Test statistic
Variables Mean SD Mean SD Mean SD F (2,306) Sig.
Perceived personalization 1.522 0.505 4.182 0.751 6.388 0.490 930.214 0.000
Consumer brand engagement 1.783 0.941 3.982 1.197 5.959 0.973 241.179 0.000
Consumer brand identification 1.826 1.102 4.030 1.479 5.694 1.247 130.804 0.000
Brand relationship quality 2.500 1.130 4.061 1.305 5.592 1.225 102.086 0.000
Brand loyalty 2.804 1.293 4.345 1.286 5.673 1.173 86.275 0.000
Brand usage intent 2.761 1.401 4.242 1.148 5.704 1.096 106.130 0.000
Age 2.717 0.911 2.521 0.860 2.755 1.016 2.265 0.106
Sex 1.587 0.498 1.503 0.502 1.571 0.497 0.844 0.431
Students vs non-students 1.783 0.417 1.703 0.458 1.694 0.463 0.666 0.514
Education 1.957 0.788 1.988 0.732 2.133 0.727 1.440 0.238
Facebook time 1.804 0.934 2.055 0.977 2.500 1.067 9.483 0.000
Income 3.391 1.706 2.855 1.654 2.255 1.326 9.121 0.000
Location 1.196 0.833 1.370 0.726 1.633 0.694 6.650 0.001
Categorical variables Count (%) Count (%) Count (%)
Sex
Male 19 0.13 82 0.57 42 0.29
Female 27 0.16 83 0.50 56 0.34
Students vs non-students
Students 10 0.11 49 0.55 30 0.34
Non-students 36.00 0.16 116 0.53 68 0.31
Location
The USA 43 0.21 116 0.58 42 0.21
India 1 0.01 45 0.45 54 0.54
Other 2 0.25 4 0.50 2 0.25

Conclusions Study 2
Based on the online data, Study 1 confirms that all hypotheses
Hypotheses testing
capturing the impacts of personalized ads on customer
perceptions of brands advertised on Facebook are supported In Study 2, three questions were added as control variables that
except H4. It further supports the mediation effects of measured a customer’s past interaction with a brand, including
consumer-brand identification and engagement. Three market whether a respondent had purchased (labeled as “purchase” in
segments with their typical attitudes toward the ads are Table VIII) or liked (“like”) a particular brand or whether his or
identified, including “ad Lovers”, “ad haters” and “ad her past experience (“experience”) had been positive or
adjusters”. negative. The rationale was that if a respondent had bought and

481
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

liked a brand in the past, his or her attitude toward that brand significant, further supporting that personalized ads had a
may have been positive. The rest of the survey was the same as positive impact on customer perception of brands advertised on
Study 1. Out of 590 online responses, 343 responses met the Facebook. The results of the multigroup analysis indicated that
requirements and were used for data analysis. no significant differences existed among corresponding groups
A measurement model was assessed in Study 2 in the same across those variables.
procedure as in Study 1[1]. The results confirmed that reliability
and validity (convergent and discriminant) were established. The
General discussion
structural model assessment showed that all eight hypotheses
capturing the main effects were supported (Table VII). The current research contributes to the marketing literature in
A mediation test was conducted, showing that consumer– several ways. First, although research has reported various
brand engagement and identification partially mediated the effects of personalization, the results remain inconclusive.
effect of personalization on the brand relationship (VAF = 74 Positive effects are reported in some research that includes
per cent). Therefore, H9 was supported. improving customer retention, customer satisfaction, brand
loyalty and brand awareness (Sahni et al., 2018; Shanahan
Multigroup analysis et al., 2019). However, negative effects are found in other
Multigroup comparison was implemented through PLS research concerning privacy violation and customer reactance
nonparametric procedure (Hair et al., 2016) on three control are also found (Chen et al., 2019). It is important to understand
variables mentioned earlier. Then, the findings indicated that why inconsistencies exist. Although several types of
no significant differences existed among the groups personalization have been used in research, this paper centers
(Table VIII). on a specific type created after a customer searches a brand
online. The findings confirm that its effects are positive. This
Conclusions paper sheds light on a mechanism underlying the role of
Study 2 was similar to Study 1, except that potential variables personalized ads on Facebook. If applied properly, they can be
were controlled. The findings showed that all main effects were an effective tool for online marketers.

Table VII Results of hypotheses testing (Study 2)

Hypotheses Coefficients Sample mean Standard error t-statistics p-values
PER fi CBE H1: supported 0.700 0.700 0.041 17.180 0.000
PER fi CBI H2: supported 0.143 0.140 0.084 1.698 0.090
CBE fi CBI H3: supported 0.526 0.530 0.095 5.534 0.000
CBE fi REL H4: supported 0.156 0.159 0.080 1.959 0.050
CBI fi REL H5: supported 0.534 0.536 0.065 8.213 0.000
REL fi LOY H6: supported 0.696 0.696 0.043 16.010 0.000
REL fi BUI H7: supported 0.515 0.516 0.061 8.406 0.000
LOY fi BUI H8: supported 0.363 0.363 0.060 5.999 0.000
PER fi CBE and CBI fi REL H9: supported Mediation test (VAF = 74%)
R2 (Q2) for BUI 0.815 (0.658)
R2 (Q2) for CBE 0.805 (0.675)
R2 (Q2) for CBI 0.708 (0.589)
R2 (Q2) for LOY 0.732 (0.564)
R2 (Q2) for REL 0.812 (0.629)

Table VIII Multigroup analysis results (Study 2)

Purchase Experience Like
Path Pooled data No Yes p (diff) Positive Negative p (diff) No Yes p (diff)
N 343 77 266 171 172 118 225
1 PER fi CBE 0.700 0.851 0.865 0.391 0.836 0.858 0.709 0.818 0.790 0.711
2 PER fi CBI 0.143 0.421 0.083 0.935
0.021 0.290 0.970 0.131 0.206 0.328
3 CBE fi CBI 0.526 0.437 0.708 0.121 0.711 0.565 0.188 0.569 0.577 0.486
4 CBE fi REL 0.156 0.252 0.168 0.624 0.274 0.158 0.236 0.317 0.060 0.963
5 CBI fi REL 0.534 0.529 0.610 0.305 0.585 0.569 0.458 0.522 0.668 0.099
6 REL fi LOY 0.696 0.841 0.802 0.784 0.712 0.831 0.989 0.802 0.768 0.749
7 REL fi BUI 0.515 0.520 0.553 0.414 0.466 0.609 0.899 0.430 0.565 0.138
8 LOY fi BUI 0.363 0.383 0.379 0.514 0.454 0.295 0.082 0.469 0.356 0.816
Note: The multi-group comparison is based on a non-parametric approach (MGA)

482
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

Second, the action of searching for a brand online itself is requirements, they also benefit from this policy. As Daniel
pertinent to branding research because consumers typically Gilbert of Econsultancy pointed out, GDPR has “a positive
tend to search for brands with which they feel a certain impact on the quality of the data used for targeting, the
connection. When consumers feel a connection between their relevance of ads, and the attitude toward those ads on behalf of
personality and that of a brand, self-congruence is established. the consumer” (Bannerflow, 2017, p. 15). Because companies
This effect not only enhances positive feelings about the could send ads to customers who approve, they will not waste
brands, such as brand preferences, attitude and loyalty (Hosany time and money promoting ads to those who do not. Resources
and Martin, 2012; Marshall et al., 2008) but also generates saved could be used to improve the quality of advertising
psychological benefits such as higher self-esteem, social campaigns through personalization mechanisms.
approval needs and social consistency (Aguirre-Rodriguez Finally, the information on three segments – “ad lovers,” “ad
et al., 2012; Hosany and Martin, 2012). Building on self- adjusters” and “ad haters” – is critical. The first segment, “ad
congruence literature, the current research posits that lovers” (32 per cent), has the highest potential. They like
consumer–brand self-congruence can be established online personalized ads, and their attitude is more favorable.
when a customer sees a brand advertised on Facebook after
Therefore, companies should invest marketing resources in
searching that brand online. Consistent with findings from the
converting this segment into customers. The second segment,
self-congruence research, this paper finds that through self-
“ad adjusters” (53 per cent), has a neutral attitude toward the
congruence, personalized advertising has a positive impact on
ads, less favorable than that of the “ad lovers” but more
brand-related outcomes such as consumer–brand
favorable than that of the “haters.” The last segment, “ad
identification, engagement, relationship, loyalty and purchase
intention. haters” (15 per cent), does not like personalized advertisements
Finally, brands not only provide utilitarian and emotional and has a generally negative attitude. Investing money in this
benefits but also help customers communicate something segment could be wasteful. It will be even more difficult once
about themselves (Tho et al., 2016). Marketers have spent the GDPR is in place. Companies will be subject to a
much time identifying drivers and outcomes of brand noncompliance penalty if an advertisement is posted on a user’s
relationships (Chang and Chieng, 2006). Particularly, the social media account without his or her approval. Identifying
brand relationship is driven by brand image, attitude or this segment promptly will not only enable companies to save
association (Chang and Chieng, 2006) and leads to a higher resources but also help avoid legal complications linked to
likelihood of sharing or purchasing again and a lower tendency privacy concerns.
to switch brands (Smit et al., 2007). However, this study goes
one step further by investigating the causes and effects of brand Limitations and future research
relationships in the social media context. Not only does the This research is not without limitations. First, the data
paper propose new antecedents (customer–brand engagement collected from Amazon Mechanical Turk are a questionable
and identification) and outcomes (brand loyalty and purchase issue brought up by literature, although other researchers still
intention) of brand relationships but also it shows a mediation recommend it because of low cost and accessibility (Huff and
mechanism that alters the effect of personalization on the brand Tingley, 2015; Tran, 2017). Future research could examine
relationship. other, potentially more representative data sources, such as
Managerially, this research provides specific students or actual users of social media.
recommendations for marketers. First, although growing Second, cross-sectional surveys used in this paper could
research shows the positive outcomes of online advertising cause common method bias (Podsakoff et al., 2003). The bias
(Tucker, 2014; Xu, 2006), other research criticizes should be addressed through a more rigorous analytical
personalization. Particularly, Hoffman and Fodor (2010) method. An accurate way to test the common method bias is
argued that social media sites have shifted the control power of through a marker variable, which is expected not to correlate
brands from firms to users, who are more involved in brand with the other latent constructs (Podsakoff et al., 2003). This
communication through comments and interactions. The variable should be considered in the survey design stage and
authors advised that online marketers are careful when
reflected in future research.
choosing a social networking site like Facebook to promote
Finally, causality versus correlation is another concern. This
their brands. Consistently, with an increasing number of
study has used PLS-SEM to test hypotheses. However, future
fake accounts on Facebook (Krombholz et al., 2012), the
research may incorporate experimental designs to test the
effectiveness of advertising on social media is questionable.
effects of perceived personalization. This means two scenarios
Nevertheless, this skepticism could be overcome if online
(personalized and control) can be developed, respondents
marketers know how to personalize their advertisements. If
randomly assigned to each scenario, and results compared
done right, personalization could enhance customer
perceptions of the brands advertised, as evidenced by this between scenarios. As the literature suggests, experimental
study’s findings. design is a suggested method to verify a causal relationship.
Next, general data protection regulation (GDPR), recently
approved by the European Union (EU), gives EU citizens full Note
control of their personal data, including name, home address,
bank details and Internet protocol address (Bannerflow, 2017). 1 Because of the length requirement, the detailed
Companies can now use the EU’s data only after the user has explanation of Study 2 is minimized. A full description is
approved it. Although companies are facing more rigorous provided upon request.

483
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

References Garcia, K. (2018), “The future of retail in 2019: social

commerce has its day”, [blog post], available at: www.
Aaker, J.L., Fournier, S. and Brasel, A.S. (2004), “When good
emarketer.com/content/social-commerce-could-finally-
brands do bad”, Journal of Consumer Research, Vol. 31 No. 1,
happen-but-not-in-the-way-marketers-thought-it-would
pp. 1-16, doi: 10.1086/383419.
Graeff, T.R. (1996), “Using promotional messages to manage
Aguirre-Rodriguez, A., Bosnjak, M. and Sirgy, M.J. (2012),
the effects of brand and self-image on brand evaluations”,
“Moderators of the self-congruity effect on consumer
Journal of Consumer Marketing, Vol. 13 No. 3, pp. 4-18.
decision-making: a Meta-analysis”, Journal of Business
Growth from Knowledge (2019), “Personalization”, [blog
Research, Vol. 65 No. 8, pp. 1179-1188.
post], available at: www.gfk.com/landing-pages/landing-
Awad, N.F. and Krishnan, M.S. (2006), “The personalization
pages-us/the-future-of-financial-services/personalization/f
privacy paradox: an empirical evaluation of information
Hair, J.F., Hult, G.T.M., Ringle, C. and Sarstedt, M. (2016),
transparency and the willingness to be profiled online for
A Primer on Partial Least Squares Structural Equation Modeling
personalization”, MIS Quarterly, pp. 13-28.
(PLS-SEM), SAGE Publications.
Baek, T.H. and Morimoto, M. (2012), “Stay away from me”,
Hair, J.F., Sarstedt, M., Pieper, T.M. and Ringle, C.M.
Journal of Advertising, Vol. 41 No. 1, pp. 59-76, doi: 10.2753/
(2012), “The use of partial least squares structural equation
JOA0091-3367410105.
modeling in strategic management research: a review of past
Bannerflow (2017), “What the hell is GDPR? What advertisers
practices and recommendations for future applications”,
need to know”, [blog post], available at: https://blog.
Long Range Planning, Vol. 45 Nos 5/6, pp. 320-340.
bannerflow.com/gdpr-online-advertising/ (accessed).
Hasnin, E.A. (2018), “A mediating role of customer value
Belk, R.W. (2013), “Extended self in a digital world”, Journal of
between customer engagement and loyalty: an applied study
Consumer Research, Vol. 40 No. 3, pp. 477-500.
Boerman, S.C., Kruikemeier, S. and Zuiderveen Borgesius, F. in commercial banks in Egypt”, International Journal of
J. (2017), “Online behavioral advertising: a literature review Marketing Studies, Vol. 10 No. 1, pp. 136-144.
and research agenda”, Journal of Advertising, Vol. 46 No. 3, Hayes, A.F. (2013), Introduction to Mediation, Moderation, and
pp. 363-376. Conditional Process Analysis: A Regression-Based Approach,
Chang, P.L. and Chieng, M.H. (2006), “Building Guilford Press.
consumer–brand relationship: a cross-cultural Hoffman, D.L. and Fodor, M. (2010), “Can you measure the
experiential view”, Psychology and Marketing, Vol. 23 ROI of your social media marketing?”, MIT Sloan
No. 11, pp. 927-959. Management Review, Vol. 52 No. 1, p. 41.
Chen, Q., Feng, Y., Liu, L. and Tian, X. (2019), Hollebeek, L.D., Glynn, M.S. and Brodie, R.J. (2014),
“Understanding consumers’ reactance of online personalized “Consumer brand engagement in social media:
advertising: a new scheme of rational choice from a conceptualization, scale development and validation”,
perspective of negative effects”, International Journal of Journal of Interactive Marketing, Vol. 28 No. 2, pp. 149-165,
Information Management, Vol. 44, pp. 53-64. doi: 10.1016/j.intmar.2013.12.002.
Christodoulides, G. (2008), “Breaking free from the industrial Hosany, S. and Martin, D. (2012), “Self-image congruence in
age paradigm of branding”, Journal of Brand Management, consumer behavior”, Journal of Business Research, Vol. 65
Vol. 15 No. 4, pp. 291-293. No. 5, pp. 685-691.
Coelho, P.S., Rita, P. and Santos, Z.R. (2018), “On the Huang, Z. and Benyoucef, M. (2013), “From e-commerce to
relationship between consumer–brand identification, brand social commerce: a close look at design features”, Electronic
community, and brand loyalty”, Journal of Retailing and Commerce Research and Applications, Vol. 12 No. 4,
Consumer Services, Vol. 43, pp. 101-110. pp. 246-259.
Cova, B. and Dalli, D. (2009), “Working consumers: the next Huff, C. and Tingley, D. (2015), “Who are these people?’
step in marketing theory?”, Marketing Theory, Vol. 9 No. 3, Evaluating the demographic characteristics and political
pp. 315-339. preferences of MTurk survey respondents”, Research &
De Keyzer, F., Dens, N. and De Pelsmacker, P. (2015), “Is this Politics, Vol. 2 No. 3, doi: 10.1177/2053168015604648.
for me? How consumers respond to personalized advertising Imhoff, C., Loftis, L., Geiger, J.G. and Imhoff, D.C. (2001),
on social network sites”, Journal of Interactive Advertising, Building the Customer-Centric Enterprise: Data Warehousing
Vol. 15 No. 2, pp. 124-134. Techniques for Supporting Customer Relationship Management,
Dehghani, M. and Tumer, M. (2015), “A research on Wiley, New York, NY.
effectiveness of Facebook advertising on enhancing purchase Kim, Y.J. and Han, J. (2014), “Why smartphone advertising
intention of consumers”, Computers in Human Behavior, attracts customers: a model of web advertising, flow, and
Vol. 49 No. 8, pp. 597-600, doi: 10.1016/j.chb.2015.03.051. personalization”, Computers in Human Behavior, Vol. 33,
Eigenraam, A.W., Eelen, J., Van Lin, A. and Verlegh, P.W. pp. 256-269.
(2018), “A consumer-based taxonomy of digital customer Kim, J.W., Lee, B.H., Shaw, M.J., Chang, H.L. and Nelson,
engagement practices”, Journal of Interactive Marketing, M. (2001), “Application of decision-tree induction
Vol. 44, pp. 102-121. techniques to personalized advertisements on internet
Fornell, C. and Larcker, D.F. (1981), “Evaluating structural storefronts”, International Journal of Electronic Commerce,
equation models with unobservable variables and Vol. 5 No. 3, pp. 45-62.
measurement error”, Journal of Marketing Research, Vol. 18 Kozinets, R.V., De Valck, K., Wojnicki, A.C. and Wilner, S.J.
No. 1, pp. 39-50, doi: 10.2307/3151312. (2010), “Networked narratives: Understanding word-of-

484
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

mouth marketing in online communities”, Journal of Shanahan, T., Tran, T.P. and Taylor, E.C. (2019),
Marketing, Vol. 74 No. 2, pp. 71-89. “Getting to know you: social media personalization as a
Krombholz, K., Merkl, D. and Weippl, E. (2012), “Fake means of enhancing brand loyalty and perceived
identities in social media: a case study on the sustainability of quality”, Journal of Retailing and Consumer Services,
the Facebook business model”, Journal of Service Science Vol. 47, pp. 57-65.
Research, Vol. 4 No. 2, pp. 175-212. Shimp, T.A. (1981), “Attitude toward the ad as a mediator of
Liang, T.P. and Turban, E. (2011), “Introduction to the consumer brand choice”, Journal of Advertising, Vol. 10
special issue social commerce: a research framework for No. 2, pp. 9-48.
social commerce”, International Journal of Electronic Sirgy, M.J. and Danes, J.E. (1982), Self-Image/Product-Image
Commerce, Vol. 16 No. 2, pp. 5-14. Congruence Models: testing Selected Models, ACR North
Liang, T.P., Ho, Y.T., Li, Y.W. and Turban, E. (2011), “What American Advances.
drives social commerce: the role of social support and Smit, E., Bronner, F. and Tolboom, M. (2007), “Brand
relationship quality”, International Journal of Electronic relationship quality and its value for personal contact”,
Commerce, Vol. 16 No. 2, pp. 69-90. Journal of Business Research, Vol. 60 No. 6, pp. 627-633.
Malheiros, M., Jennett, C., Patel, S., Brostoff, S. and Sasse, M. Srinivasan, S.S., Anderson, R. and Ponnavolu, K. (2002),
A. (2012), “Too close for comfort: a study of the “Customer loyalty in e-commerce: an exploration of its
effectiveness and acceptability of rich-media personalized antecedents and consequences”, Journal of Retailing,
advertising”, Proceedings of the SIGCHI Conference on Human Vol. 78 No. 1, pp. 41-50, doi: 10.1016/S0022-4359(01)
Factors in Computing Systems, ACM, pp. 579-588. 00065-3.
Marshall, R., Na, W. and Deuskar, S. (2008), “Endorsement Stephen, A.T. and Toubia, O. (2010), “Deriving value from
theory: how consumers relate to celebrity models”, Journal of social commerce networks”, Journal of Marketing Research,
Advertising Research, Vol. 48 No. 4, pp. 564-572. Vol. 47 No. 2, pp. 215-228.
Peppers, D. and Rogers, M. (1998), Enterprise One-to-One: Stokburger-Sauer, N., Ratneshwar, S. and Sen, S. (2012),
Tools for Building Unbreakable Customer Relationships in the “Drivers of consumer–brand identification”, International
Interactive Age, Piatkus. Journal of Research in Marketing, Vol. 29 No. 4, pp. 406-418,
Pierrakos, D., Paliouras, G., Papatheodorou, C. and doi: 10.1016/j.ijresmar.2012.06.001.
Spyropoulos, C.D. (2003), “Web usage mining as a tool for Tho, N.D., Trang, N.T.M. and Olsen, S.O. (2016), “Brand
personalization: a survey”, User Modeling and User-Adapted personality appeal, brand relationship quality and WOM
Interaction, Vol. 13 No. 4, pp. 311-372, doi: 10.1023/ transmission: a study of consumer markets in Vietnam”, Asia
A:1026238916441. Pacific Business Review, Vol. 22 No. 2, pp. 307-324.
Pillai, A. and Chauhan, K. (2015), “A netnographic analysis of Tran, T.P. (2017), “Personalized ads on Facebook: an effective
Facebook content strategy of world’s top 10 management marketing tool for online marketers”, Journal of Retailing and
institutes”, International Journal of E-Business Research, Consumer Services, Vol. 39, pp. 230-242.
Vol. 11 No. 3, pp. 1-17. Tucker, C.E. (2014), Social Networks, Personalized Advertising,
Podsakoff, P.M., MacKenzie, S.B., Lee, J.Y. and Podsakoff, and Privacy Controls, American Marketing Association.
N.P. (2003), “Common method biases in behavioral Wallace, E., Buil, I. and de Chernatony, L. (2017),
research: a critical review of the literature and “Consumers’ self-congruence with a ‘liked’ brand: cognitive
recommended remedies”, Journal of Applied Psychology, network influence and brand outcomes”, European Journal of
Vol. 88 No. 5, p. 879. Marketing, Vol. 51 No. 2, pp. 367-390.
Punj, G. and Stewart, D.W. (1983), “Cluster analysis in Wind, J. and Rangaswamy, A. (2001), “Customerization: the
marketing research: review and suggestions for application”, next revolution in mass customization”, Journal of Interactive
Journal of Marketing Research, Vol. 20 No. 2, pp. 134-148, Marketing, Vol. 15 No. 1, pp. 13-32, doi: 10.1002/1520-
doi: 10.2307/3151680. 6653(200124)15:1<13::AID-DIR1001>3.0.CO;2-#.
Ringle, C.M., Sarstedt, M. and Straub, D.W. (2012), “Editor’s Xu, D.J. (2006), “The influence of personalization in affecting
comments: a critical look at the use of PLS-SEM in MIS consumer attitudes toward mobile advertising in China”,
quarterly”, MIS Quarterly, Vol. 36 No. 1, pp. 3-14. Journal of Computer Information Systems, Vol. 47 No. 2,
Ringle, C.M., Wende, S. and Becker, J.M. (2014), SmartPLS pp. 9-19.
3, SmartPLS, Hamburg.. Yoo, B. and Donthu, N. (2001), “Developing and validating a
Roberts, M. and Zanhay, D. (2012), Internet Marketing: multidimensional consumer-based brand equity scale”,
Integrating Online and Offline Strategies, Cengage Learning. Journal of Business Research, Vol. 52 No. 1, pp. 1-14, doi:
Ross, N. (1992), “A history of direct marketing”, unpublished 10.1016/S0148-2963(99)00098-3.
paper, Direct Marketing Association, New York, NY. Zephoria Digital Marketing (2019), “The top 20 valuable
Rudolph, S. (2015), “Facebook advertising statistics and Facebook statistics”, [blog post], available at: https://zephoria.
trends”, [blog post], available at: www.business2community. com/top-15-valuable-facebook-statistics/ (accessed).
com/infographics/facebook-advertising-statistics-trends- Zeugner-Roth, K.P., Žabkar, V. and Diamantopoulos, A.
infographic-01238796#Cu1mYI4Lh13ZI3Op.97 (2015), “Consumer ethnocentrism, national identity, and
Sahni, N.S., Wheeler, S.C. and Chintagunta, P. (2018), consumer cosmopolitanism as drivers of consumer behavior:
“Personalization in email marketing: the role of a social identity theory perspective”, Journal of International
noninformative advertising content”, Marketing Science, Marketing, Vol. 23 No. 2, pp. 25-54, doi: 10.1509/
Vol. 37 No. 2, pp. 236-258. jim.14.0038.

485
 A mediation perspective Journal of Consumer Marketing
Trang P. Tran, Michelle van Solt and James E. Zemanek, Jr Volume 37 · Number 5 · 2020 · 473–486

About the authors Management, and has successfully presented her work at
multiple conferences including Association for Consumer
Trang P. Tran (PhD, University of North Texas) is an Research and Association of Marketing Theory and Practice. She
Assistant Professor of Marketing at East Carolina University. earned her PhD and her Master of International Business
His research interests include international marketing, degree at Florida International University and her bachelor’s
customer behavior and service marketing. He has been degree at Florida State University. Her previous work
accepted or published in the Journal of Macromarketing, experience is in digital marketing.
Journal of Retailing and Consumer Services, Services Marketing
Quarterly, Computers in Human Behavior, International Journal James E. Zemanek, Jr (PhD, Texas A&M University) is a
of Nonprofit and Voluntary Sector Marketing, International Professor of Marketing and Supply Chain Management at
Journal of Bank Marketing, etc. Also, his name appeared in East Carolina University. Prior to joining East Carolina, he
several proceedings of American Marketing Association, was the Director of Undergraduate Curriculum and Interim
Academy of Marketing Science, Society of Marketing Advances, Program Coordinator for the Industrial Distribution Program
Association of Marketing Theory and Practice and Decision in the College of Engineering at Texas A&M University. His
Science Institute. Trang P. Tran is the corresponding author areas of expertise include sales force training, manufacturer-
and can be contacted at: trantra17@ecu.edu distributor relations, sports marketing, electronic commerce
and marketing strategy. He has conducted numerous seminars
Michelle van Solt (PhD, Florida International University) is for both manufacturers and distributors, and has consulted for
a Marketing Assistant Professor at Valparaiso University. Her a variety of firms concerning sales force development, sports
scholarly research focuses on consumer motivations and marketing, branding strategy, electronic commerce and
emotions in experiential consumption, personalized online distribution strategy. He earned his BBA (1985), MS
advertising and consumer perceptions of business ethicality. (Marketing; 1987) and PhD in Business (1992) all from Texas
She has published her work at Society and Business Review, A&M University, College Station.
Social Responsibility Journal and Cross Cultural and Strategic

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

486