Celebrity Endorsement
An empirical study that examines whether
celebrity endorsed advertising is most
effective, in terms of purchase intentions,
via Snapchat, Twitter or Instagram, based
on two different characterizations; duration
and layout.
BSc Thesis BA by
Else van Oord
10783814
25-06-2018
2017/2018
Supervisor: Hesam Fasaei
Reputation, CSR, and celebrity and their effects on
innovation and other strategic decisions in
organizations
Statement of Originality
This document is written by Else van Oord who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business of the University of Amsterdam is responsible solely for the supervision of completion of the work, not for the contents.
Content
1. Introduction 4
2. Literature review 6
2.1 Celebrity endorsement 7
2.2 Purchase intention 10
2.3 Visual content versus textual content 11
2.4 Permanent visibility versus temporary visibility 13
2.5 Layout x Duration 15 3. Methodology 17 3.1 Research design 17 3.2 Participants 17 3.3 Stimulus material 18 3.4 Procedure 18 3.5 Measures 19 3.6 Control variables 20 4. Results 20 4.1 Preliminary analysis 20 4.2 Hypothesis testing 21
4.3 Checking for control variable 23
5. Conclusion 24
5.1 Discussion and link to former research 24
5.2 Limitations and future research 27
5.3 Theoretical relevance 28
6. References 29
1. Introduction
Nowadays, the success, and therefore the future, of a company depends largely on how they
are viewed by their stakeholders. To reach and influence them, organizations can use different ways
of communication. With the arrival of the internet and social media, the way brands communicate
with their customers changed. Today, more than 2 billion people are actively using social media (eg.
Instagram has over 500 million daily active users (Statista, 2017), Snapchat over 191 million and
Twitter over 336 million (Statista, 2018)). Due to the significant growth in the use of internet and
social media among consumers, these channels have become very important for marketers and are
relevant and interesting to study (Swedowsky, 2009).
It is difficult for marketers and advertisers to reach and convince customers because
they are constantly exposed to advertising. This sometimes leads to skepticism under customers
(Helm, Moulard & Richins, 2015). Therefore marketers have to think of creative ways to gain
attention from customers. One way to do this for an organization is to endorse with a celebrity.
Celebrities are characterized as individuals with a wide range of personal connections, who play an
important influential role and are considered as a source and advisor at the same time. According to
Roberts (2016), celebrities can be seen as opinion leaders. He defines them as “influential
individuals who translate mass media content into persuasive personal messages that impact the
people in their social circles” (p.170). As an independent third party, celebrities, have the power to
make the link between companies and the specific public they reach (Kip & Uzunoğlu, 2014).
Because people choose who they want to follow on social media, more interest will be shown in
these posts. In this study, we look into social media posts of celebrities. By tapping into these
celebrities’ posts marketers and advertisers can reach and attract the attention of these followers.
When a company decides to endorse their products/services with a celebrity, it is very
important which person they choose for this. Organizations can use celebrities to transfer specific
cooperation between Tommy Hilfiger and Gigi Hadid. Since two years Gigi is the endorser of the
brand. According to Tommy Hilfiger, they want to associate the brand with Gigi’s “positive,
down-to-earth-energy and cool, nonchalant style.” They say that “with her 33,9 million Instagram
followers Gigi is a force in the fashion industry and the ultimate Tommy Girl” (Gigi Hadid verlengt
samenwerking met Tommy Hilfiger, 2017). When customers can identify themselves with the
endorsed celebrity they will be more loyal and dedicated. So the organizations should pick someone
who has the norms and values they want to convey. If an organization manages to form this loyal
customer base, they will see this reflected in the behavior of consumers. They will buy products
more easily and pay less interest in competitors, which will have a positive effect on the overall
reputation and brand attitude (Lam, Ahearne, Mullins & Schillewaert, 2013; Stokburger Sauer,
Ratneshwar & Sen, 2012).
According to Steffes and Burgee (2009), the way organizations communicate and through
which channel determines how consumers perceive the information and reflect on the
organizational identity. The credibility of the source depends on the characteristics of the different
communication channels. In this study, we look into three different communication channels;
Snapchat, Instagram, and Twitter. Snapchat is a mobile application for sharing photos and videos
and is characterized as ephemeral social media, a platform that displays shared content for a limited
period of time (Bayer et al., 2016). Instagram is a mobile photo/video capturing and sharing service
that focuses on visual storytelling. Twitter is a microblogging platform. Users can post short public
messages/updates. They can decide who they want to receive messages from, not necessarily who
can receive their messages. Above three social media platforms are often used to launch celebrity
endorsed advertisements. To date, little research has been done on via which platform this leads to
the highest purchase intentions. These platforms were chosen, because of their popularity. It is
important to look at how celebrity endorsement via these different types of platforms affect
(Djafarova & Rushworth, 2017). This success is measured by purchase intention, as it is the most
common way to measure advertising effectiveness (Beerli & Santana, 1999). Thus, it is important to
distinguish between different types of platforms, as the effect they can have on consumers may
differ. In this study I will look into whether celebrity endorsed advertising is most effective, in
terms of purchase intentions, via Snapchat, Twitter or Instagram, based on two different
characterizations; duration and layout.
RQ: Does celebrity endorsement leads to the highest purchase intentions via Snapchat, Instagram or Twitter?
After reviewing the literature, we describe our hypotheses and methodology in the next
session. We then discuss our data and present the empirical results. The conclusion follows.
2. Literature review
As previously described in the introduction, this study focusses on the effect of celebrity
endorsement via different social media platforms (Snapchat, Twitter, and Instagram) on the
dependent variable; purchase intention. In this section, we look at existing literature regarding
celebrity endorsement, purchase intention and into two characteristics that distinguish the platforms;
layout and duration. In this study, duration means for how long the advertisement is visible, this
could either be permanent or temporary. The layout is divided into two options, visual content
versus textual content. The outcome variable is purchase intention. This variable is chosen, as
previous studies have proven this to be a valid measure in determining the effectiveness of
advertisements (Dehghani, Niaki, Ramezani & Sali, 2016; Djafarova & Rushworth, 2017). The
following section will focus on existing literature and theories relevant to determining how celebrity
endorsement via Snapchat, Instagram and Twitter may differ in the effects they have on the outcome
2.1 Celebrity Endorsement
A celebrity endorser is defined as “any individual who enjoys public recognition and who
uses this recognition on behalf of a consumer good by appearing with it in an advertisement”
(McCracken, 1989, p.310). According to Gwinner (1997), celebrities can be seen as associative
networks. When a celebrity is used to endorse a brand, those associative networks become linked to
each other (Gwinner, 1997). Centeno & Wang (2016) have investigated how this network is created.
They state in their research that celebrities can function as human brands. Every human brand has
its own identity. According to them, a brand identity is “a unique set of brand associations that
brand strategists aspire to create or maintain” (Centeno & Wang, 2017, p.133). Human brand
identities form through a collaborative of different stakeholder actors (Holt, 2004). These different
stakeholders actors are the Focal Stakeholder-Actor, Primary Stakeholder-Actor, and Instrumental
Stakeholder-Actors. The Focal Stakeholder-Actor is the celebrity. His or her everyday life choices
and values can reach millions of people. With the arrival of social media, celebrities can influence
this process through their social media platforms. According to Centeno & Wang (2016, p.133)
creates social media “an outlet for celebrity stakeholders to create, re-create, persuade, and
negotiate identities for social and economic purposes”. This effect is reinforced by the fact that
participation within a community or with like-minded customers, their followers, is an important
factor contributing to the success of a brand (Milde & Yawson, 2017, 20). The followers, fans, and
consumers are the Primary Stakeholder-Actor. They co-create by interaction and socialization with
the celebrity. Because of these parasocial interactions they can feel a greater sense of social
intimacy and relatedness. This increases the likelihood that they will become active members of the
brand community (Centeno & Wang, 2017, p.135). Instrumental Stakeholder Actors are television
networks, the press, talent agencies etc. They lend information, are a source and co-create by
mediating between celebrities and stakeholders (Centeno & Wang, 2017, p. 135). This co-creation
including people, organizations, agents, machines and other objects transform and translate into a
common project (Law, 1991).
There has been a lot of research into the effectiveness of celebrity endorsement. One reason
for its effectiveness is explained by the Persuasion Knowledge Model by Friestad and Wright
(1994). This model shows that when people recognize a message as an advertisement, their
persuasion knowledge will be activated. Customers can become skeptical and sometimes show
resistance towards the product/service that occurs (Friestad & Wright, 1994). In addition, research
of Just & Carpenter (1992) shows that because of consumers’ selective attention, they are unable
and unwilling to pay attention to all the ads they encounter. Celebrity endorsement is proved to be
an effective strategy to attract customers (Belch & Belch, 2004). Advertisements containing
celebrity endorsement showed to have a better chance of breaking through the advertising clutter
(Choi & Rifon, 2007). Other research has shown that consumers believe that celebrities genuinely
like and use the actual endorsed products, without receiving money for it (Silvera & Austad, 2004;
Dekker & Van Reijmersdal, 2010). This is in contrast to advertisements that are posted by a brand,
by which consumers might associate exposure to a brand with commerce (Bhatnagar, Aksoy, &
Malkoc, 2004).
Kelvin (1961) explains the effectiveness of celebrity endorsement with the Internalization
Process. He states that the Internalization Process of social influence occurs "when an individual
accepts influence because the induced behavior is congruent with his value system" (Kelman, 1961,
p. 65). Important to the acceptance of influence is the credibility of the source of influence. If a
source is perceived to be credible, a consumer will accept the influence (information) as accurate
and use it. It is accepted because it is ‘demanded’ by one's own values (Kelman, 1961). This
argument is confirmed by Petty & Cacioppo (1986) with the Source Credibility Theory. This theory
states that communication's persuasiveness is affected by the perceived credibility of the source of
distinct dimensions, namely expertise, trustworthiness, and attractiveness. So when the audience
sees the celebrity more as an expert, trusts the celebrity and find the celebrity attractive, the
celebrity will be more valued as a credible source and of better persuasive influence than when the
product is only presented by the brand itself. A person’s family and friends are perceived as
credible. Celebrities do not have the same ties to customers as family and friends. However,
previously discussed literature provides evidence that celebrities can be perceived as trustworthy
and credible (Petty, & Cacioppo, 1986). Friedman and Friedman (1979) found in their research that
products high in psychological and/or social risk endorsed by a celebrity were evaluated more
favorably than those using an expert or a typical consumer; and that the celebrity was more
effective in getting consumers to buy the product. Thus, an endorsement of a product by a credible
source may influence purchase behavior.
Another reason for celebrity endorsements effectiveness is shown in the Meaning Transfer
Model (McCracken, 1989), which states that through associations, certain qualities of celebrities
can be transferred to the brand. When people like the endorsed celebrity, this can lead to a greater
liking of the brand. Hsu & McDonald (2002) state that the fit between the endorser and the brand is
a key factor in the success of celebrity endorsement. An organization should choose a celebrity that
has the associations they want to associate with the brand. This is in line with the Match-Up
Hypothesis (Kamins, 1990), which suggests that celebrity endorsement is only effective if the
endorser has a congruent fit with the endorsed product or brand. Friedman & Friedman (1979) use
the Identification Process to explain this. This theory suggests that a person is more likely to adopt
an attitude or behavior of another person or a group if he/she can identify with the person (Kelman,
1961). Research shows that when there is a match between the celebrity and the brand, this leads to
2.2 Purchase Intention
The reputation of a company is very important when it comes to the financial benefits a
company is seeking to obtain. To succeed, it is necessary for companies to focus on how they are
perceived in consumers’ minds. A customers’ attitude towards an organization or its products/
services is reflected in their actual behavior. Fishbein and Ajzen (1975) define actual behavior as;
"the intention to behave in a certain way”.As stated in the introduction, a company can influence
how they are perceived in the consumer’s minds. An example of this is described by Green &
Peloza (2011). They state that companies that are active in socially responsible activities motivate
customers to buy a product because they have a rewarding feeling of being socially responsible.
This same motivation leads to an increase in the brand loyalty of the customer and a decrease in
price sensitivity. A decrease in price sensitivity means that customers are willing to pay more for a
product/service or are less sensitive towards price changes.
Another way for organizations to influence the way they are perceived in consumers’ minds
is advertising. As a result of advertising, consumers can get to know a brand and be able to identify
it. Hutter et al. (2013) define this as brand awareness; “a consumer’s ability to identify a brand and
the strength of the brand’s presence in a consumer’s mind”. Brand awareness tends to be important
because, research shows, people buy products from a brand that they have heard of much more
easily than from an unknown brand. They have more confidence and trust in a brand they know of.
For this reason, brand awareness can be seen as a step towards developing purchase intentions
(Hutter et al., 2013; Lu, Chang, W.P. & Chang, H. H., 2014). This is confirmed in research by
Djafarova & Rushworth (2017) who found that brand awareness is a key predictor of an
individual’s purchase intentions. This is quite logical as people tend to purchase products from a
brand they have positive associations with. Purchase intentions are defined as “the likelihood a
consumer would consider purchasing a certain product” (Lafferty, Goldsmith & Newell, 2002).
factors that influence consumers’ purchase intentions are a brand name, product quality, price,
product packaging and advertising have a significant impact on it (Mirabi et al.,2015).
2.3 Visual content versus textual content
As said before, the characteristics of a platform determine how consumers perceive
information and react to celebrity endorsement. Twitter is a social networking service that allows
users to publish short messages of up to 280 characters. A Twitter message is called a ‘tweet’. In
contrast to this, Instagram and Snapchat are based on visual storytelling. Instagram and Snapchat
are photo and video-sharing social networking services, which can be edited with various filters,
and organized with tags and location information. A lot of research has been done on the effect of
visual versus textual advertisements. Most of these research arguments that visual advertisements
are more effective.
An argument for the fact that visual advertisements are more effective is explained by the
Motion Effects Theory (Sundar & Kalyanaraman, 2004). The Motion Effect Theory states that
people have an innate preference for moving objects over static objects. The reason for this is that
movement seizes our attention and creates excitement. Other research confirmed that animated ads
caused higher levels of excitement and better memory of the content (Diao & Sundar, 2004; Day,
Shyi & Wang, 2006). A reason why visuals are easier to memorize than words is their capacity to
arouse mental images.
The second argument for the effectiveness of visual advertisements is shown in the Dual
Coding Model (Singh et al., 2002). This model shows that pictures are encoded in imaginal codes
and texts in verbal codes, so pictures are labeled more spontaneously as the creation of the code is
more likely for pictures than for words. Thus, pictures will have a greater number of codes that will
act as multiple retrieval routes in the memory, which means that pictures will be memorized faster
and add that pictures can transmit specific beliefs. According to the authors, pictures are ambiguous
and open, characteristics that make them more persuasive as they can lead to different
interpretations, so people can create their own story about the picture and give them different
meanings.
Joffe (2008) argues that the difference between the effects of texts and visuals is because of
the emotive impact. According to him, pictures are more emotive while texts are more rational,
logical and linear. In his study, he shows that visual elements inculcate stronger personal concerns
and evoke stronger emotional engagement than texts as they represent life situations and are
subjectively relevant. The emotions transmitted by the pictures make them more vivid and they
leave a stronger memory trace than text due to their salience effect. When attention is directed to a
part of the environment, it tends to be remembered when making judgments.
The last argument of why visual ads are more effective is according to Powell et al., (2015)
that visuals are highly salient and are perceived quickly. This is because they are easier to
understand than text, and therefore, the frame provided by the image may be stronger than the frame
provided by the text. Thus, images are expected to have stronger effects than text and to have a
stronger effect on behavioral intentions than text. Therefore, visuals have a strong effect on
advertising and persuasion, and companies should use them to persuade. Moreover, visuals can help
companies gain endorsement as pictures are more persuasive than text.
Based on the above arguments, it seems that celebrity endorsement should be more effective
via platforms characterized by visual content. That is why I propose the following hypothesis:
H1: Celebrity endorsement leads to higher purchase intentions via social media platforms characterized by visuality than via platforms characterized by textuality.
2.4. Permanent visibility versus temporary visibility
Another characteristic that distinguishes Snapchat, Instagram and Twitter is the duration of
the shared content. Snapchat allows users to create multimedia messages, also called ‘snaps’. These
snaps can either be a photo or a short video. Snapchat users can create multimedia messages, also
known as ‘snaps’; these can either be a photo or a short video. It is possible to edit ‘snaps’ with
several filters and effects, drawings and text captions. Because Snapchat is a relatively new
platform, the effectiveness of celebrity endorsement via this platform has not been researched
empirically in previous studies. The main characteristic that divides Snapchat from Twitter and
Instagram is that the shared content is only visible for a certain amount of time, after which it
becomes inaccessible. Receivers can screenshot the snaps they are receiving, but the sender gets a
notification of this. This shapes the uses and perceptions of the medium (Kaun & Stiernstedt, 2014).
Research shows that photography and audio recording is normally done to capture a specific
moment and make them accessible over time and across space (Treem & Leonardi, 2012). Zhao &
Lindley (2014) found in their research that social media is partly valued for their archiving capacity.
If the shared content on social media is permanent, this allows people to document, organize and
remember specific moments from the past (Ozkul & Humphreys, 2015). For example, Instagram is
seen as a place for archiving memorable events or sharing important things, this does not apply to
Snapchat. A study by Bayer et al. (2016) shows for what reasons people do use Snapchat. Users
state that it is more for fun; sharing the small moments with your friends. A participant in their
research used Snapchat to communicate with a childhood friend who lived in another state. For
them, Snapchat provided a window into each other’s daily lives. People sent stuff to each other like
‘toilet selfies’, that they would never post on for example Instagram.
Research shows that Snapchat is mostly used to interact and check-in with close ties
throughout the day through the sharing of mundane content (Bayer et al., 2016). Celebrities do not
much more difficult than on other platforms. Unlike other platforms, you can not just type in the
name of the celebrity you are looking for. You need to know the exact username of the celebrity to
add him or her. In addition, the specific audience on other platforms is often unknown (Bernstein et
al., 2013), Snapchat gives the sender explicit control over the audience by allowing the user to
choose who the snap is sent to and notifying the user when the snap is viewed. For this reason, users
are less likely to follow someone, because they might feel uncomfortable not being totally
anonymous.
Another consequence of the temporal character of Snapchat is that it is not visible to the
future audience, unlike for example an Instagram post. Because of its non-archival content, it does
not have a like, react or favorite feature. If you receive a Snapchat you can not see other people’s
reaction or emotion towards it. Consumers are sensitive to the opinion of other people. If a product/
service has a lot of bad reviews, it is less likely to be bought, as opposed to when people are
enthusiastic about something, what regularly leads to purchase intentions. The only record of a
Snapchat interaction for the sender is the seen confirmation (Bayer et al., 2016).
In a study by Bayer et al. (2016) another consequence of the temporary character is
highlighted. They found that Snapchat was more likely to command the full attention of the receiver
as opposed to other social media platforms. Nevertheless, because of the mundane nature of the
content, users often do not remember the specific contents of their recent exchanges. For example,
when asked about his last Snapchat use, P1 replied: “They were just literally a picture of my face
and a couple were, just something, I don’t know exactly what it was but it was nothing.” So users
have full attention but won’t remember the specific content in the end. It is expected that this will
lead to less effectiveness of celebrity endorsement through Snapchat.
Because of the consequences of the temporal visible shared content on Snapchat, it does not
seem to be conducive to the transfer and effectiveness of celebrity endorsement. For these reasons I
H2: Celebrity endorsement leads to higher purchase intentions via social media platforms characterized by permanently visible content than platforms characterized by temporary visible content.
2.5 Layout x Duration
In this study, we link textual-permanent content to Twitter, visual-permanent content to
Instagram and visual-temporary content to Snapchat. Assuming that hypotheses 1 and 2 are
significantly correct, this would mean that both permanently (versus temporarily) visible and visual
(versus textual) advertisements lead to higher purchase intentions. Instagram has a combination of
these two characterizations. As mentioned before, Instagram is a social media platform that focuses
on visual storytelling which is permanently available. Research has been done on the effectiveness
of advertising through Instagram and proven efficient (Djafarova & Rushworth, 2017; Miles, 2014).
Beside Instagram is third on the list of most popular social media channels to learn about new
products according to consumers in the United States as of July 2016 (Statista, 2016), and is
expected to have 32.3 million brand sponsored influencer posts on Instagram in 2019 (Statista,
2016). But none of these researches focused on a comparison between Instagram, Twitter, and
Facebook, and which platform is most efficient. Since we propose the hypothesis that platforms
characterized by permanence and textuality lead to the highest effectiveness, pairing these two
characterizations should lead to even higher purchase intentions. Therefore I hypothesize:
H3: The positive effect of visuals rather than texts is improved by permanence (compared to temporality).
Figure 2. Effectiveness tree
Celebrity endorsed advertisement Temporal Snapchat Instagram Twitter Text Visual Twitter Instagram Permanent
3. Methodology
In order to answer the research question, I want to do a quantitative, exploratory research.
To build the first and second chapter, use is made of secondary data, found in studies that refer to
the research question (Saunders et al., 2009). The fourth chapter will be dedicated to the
quantitative study. It has been decided to collect data of at least 200 participants. Then in the fifth
chapter, the results will be discussed and compared with the existing literature. Since we will use
existing literature at the start of the research, but will also discuss the results, this will lead to a
combination of both deductive and inductive research (Saunders et al., 2009). In the sixth chapter,
the implications, ideas for future research and limitations will be discussed. Finally, a conclusion
will be drawn.
3.1. Research design
This research has been conducted by a randomized experimental, 2 (temporary versus
permanent advertisement) x 2 (textual versus visual advertisement) between subject design. This
design allows to investigate the possible differences on the scores on the dependent variable
‘purchase intention of the brand’ and to discover possible causal relations between the independent
variables duration (temporary versus permanent advertisement) and layout (textual versus visual
advertisement (Boeije, Hart & Hox, 2009).
3.2. Participants
A sample was recruited in May 2018. 210 Dutch participants were recruited through a
convenience sampling method. The researcher shared the questionnaire on her social media
platforms, and via a mail-list. For legal reasons, only responses from participants aged 18+ were
taken into account. Participating in the research was voluntary. Only 200 respondents were used for
the final analysis, as 9 respondents did not complete the survey and 1 respondent was a minor and
gender, 76.1% were females (N=152), whereas males (N=48) represented 23.9%. Regarding their
education, 58.5% of the participants studied at University, whereas 26% of them studied at HBO,
4% studied MBO and 11.5% of the respondents were still attending secondary school. Respondents
were randomly assigned through Qualtrics to one of the four conditions. Condition one contained
49 respondents, condition two 49 respondents, condition three 51 respondents and condition four 51
respondents. Almost as many participants in each group. Hence, randomization of the respondents
was successful χ2 (3) = 0.080, p = 0.994.
3.3. Stimulus material
The celebrity endorsed advertisement that is used in this research is the one of George
Clooney in the Nespresso campaign; ‘Nespresso, what else?’. This ad has been shown in four
different ways. (1) a permanent visible photo of George with a cup of Nespresso coffee, (2) a
temporary (5 seconds) visible photo of George with a cup of Nespresso coffee, (3) a permanent
visible positive statement from George Clooney about Nespresso and (4) a temporary (5 seconds)
visible positive statement by George Clooney about Nespresso. This advertisement is chosen
because firstly, the product is unisex. Consequently, I expect to prevent inequalities based on the
attractiveness of the product and therefore avoid an effect that is caused by gender. Secondly, it
appeals to a wide range of ages. So that it could be interesting to almost every participant.
3.4. Procedure
Respondents were recruited through a link that was shared on the social media platform
Facebook, via WhatsApp and e-mail. The survey started with a brief introduction. Firstly, they were
thanked for their participation. Then the research was explained. Finally, a declaration of consent
followed, which they agreed on as soon as they pushed the continue button. After the introduction,
respondents had to fill in some demographic information; gender, age, and highest education level.
visible textual celebrity endorsed advertisement, (2) a temporary visible visual celebrity endorsed
advertisement, (3) a permanent visible textual celebrity endorsed advertisement and (4) a permanent
visible visual celebrity endorsed advertisement. After the participants were exposed to one of the
four conditions, the purchase intention towards the brand was measured. Finally, the control
variables were tested. This included the familiarity with the celebrity (George Clooney), the brand
(Nespresso) and the campaign (‘Nespresso, what else?’). The average duration of the survey was
62,95 (SD=45,285) seconds per respondent. 12 respondents were detected who took more than 5
minutes. After all the participants were thanked again for their participation.
3.5. Measures
To test respondents’ ‘purchase intention’ a 5-point semantic differential scale is used based
on Barber, Kuo, Bishop and Goodman (2012) ranging from disagree/agree. This variable was
recoded so that high scores actually meant that individuals had high purchase intentions. The items
are as follows: ‘I would consider purchasing an item of Tommy Hilfiger’, ‘I am planning to
purchase an item of Tommy Hilfiger’, ‘I am interested in buying an item of Tommy Hilfiger’, ‘I
would definitely purchase an item of Tommy Hilfiger’. Besides that, a 5 point semantic differential
scale is used based on Close, Finny, Lacey and Sneath (2006) ranging from disagree/agree with the
item: ‘I am more likely to purchase an item of Tommy Hilfiger than of any other brand’. This
variable is also recoded so that high scores actually meant that individuals had high purchase
intentions. To test the reliability of these several items, a reliability analysis was conducted. It is
stated that when your Cronbach’s α >0,70 your scale is reliable. This means that the questions
measure the same. The answers that people give are related. In this case, Cronbach’s α for the
intention to purchase items= 0.91, which indicates that the scale was very reliable (M = 3.44; SD =
3.6. Control Variables
A few questions were asked in order to control for the dependent variables. The
demographic data; gender, age and highest education were asked. In addition, questions were asked
about the participant's familiarity with the brand Nespresso, the campaign; Nespresso, what else?
and the endorser; George Clooney was asked. Respondents could indicate their awareness of the
celebrity, brand, and campaign with the answer options ‘yes’ or ‘no’ (Tieman, 2017).
4. Results
4.1 Preliminary Analysis
To ensure internal validity, the variables gender, education, and age were checked in order to
see whether respondents were sufficiently randomized during the experiment. As shown in table 1
all variables do not deviate significantly. This means that, based on gender, age and education,
participants were randomly assigned to one of the four conditions. It was also checked whether age,
gender, education, familiarity with the brand, campaign or celebrity influences the purchase
intention (table 2). The only notable connection was found between familiarity with the celebrity
and purchase intention. The familiarity with the celebrity is significant, so influences the intention
to purchase.
Table 1. Randomization checks
Layout * Duration Statistic Value (V) Degree of
Freedom (Df)
Probability (P)
Gender χ2 4,137 3 0,247
Age F 0,995 3,196 0,396
Table 2. Bivariate Correlations between familiarity and intention to purchase
4.2 Hypotheses Testing
In order to test the three proposed hypotheses, a two-way Analysis of Variance (ANOVA)
was conducted. The first hypothesis (H1) predicted that respondents who were being exposed to a
visual celebrity endorsed advertisement will have a higher purchase intention compared to
respondents who were being exposed to a textual celebrity endorsed advertisement. Results showed
that there was a significant main effect of the type of layout on purchase intention (F(1, 196) =
22.223, p < 0.001). In other words, participants assigned to one of the visual conditions (M = 3.79;
SD = 0.966) were more likely to purchase the product then participants assigned to one of the
textual conditions (M = 3.10; SD = 1.099), (95% Bonferroni CI = [0.257; 1.073]). Therefore, H1 is
supported.
The second hypothesis (H2) predicted that respondents who were being exposed to a
permanently visible celebrity endorsed advertisement would have higher purchase intention
compared to respondents who were being exposed to a temporarily visible celebrity endorsed
advertisement. Results showed that there was no significant main effect of the type of duration on
purchase intention (F (1, 196) = 2.612, p = .108). In other words, participants who were assigned to
Correlation Coefficient (R) Number of Observations (N) Probability (P) Age 0,086 200 0,228 Gender -0,029 200 0,680 Education -0,107 200 0,130
Familiarity with the brand
-0,001 200 0,990
Familiarity with the campaign
0,088 200 0,215
Familiarity with the celebrity
one of the permanent conditions (M = 3.560; SD = 1.140) were as likely to purchase the product
compared to participants who were assigned to one of the temporary conditions (M = 3.320; SD =
1.029), (95% Bonferroni CI = [-0.192; 0.616]). Therefore, H2 is not supported.
The third hypothesis (H3) predicted that the layout (visual content versus textual content)
would have a significant main effect on consumers’ purchase intention, moderated/improved by the
type of duration (permanent visible versus temporary visible). A two-way Analysis of Variance
(ANOVA) was conducted and results showed no significant association of layout on consumers’
purchase intention, moderated by the duration (F (1, 196) = 0.029, p = .866). In other words,
participants who were assigned to the condition; permanent visible visual content (M = 3.92; SD =
1.042) were as likely to purchase the product compared to participants who were assigned to one of
the three other conditions; temporary visible visual content (M = 3.66; SD = 0.875), permanent
visible textual content (M = 3.21; SD = 1.130), temporary visible textual content (M = 3.00; SD =
!
Figure 3: Interaction effects between the type of layout and type of duration on consumers’ purchase intention.
4.3 Checking for control variables
The two-way ANOVA is repeated with the previously significantly proven variable (table 2).
No changes of significance can be identified (Table 3). The results reported in the hypothesis testing
seem robust - at least against the here introduced control variable; familiarity with the celebrity.
Table 3: significance of the three control variables
Familiarity with the celebrity
F df p
Layout 19,956 1 <0.001
Duration 2,266 1 0,134
Layout * Duration 0,046 1 0,830
5. Conclusion
5.1 Discussion and link to former research
This research investigated whether celebrity endorsed advertising is most effective, in terms
of purchase intentions, via Snapchat Twitter or Instagram, based on two different characterizations;
duration and layout. The main research question is; Does celebrity endorsement leads to the highest
purchase intentions via Snapchat, Instagram or Twitter? Decades of studies focused on the
effectiveness of celebrity endorsement, which is described in the theoretical framework. Centeno &
Wang (2016) described the co-creation process of celebrity endorsement in the Actor-Network
Theory. One of the reasons for its effectiveness is that celebrity endorsed advertisements (versus
normal advertising) are received differently from consumers. This is explained by the Persuasion
Knowledge Model by Friestad & Wright (1994). Belch & Belch (2004) proved that celebrity
endorsement is an effective strategy to attract customers. It has the power to break through
advertisement clutters (Choi & Rifon, 2007). Another reason for the effectiveness of celebrity
endorsed advertisements is determined by the credibility of the source (the celebrity). This is
described by Petty & Cacioppo (1986) in the Source Credibility Theory and explained by Kelvin
(1961) and Ohanian (1990) with the Internalization Process. The last significant reason for the
effectiveness of celebrity endorsements is determined by the fit between the celebrity and the
customers. This described by in the Meaning Transfer Model (McCracken, 1989), the Match-Up
Hypothesis Model (Kamins, 1990), the Identification Process (Friedman & Friedman, 1979) and by
Hsu & Macdonald (2002) in their study. On the basis of above research, it is concluded that
celebrity endorsed advertising is effective. Little research has been done on via which platform this
leads to the best results. This study attempted to investigate this question on the basis of 3
hypotheses.
The first hypothesis predicted that respondents who were being exposed to a visual celebrity
being exposed to a textual celebrity endorsed advertisement. Results showed a significant difference
and therefore it can be concluded that celebrity endorsed advertising via platforms characterized by
visual content are more effective than via platforms characterized by textual content. For this
research, this result means that advertising via Snapchat and Instagram would lead to a higher
purchase intention than via Twitter. This effect was expected on the basis of several studies who
argument that visuals are more effective than texts. The first reason for this is that visuals lead to
excitement (Diao & Sundar, 2004; Day, Shyi & Wang, 2006). This is explained in the Motion
Effects Theory (Sundar & Kalyanaraman, 2004). The second reason is that visuals are easier to
memorize because there are more possible interpretations (McQuarrie and Phillips, 2005). Another
explanation for this is explained in the Dual Coding Model by Singh et al. (2002). The third reason
is that visuals have a more emotive impact, which leads to a stronger effect (Joffe, 2008). Lastly,
visuals are easier to understand than text. This leads to a stronger effect on the consumer (Powell et
al., 2015).
The second hypothesis predicted that respondents who were being exposed to a permanently
visible celebrity endorsed advertisement would have higher purchase intention compared to
respondents who were being exposed to a temporarily visible celebrity endorsed advertisement. The
results were nonsignificant and hypothesis 2 was rejected. This effect was not expected, based on
the arguments of several studies described in the theoretical framework. We found several reasons
that permanent media (Instagram and Twitter) was significantly more effective than temporary
media (Snapchat). One of these reasons is that social media is valued for it’s archiving capacity
(Zhao & Lindley, 2014). In addition, the main reason people use Snapchat is for fun (Bayer et al.,
2016). Therefore, users often do not remember the content, which is not favorable for the
effectiveness of advertising. Another reason is that users of Snapchat mainly have interaction with
close ties. This is also because Snapchat did not make it easy to add celebrities, which does not
‘snap’ is not visible because there is no like, react feature (Bayer et al., 2016). Because the above
arguments it was not expected that hypothesis 2 was nonsignificant. However, in figure 3 on page
22, it can be seen that for both textual and visual advertisements the purchase intention is higher
when it was permanently visible. A possible explanation for the fact this result was not significant is
that the stimulus material was unclear. Before participants start the survey, they have to read the
introduction in which is stated that research was conducted into the differences between permanent
and temporary visible advertisements. According to Bayer et al. (2016) constraints on content
influences the attention pattern of people. They state that people pay more closely to ‘snaps’
because they know the content disappears after a few seconds. If participants have not carefully
read the introduction, this may have led to the stimulus material being unclear. This could be a
logical explanation for why participants have given different answers than expected and therefore
hypothesis 2 was rejected.
The third hypotheses predicted that respondents who are being exposed to visual celebrity
endorsed advertisements (versus textual) have more positive purchase intentions, which is improved
by permanent visibility (versus temporary visibility) of the advertisement. So we expect that when
consumers see a visual celebrity endorsed advertisement, this positive effect is enhanced if it is also
permanently visible (versus temporary). However, the results showed a nonsignificant effect and H3
was rejected. This means that the choice of duration does not significantly affect the positive
purchase intention after seeing a visual celebrity endorsed advertisement (versus textual). On the
basis of these results, it can be stated that Instagram and Snapchat are just as effective as advertising
platforms. A possible explanation for why hypothesis 3 is not significant has already been explained
in the discussion about hypothesis 2. Again briefly summarized; according to Bayer et al. (2016)
constraints on content influences the attention pattern of people. So if the stimulus material was
unclear to the participants this might have led to different, wrong or unreliable answers.
In the next section, some limitations of the study are proposed with the aim of addressing future
research in the right track.
5.2 Limitations and future research
This study has several limitations. The first limitation is represented by the research design.
Use has been made of an online experiment. This is fast and cheap but entails a drawback. The
researcher cannot control and support the participants while they are taking the survey. As a result
of this, some participants do not complete the survey and have to be excluded from the analysis.
Secondly, due to the lack of time, only 200 people participated in this research. These people
included mainly acquaintances of the researcher. For future research, it is recommended to search
for more participants and to spread the survey more widely. In addition, every organization has its
own target group. This variable was not taken into account in this study. This can lead to results that
do not apply to certain companies. For future research, it is interesting to find out for a specific
company, via what social media platform, celebrity endorsed advertising is most effective for them.
Another limitation of this study is the stimulus material. As stated in the discussion, a
possible explanation for the nonsignificant result of hypothesis 2 is unclear stimulus material. A
control variable must be added, which controls whether all participants know for sure what is being
investigated. This will lead to more reliable results, and possibly to a significant result of hypothesis
2. Besides, the stimulus material lacked some important characteristics and could be improved in
future studies. No characteristic of Snapchat, Twitter or Instagram were included in the layout of the
stimulus material. This might have influenced the purchase intention results.
As said before, this research is a good first attempt at comparison between Snapchat,
Instagram, and Twitter. However, this study is only focused on the distinctive characterizations
duration and layout. To make more concrete statements, other variables should be taken into
an advertisement clutter yet, in contrast to Twitter and Instagram. For future research, this is one of
the variables that is interesting to focus on.
5.3 Theoretical relevance
This study contributes to the scientific research on social media advertising. It is the first
empirical study that researches the differences in the effectiveness of celebrity endorsed advertising
via Snapchat, Instagram, and Twitter. The results of this study indicate that advertising via Snapchat
and Instagram has a more positive effect on purchase intention than advertising via Twitter. This
difference in effect is caused by the visual character of Instagram and Snapchat and the textual
character of Twitter. The type of duration of the advertisement does not have a significant effect on
purchase intention. The positive effect on purchase intention of platforms characterized by visual
content (versus textual) is not improved by the type of duration. This study is a good starting point
for future research when the limitations of this study are taken into account. In conclusion, it can be
stated that according to this study, celebrity endorsed advertising via Instagram and Snapchat should
6. References
Adiga, N. R., Almási, G., Almasi, G. S., Aridor, Y., Barik, R., Beece, D., ... & Bright, A. A.
(2002, November). An overview of the BlueGene/L supercomputer. In Supercomputing, ACM/IEEE
2002 Conference (pp. 60-60). IEEE
Barber, N., Kuo, P. J., Bishop, M., & Goodman Jr, R. (2012). Measuring psychographics to
assess purchase intention and willingness to pay. Journal of consumer marketing, 29(4), 280-292.
Bayer, J. B., Ellison, N. B., Schoenebeck, S. Y., & Falk, E. B. (2016). Sharing the small
moments: ephemeral social interaction on Snapchat. Information, Communication & Society, 19(7),
956-977.
Belch, G. E., & Belch, M. A. (2004). Advertising and Promotion: An Integrated Marketing
Communications Perspective,(6 Th. New York: NY: McGraw-Hill.
Bernstein, M. S., Bakshy, E., Burke, M., & Karrer, B. (2013, April). Quantifying the
invisible audience in social networks. In Proceedings of the SIGCHI conference on human factors
in computing systems (pp. 21-30). ACM.
Bhatnagar, N., Aksoy, L., & Malkoc, S. A. (2004). Embedding brands within media content:
The impact of message, media, and consumer characteristics on placement efficacy. The psychology
of entertainment media: Blurring the lines between entertainment and persuasion, 99-116.
Boeije, H. (2009). H.‘t Hart, & J. Hox, Onderzoeksmethoden (Den Haag: Boom Lemma
Uitgevers, 2009).
Centeno, D., & Wang, J. J. (2017). Celebrities as human brands: An inquiry on
stakeholder-actor co-creation of brand identities. Journal of Business Research, 74, 133-138.
CHOI, S., & Rifon, N. J. (2007). Who is the celebrity in advertising? Understanding
dimensions of celebrity images. The journal of popular Culture, 40(2), 304-324.
through event marketing: Linking attendees with the sponsor, community, and brand. Journal of
advertising research, 46(4), 420-433.
Day, R. F., Shyi, G. C. W., & Wang, J. C. (2006). The effect of flash banners on
multiattribute decision making: distractor or source of arousal?. Psychology & Marketing, 23(5),
369-382.
Dehghani, M., Niaki, M. K., Ramezani, I., & Sali, R. (2016). Evaluating the influence of
YouTube advertising for attraction of young customers. Computers in human behavior, 59, 165-172.
Dekker, K., & van Reijmersdal, E. (2010). Waarschuwingen, beroemdheden en brand
placement: De effecten van type waarschuwing en geloofwaardigheid op kijkerreacties. Tijdschrift
voor Communicatiewetenschap, 38(4), 320.
Diao, F., & Sundar, S. S. (2004). Orienting response and memory for web advertisements:
Exploring effects of pop-up window and animation. Communication research, 31(5), 537-567.
Djafarova, E., & Rushworth, C. (2017). Exploring the credibility of online celebrities'
Instagram profiles in influencing the purchase decisions of young female users. Computers in
Human Behavior, 68, 1-7.
Joffe, H. (2008). The power of visual material: Persuasion, emotion and
identification. Diogenes, 55(1), 84-93.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to
theory and research.
Friedman, H. H., & Friedman, L. (1979). Endorser effectiveness by product type. Journal of
advertising research.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with
persuasion attempts. Journal of consumer research, 21(1), 1-31.
Gigi Hadid verlengt samenwerking met Tommy Hilfiger. (2017, May 10). Retrieved April
https://www.vogue.nl/fashion/nieuws/artikel/gigi-hadid-verlengt-samenwerking-met-tommy-hilfiger
Green, T., & Peloza, J. (2011). How does corporate social responsibility create value for
consumers?. Journal of Consumer Marketing, 28(1), 48-56.
Gwinner, K., (1997). A model of image creation and image transfer in even sponsorship.
International Marketing Review, 14(3), 145-158.
Helm, A. E., Moulard, J. G., & Richins, M. (2015). Consumer cynicism: developing a scale
to measure underlying attitudes influencing marketplace shaping and withdrawal behaviours.
International journal of consumer studies, 39(5), 515-524.
Holt, D. B. (2004). How brands become icons: The principles of cultural branding. Harvard
Business Press.
Hsu, C. K., & McDonald, D. (2002). An examination on multiple celebrity endorsers in
advertising. Journal of Product & Brand Management, 11(1), 19-29.
Hutter, K., Hautz, J., Dennhardt, S., & Füller, J. (2013). The impact of user interactions in
social media on brand awareness and purchase intention: the case of MINI on Facebook. Journal of
Product & Brand Management, 22(5/6), 342-351.
Just, M. A., & Carpenter, P. A. (1992). A capacity theory of comprehension: individual
differences in working memory. Psychological review, 99(1), 122.
Kamins, M. A. (1990). An investigation into the “match-up” hypothesis in celebrity
advertising: When beauty may be only skin deep. Journal of advertising, 19(1), 4-13.
Kaun, A., & Stiernstedt, F. (2014). Facebook time: Technological and institutional
affordances for media memories. New Media & Society, 16(7), 1154-1168.
Kelvin, N. (1961). A troubled Eden; nature and society in the works of George Meredith.
Stanford University Press.
Kelman, H. C. (1961). Three processes of social influence. Public Opinion Quarterly, 25,
Lafferty, B. A., Goldsmith, R. E., & Newell, S. J. (2002). The dual credibility model: The
influence of corporate and endorser credibility on attitudes and purchase intentions. Journal of
Marketing Theory and Practice, 10(3), 1-11.
Lam, S. K., Ahearne, M., Mullins, R., Hayati, B., & Schillewaert, N. (2013). Exploring the
dynamics of antecedents to consumer–brand identification with a new brand. Journal of the
Academy of Marketing Science, 41(2), 234-252.
Law, A. M., Kelton, W. D., & Kelton, W. D. (2007). Simulation modeling and analysis (Vol.
3). New York: McGraw-Hill.
Leonardi, P. M., & Treem, J. W. (2012). Knowledge management technology as a stage for
strategic self-presentation: Implications for knowledge sharing in organizations. Information and
Organization, 22(1), 37-59.
Lu, L. C., Chang, W. P., & Chang, H. H. (2014). Consumer attitudes toward blogger’s
sponsored recommendations and purchase intention: The effect of sponsorship type, product type,
and brand awareness. Computers in Human Behavior, 34, 258-266.
McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the
endorsement process. Journal of consumer research, 16(3), 310-321.
McQuarrie, E. F., & Phillips, B. J. (2005). Indirect persuasion in advertising: How
consumers process metaphors presented in pictures and words. Journal of advertising, 34(2), 7-20.
Milde, K., & Yawson, R. (2017). Strategies for Social Media Use in Nonprofits.
Miles, J. (2014). Instagram power: Build your brand and reach more customers with the
power of pictures. New York, NY: McGraw-Hill Education.
Misra, S., & Beatty, S. E. (1990). Celebrity spokesperson and brand congruence: An
assessment of recall and affect. Journal of business research, 21(2), 159-173.
Nilashi, M., Ibrahim, O., Mirabi, V. R., Ebrahimi, L., & Zare, M. (2015). The role of
and Consumer Services, 26, 57-69.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers'
perceived expertise, trustworthiness, and attractiveness. Journal of advertising, 19(3), 39-52.
Özkul, D., & Humphreys, L. (2015). Record and remember: Memory and meaning-making
practices through mobile media. Mobile Media & Communication, 3(3), 351-365.
Petty, R. E., & Cacioppo, J. T. (1986). Message elaboration versus peripheral cues.
In Communication and persuasion (pp. 141-172). Springer, New York, NY.
Powell, T. E., Boomgaarden, H. G., De Swert, K., & de Vreese, C. H. (2015). A clearer
picture: The contribution of visuals and text to framing effects. Journal of Communication, 65(6),
997-1017.
Roberts, K. G. (2016). Communication Theory and Millennial Popular Culture. Essays and
Applications, 8, 167-173.
Saunders, M. L., & Lewis, P. (2009). P. & Thornhill, A.(2009). Research methods for
business students, 4.
Silvera, D. H., & Austad, B. (2004). Factors predicting the effectiveness of celebrity
endorsement advertisements. European Journal of marketing, 38(11/12), 1509-1526.
Statista. (2016). Number of brand sponsored influencer posts on Instagram from 2016 to
2019 (in millions) [Dataset]. Retrieved June 21, 2018, from https://www.statista.com/statistics/
693775/instagram-sponsored-influencer-content/
Statista. (2016, June). Most popular social media channels to learn about new products
according to consumers in the United States [Dataset]. Retrieved June 21, 2018, from https://
www.statista.com/statistics/641458/usa-social-media-usage-to-learn-about-new-products/
Statista. (2017, september). Number of daily active Instagram users from October 2016 to
September 2017 (in millions) [Dataset]. Retrieved May 20, 2018, from https://www.statista.com/
Statista. (2018, mei). Number of daily active Snapchat users from 1st quarter 2014 to 1st
quarter 2018 (in millions) [Dataset]. Retrieved May 20, 2018, from https://www.statista.com/
statistics/545967/snapchat-app-dau/
Statista. (2018, April). Number of monthly active Twitter users worldwide from 1st quarter
2010 to 1st quarter 2018 (in millions) [Dataset]. Retrieved May 20, 2018, from https://
www.statista.com/statistics/282087/number-of-monthly-active-twitter-users/
Statista. (2018, April). Most popular social networks worldwide as of April 2018, ranked by
number of active users (in millions) [Dataset]. Retrieved May 20, 2018, from https://
www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/
Steffes, E. M., & Burgee, L. E. (2009). Social ties and online word of mouth. Internet
research, 19(1), 42-59.
Stokburger-Sauer, N., Ratneshwar, S., & Sen, S. (2012). Drivers of consumer–brand
identification. International journal of research in marketing, 29(4), 406-418.
Sundar, S. S., & Kalyanaraman, S. (2004). Arousal, memory, and impression-formation
effects of animation speed in web advertising. Journal of Advertising, 33(1), 7-17.
Swedowsky, M. (2009). Improving Customer experience by listening and responding to
social media.
Tieman, A. N. N. E. (2017). Events, Celebrities and Instagram. Retrieved from http://
arno.uva.nl/document/651367
Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness,
expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of
advertising, 29(3), 1-13.
Uzunoğlu, E., & Kip, S. M. (2014). Brand communication through digital influencers:
Leveraging blogger engagement. International Journal of Information Management, 34(5),
Zhao, X., & Lindley, S. E. (2014, April). Curation through use: understanding the personal
value of social media. In Proceedings of the 32nd annual ACM conference on Human factors in
7. Appendix: survey
Page 1
Page 3: a) & b) the visual advertisement disappears after 3 seconds
Page 4