INTERNATIONAL JOURNAL OF ADVERTISING, 2017
VOL. 36, NO. 5, 798–828
https://doi.org/10.1080/02650487.2017.1348035
Marketing through Instagram influencers: the impact
of number of followers and product divergence on
brand attitude
Marijke De Veirman, Veroline Cauberghe and Liselot Hudders
Department of Communication Sciences, Faculty of Political and Social Science, Ghent University, Gent,
Belgium
ABSTRACT
ARTICLE HISTORY
Findings of two experimental studies show that Instagram
influencers with high numbers of followers are found more likeable,
partly because they are considered more popular. Important, only in
limited cases, perceptions of popularity induced by the influencer’s
number of followers increase the influencer’s perceived opinion
leadership. However, if the influencer follows very few accounts
him-/herself, this can negatively impact popular influencers’
likeability. Also, cooperating with influencers with high numbers of
followers might not be the best marketing choice for promoting
divergent products, as this decreases the brand’s perceived
uniqueness and consequently brand attitudes.
Received 11 August 2016
Accepted 7 June 2017
KEYWORDS
E-WOM; influencer
marketing; Instagram; social
influence; social media
marketing
Introduction
Recently, brands discovered the far-reaching impact and viral growth potential of forging
alliances with social media influencers to promote their products. Social media influencers
are referred to as people who have built a sizeable social network of people following
them. In addition, they are seen as a regard for being a trusted tastemaker in one or several niches. As brands continue to abandon traditional advertising techniques, efforts are
increasingly focused on these influencers to endorse their products among their followers
and beyond. These endorsements are likely to be interpreted as highly credible electronic
Word Of Mouth (eWOM) rather than paid advertising as they are often seamlessly woven
into the daily narratives influencers post on their Instagram accounts (Abidin 2016). This is
particularly desirable for brands as it appears to be more effective than traditional advertising tactics, due to higher authenticity and credibility, which subsequently leads to lower
resistance to the message (de Vries, Gensler, and Leeflang 2012). Therefore, by seeding a
certain message or a new product with these influencers, marketers aim to maximize the
diffusion of information through their social network (Weimann 1994; Keller and Berry
2003). Through their posts, influencers may influence a disproportionately large number
of others, possibly indirectly via a cascade of influence through their followers (Gladwell
2000). Today, 75% of marketers are using influencer marketing (Augure 2015).
CONTACT Marijke De Veirman
© 2017 Advertising Association
marijke.deveirman@ugent.be
INTERNATIONAL JOURNAL OF ADVERTISING
799
One of the major challenges for brands is to identify and select these so-called influencers who may have a strong impact on their target audience and influence them to try
and adopt new products and help diffuse them in their social network through their posts
(Momtaz, Aghaie, and Alizadeh 2011; Pophal 2016). Today, the number of followers, which
reflects network size and serves as an indication for popularity, is frequently used to identify these influential nodes. Accordingly, higher numbers of followers may result in larger
reach of the (commercial) message and may thus leverage the power of this specific type
of word-of-mouth at scale. Regarding their commercial potential, technologies have been
developed to identify and track relevant influencers for brands and connect with them
(e.g. Traackr, Little Bird,…). These platforms draw up criteria such as minimum 10.000
followers in order to be suitable as brand advocate. However, to our knowledge, no
research yet investigated how people perceive and evaluate influencers’ numbers of followers. Moreover, the reach of the message through an influencer should not be the only
criterion for successful persuasive communication. To increase the message’s impact one
should search for the most likeable, credible influencer who has a high value as an opinion
leader. The challenge for advertisers thus becomes to select the most efficient and suitable influencer, also keeping the type of product they want to promote in consideration.
Hence, in two studies, we aim to provide more insights in the characteristics that make
a social media influencer on Instagram efficient above and beyond their potential reach
through their large social network. Being a social networking site that provides users with
video- and photo-sharing possibilities, Instagram lends itself very well for eWOM purposes
because products and brands can be visually imaged and named in the caption of the
photo. Moreover, it is one of the most popular social networking sites and currently has
over 500 million active users and counting (Statista 2016a). Study 1 explores which Instagram influencer is the best marketing choice in terms of number of followers. Whereas
previous studies have examined influence and diffusion, mainly on Twitter (Cha et al.
2010; Weng et al. 2010), to identify a suitable brand advocate, it is essential to examine
whether this number of followers indeed is a suitable indicator for doing so. There is discussion about whether or not there exists a link between number of followers and opinion
leadership. Findings vary from a clear connection between number of followers and opinion leadership (e.g. Yoganarasimhan 2012; Feng 2016; Hwang 2015) to number of
followers being merely an indication for popularity rather than influence (e.g. Cha et al.
2010; Romero et al. 2011). The first study contributes to this on-going debate by shedding
more light on whether one’s number of followers contributes to his/her opinion leader
status and how it affects general likeability towards him/her. In particular, it is investigated
whether one’s number of followers may work as a cue – indicating one’s popularity – and
whether perceived popularity in turn might cause people to ascribe opinion leadership to
the person in question, which eventually affects the endorser’s overall likeability.
As relationships on Instagram do not always entail reciprocal activities, meaning that
one can freely choose to follow an account without the need to ask the other’s permission
and without the other feeling obliged to follow him/her back, besides the number of followers, also the number of followees (i.e. the number of accounts the influencer follows
him-/herself) and the combination of both may affect one’s perceptions of the influencer.
In popular literature, some ‘rules’ exist about the ideal ‘followers/followees ratio’ exist.
There are even online calculators that calculate one’s followers/followees ratio and explain
its meaning (e.g. tffratio.com). However, to our knowledge, no study has ever examined
800
M. DE VEIRMAN ET AL.
the impact of number of followees and the ratio between one’s number of followers and
followees. Therefore, it is investigated whether and how number of followees affects the
relationship between number of followers and likeability. In other words, the importance
of one’s ‘followers/followees ratio’ in terms of likeability is examined.
Whereas the commercial use of influencers is a growing global marketing phenomenon due to their capacity to shape purchase decisions, little is known about how the
endorsements they produce on social media platforms in exchange for payment or sponsored products and services affect attitudes towards the brands or products they endorse.
Abidin (2015, 2016) did some ethnographic research on how influencers insert advertisements for products and services in the textual and visual narration of their personal, everyday lives on social media. However, experimental research on the working of influencer
marketing is largely absent. Therefore, Study 2 assesses the brand effects of influencers
and examines the moderating role of number of followers on the advertising effectiveness
of influencers’ posts. In particular, effectiveness in terms of attitude towards the brand of
commercial posts containing endorsements of products with common versus divergent
product designs will be investigated. The effect of product type on brand attitudes is
expected to be affected by number of followers as this might have an impact on perceptions about the brand. Products with divergent product designs that may respond to people’s need for uniqueness, might be perceived as less unique when posted by an
influencer with a high number of followers compared to when it is promoted by an influencer with a moderate number of followers, which may eventually lower brand attitudes.
Hence, this study adds value by providing an understanding on how one’s number of
followers affects attitudes towards the influencer and the brands (s)he promotes. Insights
are of general relevance to word-of-mouth marketing and influencer marketing in particular and contribute to literature in several ways. First, this study theoretically contributes to
literature on influence and the dissemination of word-of-mouth. Also, it adds to the ongoing debate concerning opinion leadership and how to identify it. Next, this study sheds
more light on how influencers determine the perception of a brand through their
endorsements. In this sense, this study contributes to literature on heuristic processing
and naive theories by investigating how an influencer’s number of followers may affect
perceptions of the uniqueness of a product and accordingly brand attitudes. Moreover,
this study took into account the moderating impact of the empirically underexposed
metric number of followees.
Theoretical background
The emergence of influencer marketing
It has been well recognized in marketing and consumer behaviour literature that eWOM,
or the information consumers obtain from interpersonal sources, has stronger effects on
consumer decision-making than traditional advertising techniques (Goldsmith and Clark
2008). A similar message is perceived as more authentic and credible when it is communicated by a fellow consumer compared to than when it would have been put forward by
an advertiser. Consumers have always valued others’ opinions, however, the advent and
still growing popularity of social media has amplified the effects of peer recommendations, as it empowered consumers to share their opinions and experiences one-to-many.
INTERNATIONAL JOURNAL OF ADVERTISING
801
As consumers can freely create and disseminate brand-related information and voluntarily
display their brand preference to others through their social interactions, social media
such as Instagram or Facebook nowadays represent an ideal tool for eWOM (Boyd and
Ellison 2007; Jansen et al. 2009; Knoll 2016; Lyons and Henderson 2005).
The power of eWOM, defined as ‘any positive or negative statement made by potential,
actual, or former customers about a product or company, which is made available to a
multitude of people and institutions via the Internet’ (Hennig-Thurau et al. 2004, 39) has
been widely recognized and social media has amplified and accelerated its reach. Crucial
to the diffusion of eWOM, is the identification of opinion leaders, who exert a disproportionate amount of influence on others, an idea that has already been recognized decades
ago (Katz and Lazarsfeld 1955). Through their social media activities, nowadays’ digital
opinion leaders or influencers, are able to influence the attitudes, decisions and behaviours of their audience of followers (Watts and Dodds 2007; Lyons and Henderson 2005).
Moreover, as messages can be disseminated rapidly and easily, a viral effect or buzz might
be induced. This way, their influence does not only flow to their followers, but also spreads
among followers as they share the viral messages in their social networks (Thomas 2004).
At the same time, nowadays consumers are not only sceptical about traditional branddriven advertising, they are also empowered to bypass it as it is often found to be intrusive and disruptive. Consumers are able to advance forward to skip commercials or install
ad-blocking software, which makes it increasingly harder for brands to reach consumers.
As an answer that maximizes the advantages of word-of-mouth and bypasses shortcomings of traditional advertising techniques, such as avoidance and resistance (Fransen et al.
2015; Kaikati and Kaikati 2004), brands increasingly focus their efforts on so-called social
influencers. As opposed to directly targeting the target market through all kinds of advertising, brands aim to encourage highly followed and admired influencers who are
regarded as trustworthy, non-purposive opinion leaders, to talk about and recommend
their products on social media platforms. This way, brands may leverage the power of
word-of-mouth and market their products indirectly.
The working of influencer marketing
Influencers are content creators who accumulated a solid base of followers. Through blogging, vlogging or creating short-form content (e.g. Instagram, SnapChat, …) they provide
their followers an insight into their personal, everyday lives, their experiences and opinions. By involving influencers (e.g. by offering to test a product, organizing an exclusive
event, … or simply paying them), brands aim to stimulate influencers to endorse their
products and this way build up their image among influencers’ often huge base of
followers, a practice that is called influencer marketing. Unlike mainstream celebrities,
influencers are believed to be accessible, believable, intimate and thus easy to relate to as
they share the personal, usually publically inaccessible aspects of their life with their followers and interact with them in flesh (Abidin 2016; Schau and Gilly 2003). This may generate para-social interaction, which has been described as the illusion of a face-to-face
relationship with a media performer and makes consumers more susceptible to their opinions and behaviour (Colliander and Dahl
en 2011; Knoll et al. 2015). As influencers’
endorsements are highly personal and interwoven into the constant stream of textual and
visual narration of their personal lives, they will likely be perceived as the influencer’s unbiased opinions and may have relevant persuasive power (Abidin 2015). Moreover, due to
802
M. DE VEIRMAN ET AL.
its relative newness and the inexperience of consumers to influencer marketing strategies,
it is less likely to trigger persuasion knowledge which could render unfavourable attitudes
(Friestad and Wright 1994; Tutaj and Van Reijmmersdal 2012).
It is important for brands to approach an influencer who is well-liked by their audience
to endorse their products. Previous research for example found positive associations
between attitude towards the celebrity and attitude towards the brand (e.g. Amos, Holmes,
and Strutton 2008; Silvera and Austad 2004). Also, Schemer et al. (2008) found that pairing
a brand with positively evaluated artists results in positive attitudes toward the brand. Moreover, brands should be careful in picking the right influencer to endorse their brand and
decide who possesses the most appropriate and desired characteristics in relation to the
brand, as the image of the influencer may transfer to the brand by virtue of the endorsement. As consumers use brands to communicate their identity to others and evaluate
others based on their consumption behaviour (Elliott and Wattanasuwan 1998; Reed et al.
2012), the images that the brand conveys are of high importance.
Identifying influencers
Influencer marketing consists of identifying and targeting influential users and stimulate
them to endorse a brand or specific products through their social media activities. Just
like in many other word-of-mouth marketing strategies, a major challenge is the identification of a suitable opinion leader or influencer (Araujo, Neijens, and Vliegenthart 2017). As
higher numbers of followers may result in larger reach of the (commercial) message and
may thus leverage the power of this specific type of word-of-mouth at scale, today, the
number of followers is frequently used to identify influencers on social media.
Different studies have been conducted to measure online opinion leadership and identify opinion leaders. Of these, assessing one’s audience size or number of followers has
often been put forward as a first step to take in the quest for opinion leaders. Zhang and
Dong (2008) for example developed a roadmap to identify online opinion leaders in virtual communities in which the first step is finding out who are active users with large
followers. Concerning Twitter, Cha et al. (2010) proposed different types of influence a person has of which the first is the audience size of the user, referring to the number of followers a user has. A high number of followers could be advantageous to the exertion of
opinion leadership as ideas are spread more widely and rapidly and consequently, interpersonal influence is enhanced. However, it remains uncertain to what extent consumers
process this information and use it to assess an influencer on social media, in particular in
terms of opinion leadership. Therefore, the purpose of this paper is to explore the impact
of influencers’ number of followers on attitude formation, both in terms of attitudes
towards the influencer (i.e. influencer likeability) as in terms of attitudes towards the
brands (s)he promotes.
Study 1: assessing the likeability of an Instagram influencer
Hypothesis development
Assessing the influencer’s likeability
Today, consumers face a wide range of available sources to find information in their buying decision process. The internet and the advent of social media have made it possible
INTERNATIONAL JOURNAL OF ADVERTISING
803
to assimilate large amounts of information in a very short time and without substantial
costs to the user. Yet, the abundance and diversity of information also makes it difficult
for consumers to determine its value. Already in 1982, Simon stated in his theory of
bounded rationality that people have limited ability to process and evaluate available
information. Moreover, digital media have increased the complexity of determining credible sources and assessing a source in a digital context is more difficult than in traditional
face-to-face interaction decisions. Therefore, consumers are more dependent on cues and
heuristics to evaluate information sources. Due to the emergence and ever-growing popularity of social media and the plethora of information, consumers will likely use a heuristic
(or peripheral) process to assess influencers on social media. As such, when encountering
an Instagram account, consumers may base their judgments on peripheral cues, such as
number of followers (Chaiken and Maheswaran 1994; Metzger and Flanagin 2013; Petty
and Cacioppo 1986).
In line with previous research that found that people rely on cues as numbers of online
contacts, friends or followers to assess one’s popularity, we expect this to be the case on
Instagram too. In its turn, inferences about popularity may affect evaluations of the source
(Tong et al. 2008; Utz 2010; Graham 2014; Jin and Phua 2014). For instance, when a source
is found to be popular, this may elicit the reasoning that if many others think something
is good or correct, then it must be good or correct, which has been referred to as the
bandwagon heuristic (Sundar 2008). Indeed, people are inclined to believe certain sources
if others do so as well (Metzger, Flanagin, and Medders 2010). Therefore, we expect that
an influencer who is perceived as popular due to its number of followers, is likely to elicit
higher perceptions of opinion leadership compared to an influencer who is perceived as
less popular. In its turn, we expect these perceptions to positively affect the overall likeability of the influencer.
H1: The positive effect of number of followers on overall likeability of the influencer will
sequentially be mediated by perceived popularity and ascribed opinion leadership.
The moderating impact of influencers’ number of followees
Besides number of followers, also the number of followees and the combination of both
may affect one’s perceptions of the influencer. In popular literature, some ‘rules’ about
who to follow and the ideal ‘followers/followees ratio’, mostly concerning Twitter, exist.
For example, a rule of thumb is that you should especially follow people with a positive
ratio, people who have more followers than they follow accounts themselves. On the
other hand, a user who follows many accounts him-/herself has more opportunities to
learn about different topics and opinions, and thus more ability to look beyond their own
social environment, which might be beneficial in terms of opinion leadership (Williams
2006). However, following too much people is not beneficial either, because the likelihood
that one can keep track on all these account’s updates is very small. Moreover, following a
lot of people could be perceived as an attempt to be followed back by those people, thus
increasing your own number of followers (Siegler 2009). On Instagram, hashtags as #followback, #follow4follow, #instafollow and others illustrate this phenomenon. On the contrary, having a lot of followers in combination with only a few accounts following may
indicate that the followers are artificially collected or ‘fake’, which is not beneficial either
(Cresci et al. 2015).
804
M. DE VEIRMAN ET AL.
While there is a quest for the ideal ‘followers/followees ratio’, until today no study has
investigated whether number of followees is an important asset for consumers in the evaluation of an influencer. It is likely that the assumed positive effect of number of followers
on the influencer’s overall likeability (see H1) might turn negative when the influencer
follows few people him-/herself. Hence, we propose the following hypothesis:
H2: For an influencer with a high number of followees, we expect a positive effect of number of followers on influencer likeability.
For an influencer with a low number of followees, we expect a negative effect of number of followers on influencer likeability.
Method
Participants and design
The experiment used a 2 (number of followers: moderate versus high) by 2 (number of followees: low versus high) between-subjects experimental design to test the hypotheses. A
total of 117 Instagram users (74 females, MAge = 29.54 years, SDAge = 6.55) took part in the
study in return for a small payment. We recruited participants in the United States via
Amazon’s mechanical turk.
Manipulation stimuli
Instagram accounts for two fictitious influencers, a male (Stephan Jones) and a female
influencer (Stephanie Jones), using photos of actual influencers were created. The gender
of the respondent was matched to the gender of the Instagram influencers to avoid any
confounds related to gender identification. The influencer’s persona and his/her posted
photos were based on that of various, actual real life influencers’ Instagram pages. Both
influencers had a similar Instagram bio (‘Stephan(ie) Jones j 24 y/o j My life in a nutshell j
Fashion j Travel j Health j Food’). The selected photos were not identical but still similar
for both the male and female influencer and were all lifestyle related. In this way, they
were positioned in such a way that their Instagram posts could appeal to a broad
audience.
As we choose to depict an influencer, his/her number of followers had to be at least of
a moderate size, a number we choose to set at 2.100. In the high number of followers condition, this number was increased to 21.200 (21.2k). For the number of followees, we
choose to distinguish between a low and a high number of followees condition. In
the low number of followees condition, the influencer followed 32 people, whereas in the
high number of followees condition the influencer followed 32.200 (32.2k) people him-/
herself. We manipulated these numbers based on actual influencers’ Instagram pages. To
ensure that participants estimated these numbers equally, they were given an idea of the
average number of followers and followees as a benchmark. Participants were requested
to read the following text: ‘On Instagram, some people called influencers have very large
numbers of followers. Most of these influencers on average have about 2000 followers,
and follow about 300 accounts themselves. For very large numbers, Instagram uses K as
an abbreviation for thousand and M as an abbreviation for million. Please, take a moment
to look at the Instagram profile of Stephan(ie) Jones, an Instagram influencer who gives
people through Instagram a glimpse in his/her life’. Each participant was randomly
INTERNATIONAL JOURNAL OF ADVERTISING
805
assigned to one of the four conditions by being asked to view a screenshot of an influencer’s Instagram page, differing in number of followers and accounts following, and then
fill out a questionnaire. An overview of the manipulation stimuli can be found in Appendices 1 and 2.
Measures
The items for the manipulation checks measured participants’ perceptions of the number
of followers and followees. Participants were asked if they found the influencer had a very
small ( = 1) versus very large ( = 7) number of followers and if they thought the influencer’s number of followers was smaller ( = 1) versus larger ( = 7) than the average influencer’s number of followers. Using the same statements, participants were next asked to
evaluate the influencer’s number of followees. Perceived popularity was measured by a
five-point semantic differential, asking participants if they found the influencer ‘unpopular
versus popular’ (1 item). Ascribed opinion leadership was measured by Flynn, Goldsmith,
and Eastmans’ (1996) 5-point Likert-scale (4 items; 1 = strongly disagree, 5 = strongly
agree; a = .92), adjusted to review others’ opinion leadership. The influencer’s overall likeability was measured using three items of Dimofte, Forehand, and Desphande’s (2003)
scale for attitude toward the spokesperson (5-point semantic differential, a = .85). An overview of the used measurement scales in Study 1 can be found in Appendix 3.
Results and discussion
Manipulation checks
First, both the male (M = 3.41, SD = .55) and female (M = 3.55, SD = .59, t(115) = 1.24, p =
.22) endorser were perceived to be equally credible. Next, participants perceived the influencer’s number of followers to be lower in the moderate (M = 4.93, SD = 1.10) than in the
high number of followers condition (M = 5.86, SD = 1.08, t(115) = ¡4.59, p < .001). Considering the number of followees, participants perceived the influencer’s number of followees to be lower in the low (M = 2.26, SD = 1.75) than those in the high number of
followees condition (M = 6.11, SD = 1.32, t(97.52) = ¡13.25, p < .001). These results show
that the manipulations are satisfactory.
Sequential mediation analysis
To test H1 and H2, we conducted a sequential mediation analysis using Hayes’ PROCESS
macro (2013, model 6, 5.000 bootstrap resamples) with number of followers as independent
variable, perceived popularity and sequentially ascribed opinion leadership as mediators
and influencer likeability as dependent variable. The analysis showed a positive effect of
number of followers on perceived popularity (a1 = .36, SE = .15, p = .017). Next, it was found
that perceived popularity has a significant positive effect on ascribed opinion leadership (a3
= .24, SE = .11, p = .029), which consequently has a significant positive effect on likeability
(b2 = .15, SE = .06, p = .021). Bootstrapping showed a significant indirect effect for perceived
popularity (ab = .18, SE = .07; 95% CI = [.03; .33]), but not for ascribed opinion leadership (ab
= .01, SE = .03; 95% CI = [¡.04; .08]). Important, the serial indirect effect was significant, however small (ab = .01, SE = .01; 95% CI = [.00; .05]). From the above analysis, it appears that the
indirect effect of number of followers on likeability is for the largest part explained by
806
M. DE VEIRMAN ET AL.
Figure 1. Illustration of the effect of number of followers on likeability through perceived popularity
and ascribed opinion leadership.
Note: p < .05; p < .01; p < .001.
perceived popularity only, and for a small part by the sequential mediation through perceived popularity and ascribed opinion leadership (see Figure 1).
Moderation analysis
Addressing H2, a moderation analysis using Hayes’ PROCESS macro (2013, model 1, 5.000
bootstrap resamples) with number of followers as independent variable, number of followees as moderator, likeability as dependent variable and perceived popularity and
ascribed opinion leadership as covariates was conducted. We examined whether
the effect of number of followers on likeability was moderated by number of followees.
The interaction term was statistically significant indicating that the effect of number of
followers on likeability is contingent number of followees, B = .46, SE = .23, p = .05 .We
further examined the conditional effect of number of followers on likeability at the two
numbers of followees. When number of followees was low (32), there was a significant
negative effect of number of followers on likeability, B = ¡.38, SE = .17, p = .03. When
number of followees was high (32.2k), no significant effect of number of followers on likeability was found, B = .08, SE = .16, p = .63. These data suggest that the influencer’s number of followers may have a negative effect on the influencer’s likeability, but this is only
the case when the influencer follows few accounts, not when the influencer follows many
accounts. Both ascribed opinion leadership (B = .14, SE = .06, p = .03) and perceived popularity (B = .44, SE = .07, p < .001) had a significant positive effect on likeability (see Table 1).
Additional analyses on the moderating role of gender
As previous studies have shown that gender differences may appear when individuals
evaluate luxury consumers (e.g. Dunn and Searle 2010; Dunn and Hill, 2014), this may suggest that luxury purchases of significant others may have different effects on men versus
women. Hence, it was investigated whether the above findings differ between men and
women. First, results reveal that the serial indirect effect of number of followers on likeability through perceived popularity and opinion leadership was not moderated by
INTERNATIONAL JOURNAL OF ADVERTISING
807
Table 1. Tested moderation results Study 1.
Coefficient (SE)
t
95% confidence interval
1.75(.33)
¡.16(.16)
¡.38(.17)
.46(.23)
.14(.06)
.44(.07)
5.31
1.00
¡2.24
1.98
2.14
5.96
[1.10; 2.40]
[¡.49; .16]
[¡.72; ¡.04]
[.00; .92]
[.01; .26]
[.29; .59]
Conditional effect of X on Y at values of the moderator
N followees = low
¡.38(.17)
N followees = high
.08(.16)
¡2.24
.48
[¡.72; ¡.04]
[¡.24; .39]
Outcome: likeability
Constant
N followees
N followers
int_1
Opinion leadership
Perceived popularity
Note: int_1 = N followers N followees.
gender (B = ¡.00, SE = .01, 95%CI: = [¡.04; .03]). Next, an analysis using Hayes’ PROCESS
macro (2013, model 3, 5.000 bootstrap resamples) found that the negative effect of number of followers on likeability when the influencer followed few accounts (32) him-/herself
which was found in the second analysis, only applied to females (B = ¡.58, SE = .21, p =
.01). In contrast, this effect was not significant for males (B = .02, SE = .28, p = .93).
Conclusion Study 1
Examining the influence of different numbers of followers on an Instagram influencer’s likeability, Study 1 found that having more followers positively affects attitudes towards the
influencer, for the most through higher perceptions of popularity and for a small part
because these higher perceptions of popularity leads people to ascribe more opinion leadership to the influencer. A high number of followers may thus lead to higher perceptions of
popularity, and subsequently higher likeability, but it does not mean that the influencer is
automatically perceived as an opinion leader. Furthermore, results suggested the emergence of a negative relationship between number of followers and likeability when a popular influencer follows very few accounts him-/herself. However, additional analyses pointed
out that this might only be true for female Instagram users. This might imply that female
Instagram users are more sensitive to influencers’ ‘followers/followees ratio’.
As consumers have become savvy to traditional marketing techniques, brands increasingly partner up with Instagram influencers to reach their target audience. By convincing
influencers to include their products in their posts, brands hope to capitalize on the influencer’s status, credibility and popularity. To further elaborate on the impact of the
influencer’s number of followers on the effects for brands (s)he might promote through
his/her Instagram posts, we conduct a second study. The aim of Study 2 is thus to examine
the impact of the influencer’s number of followers on consumers’ attitude towards the
brand of the promoted product. Moreover, the relation between number of followers and
influencer effectiveness in terms of brand attitude will be investigated for products that
have a standard versus divergent design.
Study 2: assessing the brand effects of Instagram influencers
Hypothesis development
As the number of followers represents the audience with whom influencers share their
ideas, a higher number of followers might elicit stronger brand effects. Jin and Phua
808
M. DE VEIRMAN ET AL.
(2014) recently illustrated this idea and found that positive tweets from celebrities with a
high number of followers result in higher product involvement and buying intentions
than tweets from less popular celebrities. However, we expect that the impact of the number of followers might be different according to the type of product.
Consumers in the first place evaluate products to decide whether they respond to their
needs. In particular, sometimes consumers want to buy what others have bought, but at
other times, they might be very attracted to unique products that are not obvious to obtain
(Steinhart et al. 2014). Two opposite social needs explain these preferences, i.e. the need for
uniqueness and the need for conformity (Tian, Bearden, and Hunter 2001). However, consumers rarely have complete product information, which makes evaluation difficult. Therefore, they often make inferences to fill these gaps. These inferences have been referred to
as naive theories and serve as common-sense explanations to evaluate and interpret mar~ol, Rucker, and Petty 2015; Deval et al.
keting communication, products and brands (Brin
2013; Gunasti and Ross 2009; Kardes, Posavac, and Cronley 2004). Consumers use these
naive theories in forming product judgments and deciding whether a product responds to
their needs. Accordingly, Deval et al. (2013) found that subtle primes in a consumer context
can activate naive theories that guide consumers’ beliefs about market related phenomena
such as pricing, sales promotion, product popularity versus scarcity, and technical language.
Marketers exploit these naive theories by emphasizing product characteristics that are likely
to trigger these naive beliefs associated with desirable consumer responses when developing their communication strategies (Duncan 1990; Lynn 1992). Posavac et al. (2010), for
example, found that firms that are presented as profitable, are evaluated more positively,
their advertisements are considered more credible and their advertised products evoke
more positive brand attitudes, and increase purchase intention. Similar, Steinhart et al.
(2014) found that exposing consumers to functional products (i.e. products that enable one
to achieve a certain goal or complete a practical task) triggers the naive theory of popularity, or the belief that popular products are desirable, similar to ‘bandwagon’ effects (Cialdini
and Goldstein 2004; Deval et al. 2013, Henshel and Johnston 1987). Exposing them to selfexpressive products (i.e. unique products, or products that enable one to diverge from
others) induces beliefs in the naive theory of exclusivity, or the belief that exclusive products are desirable (Berger and Heath 2007, 2008; Snyder and Fromkin 1980).
Following the naive theory of exclusivity, we expect consumers to have a better attitude towards brands with divergent product designs compared to brands with standard
designs because they are perceived as more unique.
H3: Products with a divergent design evoke higher attitudes towards the brand compared
to products with a standard design.
H4: The positive effect of product divergence on attitude towards the brand is mediated
by perceived brand uniqueness.
However, if the product is posted by an influencer with a high number of followers, this
might trigger the naive theory of popularity and thoughts that the product is rather common instead of unique. When such an influencer promotes a divergent product, product
uniqueness might diminish due to the idea that many others might be interested in the
product as well (Hui and Bateson 1991; Machleit, Eroglue, and Mantel 2000). Hence, we
expect the positive relationship between product divergence and attitude towards the
brand through perceived brand uniqueness to be weakened when the product is posted
by an influencer with a very high number of followers. We hypothesize,
INTERNATIONAL JOURNAL OF ADVERTISING
809
H5: The indirect positive effect of divergence of the product design on attitude towards
the brand through perceived brand uniqueness, is weaker when the brand is promoted
by an influencer with a high number of followers compared to when it is promoted by an
influencer with a moderate number of followers.
Method
Participants and design
The study used a 2 (number of followers: moderate versus high) by 2 (product divergence:
low versus high) between-subjects experimental design. Contrary to Study 1 in which
male participants were exposed to a male influencer’s profile and vice-versa, in Study 2
we decided to only include female participants, as the findings in Study 1 mainly applied
to female participants and the majority of Instagram users is female (Statista 2016b).
Hence, any gender confounds are avoided and females are more susceptible to social
influence than males (Eagly 1983). One hundred eighteen female Instagram users from
the United States recruited from Amazon’s mechanical turk completed the study in
exchange for a small payment (MAge = 26.92 years, SDAge = 4.24).
Manipulation stimuli
Participants were exposed to the same profile of Stephanie Jones as in the first study and
were instructed to view her profile carefully. As in Study 1, in the moderate number of
followers condition, the influencer was given 2.100 followers, whereas in the high number
of followers condition, this number was increased to 21.200 (21.2k). The number of followees (N = 320) and number of posts (N = 366), was kept constant over all conditions. Again,
participants were given an idea of the average number of followers and followees of an
influencer as a benchmark. After viewing the profile, participants read that the influencer
recently posted a picture on Instagram and participants were again instructed to view the
post carefully. To strengthen the manipulation, the picture that was posted by the influencer with a moderate number of followers, had a lower number of likes (N = 210) than
the picture that was posted by the influencer with a high number of followers
(N = 2.120), as this controls for discrepancies between the number of followers versus
amount of likes per post ratio. A post of a very popular influencer is likely to be liked by
more people compared to a less popular influencer.
Divergence of the product design was manipulated based on the study of Warren and
Campbell (2014). In the low divergence condition, the influencer promoted a bottle of
water with a standard design of a in the United States unknown brand named NZO,
whereas in the high divergence condition, the influencer promoted a drop-shaped bottle
of water of the same brand. Each participant was randomly assigned to one of the four
conditions. After being exposed to the manipulation stimuli, participants filled out a questionnaire. An overview of the manipulation stimuli can be found in Appendix 4.
Measures
First, as a manipulation check participants were asked whether they found the influencer
had a very small ( = 1) versus very large ( = 7) number of followers. Participants’ perceptions of the divergence of the endorsed product’s design was measured by Warren and
Campbell’s (2014) three-item 5-point Likert scale (a = .88). Participants had to evaluate
810
M. DE VEIRMAN ET AL.
whether the design ‘is different from the norm’, ‘is unique’ and ‘shows independence’. Attitude towards the brand was measured by Spears and Singh’s (2004) five-item, 5-point
semantic differential (a = .93). Perceived brand uniqueness (a = .93) was measured with
four items that measure the uniqueness dimension of Netemeyer et al. (2004) 10-item 5point Likert Brand Equity scale (1 = strongly disagree, 5 = strongly agree). An overview of
the used measurement scales in Study 1 can be found in Appendix 5.
Results and discussion
Manipulation checks
In the average number of followers condition, participants perceived the influencer’s number of followers to be lower (M = 5.56, SD = 1.08) than those participants in the high number of followers condition (M = 6.02, SD = 1.17, t(116) = ¡2.19, p = .03). The divergence of
the endorsed product’s design was also correctly perceived, as the design of the standard-shaped bottle (M = 2.91, SD = .97) was evaluated significantly lower in divergence
than the design of the drop-shaped bottle (M = 4.25, SD = .60, t(116) = ¡8.94, p < .001).
The manipulations were effective.
Post-test on the manipulation of the bottle design
To ensure that perceived attractiveness of the standard-shaped bottled water and that of
the drop-shaped bottled water are similar and only differed in terms of perceived divergence, a post-test (N = 49; 26 females; MAge = 34.53 years, SDAge = 11.49) was set up. As
such, next to perceived divergence (Warren and Campbell 2014; three items, 5-point Likert
scale; a = .91), in a between-subjects design, participants’ attitude towards the design of
the bottle (Warren and Campbell 2014; two items; a = .84, 5-point Likert scale), aesthetic
evaluation of the bottle (Bell, Holbrook, and Solomon 1991; four items; 5-point semantic
differential; a = .94), attitude towards the product/brand (Aaker, Brumbaugh, and Grier
2000; three items; 5-point semantic differential, a = .94), perceived quality of the brand
(Taylor and Bearden 2002; four items, 5-point semantic differential; a = .89), evaluation of
product packaging (Ghoshal, Boatwright, and Cagan 2011; one item, 5-point semantic differential), perceived product attractiveness (Page and Herr 2002; two items, 5-point
semantic differential; a = .89) and perceived expressive aesthetics of the bottle (Cai and
Xu 2011; five items, adapted to the packaging context, 5-point Likert scale; a = .95) were
measured. An overview of the used measurement scales can be found in Appendix 6. It
was found that the two bottle designs were only evaluated as significantly different in
terms of perceived divergence (see Table 2).
Simple mediation analysis
To test H3 and H4, we conducted a simple mediation analysis using Hayes’ PROCESS
macro (2013, model 4, 5.000 bootstrap resamples) with the divergence of the product
design condition as independent variable, perceived brand uniqueness as mediator and
attitude towards the brand as dependent variable. The analysis revealed a significant indirect effect of divergence of the product design on attitude towards the brand through
perceived brand uniqueness (ab = .57, SE = .11, 95% CI: = [.37; .81]). Exposure to the dropshaped bottle led to higher perceived brand uniqueness (a = 1.02, SE = .17, p < .001),
INTERNATIONAL JOURNAL OF ADVERTISING
811
Table 2. Results of the post-test on the manipulation of the bottle design.
Perceived divergence
Attitude towards the bottle design
Aesthetic evaluation
Attitude towards the product/brand
Perceived quality of the brand
Evaluation of product packaging
Perceived product attractiveness
Perceived expressive aesthetics
Standard design M (SD)
3.11 (1.11)
3.77 (.91)
3.72 (.91)
3.76 (.92)
3.62 (.68)
3.73 (.96)
3.51 (1.15)
3.53 (.99)
Divergent design M (SD)
4.04 (.97)
3.33 (1.14)
3.91 (1.05)
3.72 (1.07)
3.88 (.91)
4.00 (1.08)
.78 (1.02)
3.62 (1.19)
t (df)
¡3.11 (46.99)
1.49 (41.90)
¡.68 (43.97)
.11 (43.72)
¡1.10 (40.38)
¡.91 (44.31)
¡.78 (45.70)
¡.25 (43.03)
p
.003
.144
.502
.912
.277
.366
.440
.803
which, in turn, positively affected attitude towards the brand (b = .56, SE = .06, p < .001).
These results confirm H3 and H4 (see Figure 2).
Conditional process analysis
To test H5, we conducted a moderated mediation analysis using Hayes’ PROCESS macro
(2013, model 7, 5.000 bootstraps; 95% bias-corrected confidence intervals) with number
of followers (i.e. moderate versus high) as the moderator of the effect of divergence of the
product design on attitude towards the brand through perceived uniqueness of the
brand. The moderated mediation index was significant (ab = -.50, SE = .20, 95% CI: =
[¡.93; ¡.15]). Based on this result, we can infer that the indirect effect of divergence of
the product design on attitude towards the brand through perceived brand uniqueness
differs significantly across different levels of followers. The interaction term was significantly indicating that the effect of divergence of the product design on perceived brand
uniqueness is contingent on the number of followers of the influencer (B = ¡.91, SE = .33,
p = .01).
We further examined the conditional indirect effects of divergence of the product
design on perceived brand uniqueness for the two numbers of followers (Hayes 2013).
Moderated mediation analysis revealed that when the number of followers was moderate,
there was a significant positive effect of divergence of the product design on perceived
brand uniqueness (ab = .80, SE = .15, 95% CI: = [.53; 1.12]). When the number of followers
was high, the positive effect of divergence of the product design on perceived brand
uniqueness was weaker (ab = .29, SE = .15, 95% CI: = [.01; .58]). Therefore, divergence of
the product design has a positive effect on perceived brand uniqueness, and that effect is
Figure 2. Tested simple mediation model.
Note: p < .05; p < .01; p < .001.
812
M. DE VEIRMAN ET AL.
Table 3. Tested moderated mediation model results Study 2.
Outcome: perceived brand uniqueness
Constant
Divergence
N Followers
int_1
Outcome: attitude towards the brand
Constant
Perceived brand Uniqueness
Divergence
Direct effect of X on Y
Coefficient (SE)
t
95% confidence interval
2.69(.15)
1.43(.22)
.52(.24)
¡.91(.33)
17.30
6.41
2.12
¡2.75
[2.38; 3.00]
[.99; 1.87]
[.05; .99]
[¡1.56; ¡.25]
1.87(.20)
.56(.06)
¡.29(.13)
9.12
8.63
¡2.16
[1.47; .68]
[.43; .68]
[¡.55; ¡.02]
¡.29(.13)
¡.16
[¡.55; ¡.02]
Conditional indirect effects of X on Y at values of the moderator
N followers = moderate
.80(.15)
N followers = high
.29(.15)
[.53; 1.12]
[.01; .58]
Note: Index of moderated mediation: ab = ¡.50, SE = .20, 95% CI: = [¡.93; ¡.15]); int_1 = N followers N followees.
stronger when the influencer’s number of followers is moderate compared to high. Our
data suggest that exposure to a product with a divergent design leads to higher perceived
brand uniqueness, which, in turn, increases attitude towards the brand. However, this process is conditional on number of followers of the influencer: if the product is endorsed by
an influencer with a moderate number of followers, this effect is stronger than if the
product is endorsed by an influencer with a high number of followers, confirming H5
(see Table 3 and Figure 3).
General discussion
Summary of major findings and implications
This study adds value in shedding light on a specific type of endorser that has been gaining traction in marketing lately, namely social media influencers. As social media continue
to gain in popularity and concerns about ad-blocking grow, influencer marketing has
Figure 3. Tested moderated mediation model: effect of divergence on Aab via perceived brand
uniqueness, moderated by number of followers.
Note: p < .05; p < .01; p < .001.
INTERNATIONAL JOURNAL OF ADVERTISING
813
become increasingly important to brands. Rather than pushing their (whether or not personalized) ads to their target audience, brands are turning to trusted online personas to
get their products and messages out to the consumer. However, despite its growing use,
there has been little experimental research on the phenomenon of influencer marketing.
One of the biggest challenges in influencer marketing is the identification of the right
influencers. As a metric for potential reach, today number of followers is often used as a
starting point in the search for influencers. This study’s findings contribute to the quest
for influencer selection by indicating that an influencer’s number of followers affects consumers’ attitudes towards him/her (i.e. likeability) and that this is mainly explained by perceptions of popularity, in line with the findings of Quercia et al. (2011) and Romero et al.
(2011). Only a small serial effect was found, indicating that these perceptions of popularity
caused consumers to ascribe opinion leadership to the influencer in question. In other
words, a high number of followers may not always translate into true influence. Accordingly, this study theoretically contributes to the on-going debate concerning opinion leadership and how to identify it and word-of-mouth diffusion literature.
Moreover this study is the first to include the importance of one’s ratio of followers versus followees in the assessment of an influencer. It was found that a high number of followers may negatively impact influencer likeability for influencers who are following few
accounts themselves. As such, this finding confirms the negative implications of the inpopular-literature-described ‘hugely positive ratio’s’ (Siegler 2009). The current study
could not provide evidence for the underlying mechanism of this effect. Future research is
needed to further disentangle this effect. We expect that a low number of followees may
have a detrimental impact on the trustworthiness and credibility of the endorser. A high
number of followers combined with a low number of followees can be an indicator of a
false account created for advertising purposes or might elicit perceptions that the influencer is mainly aiming at commercial collaborations, thus being less authentic, whereas
authenticity should be precisely the strength of collaborating with influencers.
Next, this study theoretically contributes to literature on heuristic processing and naive
theories. In particular, Study 2 found that consumers’ attitudes towards a new, unknown
brand which has not yet established brand knowledge in the consumers’ mind may be
influenced by a product’s design and the perceptions that it evokes. Specifically, it was
found that a product with a divergent design causes perceptions of uniqueness which
eventually positively affect consumers’ attitudes towards the brand, in line with the naive
theory of exclusivity (Berger and Heath 2007, 2008). These findings are not new, however
and more important, it was found that the influencer’s number of followers may change
perceptions in the sense that when a product with a divergent design is endorsed by an
influencer with a high number of followers, perceptions of uniqueness and eventually attitudes towards the brand are lower compared to when it is endorsed by an influencer with
a moderate number of followers. A high number of followers triggered the idea that the
product is not that unique after all, as many others are interested in it (Machleit, Eroglue,
and Mantel 2000). These findings are consistent with Hellofs and Jacobson’s (1999) findings that if the market share of exclusive products grows, this may infer a loss of exclusivity for consumers. If a product with a divergent design which people purchase to stand
out from the crowd, appears in a great number of Instagram feeds, the brand’s perceived
uniqueness will be lowered, leading to lower brand attitudes. Thus, important, when
searching for an appropriate influencer, marketers must also consider the type of product
814
M. DE VEIRMAN ET AL.
they want to promote. Although it is tempting to choose an influencer with a high number of followers in any case, this would not be the best marketing option for each product
type. These insights are very important as they indicate once again that a high number of
followers is not always a guarantee of success. Therefore, the topics influencers post and
the audience they reach in terms of interests and activities, rather than the size of their
audience might be more important to take into account.
Limitations and future research
Obviously, brands should look beyond just the number of followers and followees as they
determine their ideal pool of influencers. Especially because we found that influencers
with high numbers of followers do not necessarily evoke perceptions of opinion leadership. One should evaluate the topics they are posting on, the quality of the content they
post, their fan engagement, photography style, etc. Future research could delve into these
specific assets and how they influence the preference for a specific influencer, and more
important the influencer’s true influence on consumer decisions. Additionally, because
meaning may transfer from people to associated brands (McCracken 1986), different influencers could evoke different effects. Future research could explore whether certain types
of influencers are more likely to influence brand perceptions in a beneficial way than
others. Moreover, previous research that examined transfer effects from the endorser
onto the perception of products and brands, mainly in the domain of celebrity endorsement, has found that certain attributes or characteristics of endorsers may enhance advertising effectiveness (Bergkvist, Hjalmarson, and Magi 2016). Consumers’ perceptions of
endorsers affect the effectiveness of the message and the consumer–brand relationship
(Clow et al. 2006; Dwivedi, Johnson, and McDonald 2016). In particular, accompanied by a
relevant fit with the endorsed products (Kirmani and Shiv 1998; Misra and Beatty 1990),
personal attributes of the endorser may enhance his or her persuasiveness. As such,
endorsers should be perceived as credible, attractive and they should be well liked in
order to have positive effects on brand evaluations (see Bergkvist and Zhou 2016 for a
review). Source credibility is driven by perceived expertise and trustworthiness of the
communicator and influences consumers’ attitudes (see Pornpitakpan 2004 for a review).
Attractiveness refers to the endorser’s physical appeal and may positively affect brand
evaluations (e.g. Eisend and Langner 2010; Lord and Putrevu 2009; Till and Busler 2000),
however its importance might depend on the advertised product (e.g. Kamins 1990).
Moreover attractiveness goes beyond physical attractiveness, other aspects such as perceived familiarity, similarity and likeability may impact the endorser’s persuasiveness
(McGuire 1985; Ohanian 1991). These factors have been kept constant in both studies
reported; however, they will possibly moderate the observed effects and should be
included in future research.
Another factor that may impact the effectiveness of influencers’ endorsements is the fit
between the influencer and the brand, or the similarity or consistency between the brand
and the influencer, which has been referred to as the match-up hypothesis (e.g. Kamins
1990). Also, recently Bergkvist et al. (2016) found that an important factor in celebrity
endorsements is consumers’ attribution of motives for the celebrity’s endorsement. When
an endorser was found to be merely motivated by money, this rendered less positive
brand evaluations than when (s)he was seen as being motivated not only by money but
INTERNATIONAL JOURNAL OF ADVERTISING
815
also by product quality. In this sense, future research may investigate whether different
social media endorsements may evoke different perceptions about the influencers’
motives among consumers and how these perceptions may affect endorsement
effectiveness.
We aimed to maximize internal validity and the effects of the manipulation stimuli by
creating fictitious Instagram profiles that resembled real influencer profiles as much as
possible. Also we manipulated the influencer’s number of followers and followees based
on actual influencers’ Instagram pages. Although we gave participants an idea of a typical
influencer’s average number of followers and followees as a benchmark, it is still possible
that certain participants might have differently evaluated these numbers. Participants
might have for example used the number of followers and followees of an influencer they
know as a benchmark, or even their own number of followers and followees. Therefore, to
improve external validity, future research could use actual Instagram profiles or even set
up a collaboration with influencers and question their followers.
In Study 2 in which we focus on influencers’ endorsements, to avoid confounds and
create realistic manipulations, we deliberately adapted the Instagram post’s number of
likes according to the influencer’s number of followers. This is because we expected that a
low number of likes may negatively affect the credibility of the account with a high number of followers. This is because an influencer with a high number of followers can be
expected to have higher number of likes on his/her posts than an influencer with a low
number of followers. A low number of likes can be an indication of a low quality message
for the account with a high number of followers, while it would be normal for the account
with a low number of followers and that would also create a confound in the results.
Future research may delve into the importance of number of likes and the relationship
between one’s followers and the number of likes that (s)he is able to generate. In this
sense, it might be interesting to investigate the effects of an incongruence between number of followers and number of likes, both in terms of attitudes towards the influencer as
in terms of attitudes towards the endorsed brand or product.
The second manipulation in Study 2 was product divergence, however, be it for a lowinvolvement product, namely bottled water. Findings could differ for high-involvement
products. Furthermore, a bottle of water seems to be no product consumers use to signal
their identity and thus purposely use in public. As people use products to signal their identity, findings might be different for typical public products. Therefore, future research
should examine a wider variety of product categories and look for further differences
between product types to generalize our results. Moreover, to avoid influences of consumers’ established brand knowledge, the brand stimuli in Study 2 were unknown to the
participants. However, consumers’ evaluations may vary according to consumers’ prior
experiences or familiarity with the brand (e.g. Hong and Sternthal 2010). Therefore, future
research should investigate whether the results also hold true for known brands that have
established a strong brand image in consumers’ minds and whether results differ depending on these brand images. For instance, divergent designs may only work for brands that
are known for their creativity compared to more traditional brands. Also, consumers may
have different preferences for divergent products. Future research should take into
account these personal preferences and include certain trait variables such as “Centrality
of Visual Product Aesthetics” (Bloch, Brunel, and Arnold 2003) that may affect the preference for divergent product designs.
816
M. DE VEIRMAN ET AL.
In Study 1, both males and females participated in the study. However, Study 1 indicated that the found negative effect of number of followers on likeability when the influencer followed very few accounts, only applied to females. Looking for an explanation, we
found that this might be the case because females are more susceptible to social influence than males (Eagly 1983). Therefore, we choose to only allow female participants in
Study 2, next to the fact that mainly females use Instagram. However, further research
should delve deeper into the specific differences between men and women. Not only
females are more susceptible to social influence, this is also the case for adolescents as
they are very concerned about others’ evaluations and feel under constant scrutiny by an
imaginary audience (Steinberg and Monahan 2007). Therefore, in future research it might
be interesting to focus on young Instagram users. Moreover, Instagram has a large population of young users and is considered the most important social network by American
teenagers than any other network (Meeker 2015).
Another promising research direction is to explore whether the impact of influencer
marketing differs depending on certain personality traits of the observer (e.g. self-esteem,
need for uniqueness,…). Because people give meaning to the (commercial) messages
they receive, different people may have different preferences for different types of influencers and posts including brands, according to their characteristics (Mick and Buhl 1992;
Stern 1991).
To conclude, number of followers is an interesting metric in the quest for influencers,
however it is not the Holy Grail of influencer marketing. Equally important is the influencer’s ‘followers/followees ratio’, as influencers with a high number of followers but a
very low numbers of followees might be found less likeable. Moreover, the type of product willing to promote through an influencer should be taken into account. When considering an influencer marketing strategy to promote divergent products, partnering with an
influencer with a high number of followers might not be the best option, as this may
lower the brand’s perceived uniqueness and consequently brand attitudes.
Acknowledgments
The authors would like to thank Lisa Van den Abbeele for her help with the creation of the manipulation stimuli.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on contributors
Marijke De Veirman is a teaching assistant and PhD candidate in the Department of Communication
Sciences, Faculty of Political and Social Science, Ghent University. Her research interests include
social media marketing and consumer behavior.
Veroline Cauberghe is an assistant professor in the Department of Communication Sciences, Faculty
of Political and Social Science, Ghent University. Her research interest lays on advertising effectiveness and social marketing.
INTERNATIONAL JOURNAL OF ADVERTISING
817
Liselot Hudders is an assistant professor in the Department of Communication Sciences, Faculty of
Political and Social Science, Ghent University. Her research interests include the relation between
consumption and happiness and advertising effectiveness.
References
Aaker, J., A.M. Brumbaugh, and S.A. Grier. 2000. Nontarget markets and viewer distinctiveness: The Impact
of target marketing on advertising attitudes. Journal of Consumer Psychology 9, no. 3: 127–40.
Abidin, C. 2015. Communicative
Purchase answer to see full
attachment