MASTER THESIS
EXCLUSIVE VS. MULTIPLE CELEBRITY ENDORSEMENT
IS “ANY PUBLICITY A GOOD PUBLICITY”
FOR THE BRAND?
AUTHOR: ANTOANETA KYOSEVA STUDENT NUMBER: 10224114
SUPERVISORS: JORGE LABADIE & DRS. ROGER PRUPPERS SUBMISSION DATE: AUGUST 1ST 2014
UNIVERSITY OF AMSTERDAM
FACULTY OF ECONOMICS AND BUSINESS
MSC BUSINESS STUDIES – MARKETING TRACK
2 | P a g e
Abstract
One of the main questions for companies has been how to build strong brands, in order to
stay competitive. One of the most frequently applied strategies to achieve this has been
engaging in a celebrity endorsement deal, as it has demonstrated that famous spokespersons
do enhance brand knowledge and strengthen brand image and awareness.
While celebrity endorsement has considerably gained in popularity over the past decades,
little knowledge has been obtained with regards to the concept of multiple endorsements.
Therefore the current study sought to explore the differences in the effects of using exclusive
and multiple endorsers, and what are the best endorser-brand combinations in terms of fit for
enhancing brand attitude, perceived brand quality and purchase intentions. Incorporating
Source credibility as an important factor for facilitating this process, the study provided
insights into the higher effectiveness of exclusive endorsers, especially on improving Brand
attitude and Purchase intentions, and into the higher effectiveness of higher levels of
endorser-brand fit, especially in terms of influencing Perceived quality.
In addition, it was evident that it is a challenge for multiple endorsers to outweigh the
positive effects of their exclusive counterparts. Nevertheless, the study demonstrates that
multiple endorsers are not so disadvantageous after all, especially when promoting a highly
3 | P a g e
Acknowledgments
I would like to express my gratitude to my supervisor, Mr. Jorge Labadie and my second
reader, Mr. Roger Pruppers, for their support and constructive feedback.
4 | P a g e
Table of Contents
ABSTRACT ... 2 ACKNOWLEDGMENTS ... 3 1. INTRODUCTION ... 7 1.1 RESEARCH FRAMEWORK ... 9 1.2 PROBLEM DEFINITION: ... 131.3 DELIMITATIONS OF THE STUDY ... 14
1.4 THEORETICAL CONTRIBUTIONS ... 15
1.5 MANAGERIAL CONTRIBUTIONS ... 16
1.6 RESEARCH OUTLINE ... 17
2. LITERATURE REVIEW ... 17
2.1 BRAND KNOWLEDGE AND CUSTOMER-BASED BRAND EQUITY ... 17
2.2 DIMENSION OF BRAND KNOWLEDGE AND BRAND ASSOCIATIONS ... 18
2.3 SECONDARY BRAND KNOWLEDGE AND ASSOCIATIONS AND THEIR SOURCES ... 19
2.4.1CELEBRITY ENDORSEMENT:CHARACTERISTICS,ADVANTAGES &DISADVANTAGES ... 20
2.4.2.MODELS EXPLAINING THE CELEBRITY ENDORSEMENT PROCESS ... 22
2.4.3MULTIPLE ENDORSEMENT ... 23
2.4.4MODELS EXPLAINING MULTIPLE ENDORSEMENT ... 25
2.4.5THE CONCEPT OF CONGRUENCE ... 27
2.4.6MODELS EXPLAINING THE CONCEPT OF FIT ... 32
3. HYPOTHESES ... 33
4. METHODOLOGY ... 39
4.1 RESEARCH DESIGN ... 40
4.2. PRE-TEST ... 40
4.3 STIMULI AND MANIPULATIONS ... 41
4.4 PROCEDURE ... 41 4.5 MEASUREMENTS ... 42 4.5.1INDEPENDENT VARIABLES ... 42 4.5.2DEPENDENT VARIABLES ... 42 4.6 STIMULI DEVELOPMENT ... 44 4.6.1.PRE-TEST ... 44
4.6.2PRE-TEST RESULTS SUMMARY... 47
5 | P a g e
5.1 DATA CLEANING ... 51
5.2 RANDOM ASSIGNMENT ... 52
5.3 MANIPULATION CHECK ... 53
6. HYPOTHESES TESTING ... 55
6.1 TREATMENT EFFECT ON SOURCE CREDIBILITY... 55
6.2 TREATMENT EFFECT ON BRAND ATTITUDE ... 57
6.3 TREATMENT EFFECT ON PURCHASE INTENTIONS ... 58
6.4 TREATMENT EFFECT ON PERCEIVED QUALITY ... 59
6.5 EFFECT OF CELEBRITY-BRAND COMBINATIONS ON BRAND ATTITUDE, PURCHASE INTENSIONS AND PERCEIVED QUALITY ... 59
6.5 ADDITIONAL ANALYSIS: EFFECT OF LEVEL OF FIT ON BRAND ATTITUDE, PURCHASE INTENTIONS AND PERCEIVED QUALITY ... 61
6.6 EFFECT OF SOURCE CREDIBILITY ON BRAND ATTITUDE, PURCHASE INTENTIONS AND PERCEIVED QUALITY ... 62
6.7 SOURCE CREDIBILITY AS MEDIATOR ... 63
6.8 EFFECT OF PRODUCT USAGE AND PERCEIVED CELEBRITY INFLUENCE ... 64
7. DISCUSSION ... 66
7.1 INTERPRETATION OF THE RESULTS AND EXPLANATION OF THE UNEXPECTED OUTCOMES ... 66
7.2 THEORETICAL IMPLICATIONS ... 72
7.3 MANAGERIAL IMPLICATIONS ... 74
8. CONCLUSION... 76
8.1 SUMMARY AND ANSWER PROBLEM STATEMENT ... 76
8.2 LIMITATIONS AND SUGGESTIONS FOR FURTHER RESEARCH ... 77
REFERENCES: ... 79
APPENDIX A – SPSS OUTPUT PRE-TEST ... 87
APPENDIX B – SPSS OUTPUT MAIN RESEARCH ... 91
6 | P a g e
List of Figures
Figure 1 Secondary sources of brand knowledge (Keller, 2003) ... 19
Figure 2 Conceptual framework...34
Figure 3 Estimated marginal mean plot- Source Credibility ... 56
Figure 4 Estimated marginal mean plot- Brand Attitude ... 57
Figure 5. Estimated marginal mean plot - Purchase intentions ... 58
Figure 6. Estimated marginal mean plot - Perceived Quality ... 59
List of Tables
Table 1 Endorser brand fit ... 49Table 2 Product usage ... 50
Table 3 Data cleaning – Missing values ... 51
Table 4 Data cleaning – Frequencies ... 51
Table 5 Experimental Groups statistics ... 53
Table 6 Manipulation check – Endorser type ... 54
Table 7 Manipulation check – Level of fit... 54
Table 8 Manipulation check - Familiarity ... 55
Table 9 Means of Dependent variables per Experimental group ... 55
Table 10 Mean difference experimental groups – Source credibility ... 57
Table 11 Mean differences – Brand attitude ... 60
Table 12 Mean differences – Purchase intentions ... 61
Table 13 Mean differences – Perceived quality ... 61
Table 14 Differences in effect size ... 61
Table 15 Correlation between Credibility and Brand attitude, Purchase intentions and Perceived quality ... 63
7 | P a g e
1. Introduction
Smiling faces, beautiful bodies, luxurious appeals. Luring celebrities are peeking from
shopping windows, TV screens and ground posters up to billboards in the air. This is how our
today`s world looks like – glamorous, just like the famous faces smiling at us from all corners
of our daily life. As a result, consumers have inevitably become “conditioned” to associate
the brands they use with the celebrity figures attached to them (Till, 1998; Kim et al., 1998).
This common advertising practice has been proven to be quite successful. To illustrate this,
consider e.g. George Clooney. What do you think about when you see his picture? Maybe,
besides the movies ER and Ocean`s eleven, also brands like “NESPRESSO” and “Martini”
come to mind? And how about noticing a photo of David Beckham – probably apart from
Manchester United, also the brands “Adidas”, “Pepsi”, and “H&M” pop up?
These are only few examples of the diverse set of existing celebrity endorsements. This
branding strategy has found frequent application in the last years, because of its effectiveness
for increasing brand awareness and even though it may also have potential downsides, such
as loss of control and negative feedback effects, (www.nytimes.com; Erdogan, 1999), it is
used, as exemplified, in a wide range of business fields – FMCG, fashion & cosmetic
industry, sports, film industry, etc. This practice is understandable, since, for a long time,
establishing brand awareness and image have been key goals for brand management (Esch et
al., 2006).
Interestingly, it has been noticed, that in the branding practise, several perspectives on the
celebrity endorsement have been taken. For example, using several spokespersons in various
fields, ranging from huge companies like “L'Oréal” (several endorsers for different varieties
of the same brand), (http://www.lorealparisusa.com/), to causes such as the “Got Milk”
8 | P a g e Furthermore, some companies have stuck to their endorser, who has been promoting their
brand exclusively, without being connected to any other brands, e.g. Jennifer Hudson for
“WeightWatchers” and Lance Armstrong for “Anheuser-Busch”
(www.celebrityendorsementads.com). Moreover, very often companies choose to attach their
brand to a celebrity, which has already been engaged in endorsing other brands. One of the
most legendary examples of such a celebrity spokesperson is Michael Jordan, who has been
promoting multiple brands like “Wheaties”, “Coca-Cola”, “Nike”, “Food Lion”, “Gatorade”,
“Hanes”, etc. (Till, 1998), making an impact worth 14 billion dollars on the American economy (Erdogan et al., 2001). Moreover, the Chicago Bulls star has also endorsed
“McDonalds”, which could be found as a bit contradictory to his athletic image, thus posing the question about the fit between the brand and the pitchman and how its absence would
reflect on the brand. On the other hand, a recent example of a fit between the brand and the
spokesperson could be considered Charlie Sheen`s exclusive endorsement of the beer brand
“Bavaria” (www.rtl.nl ). Considering his scandalous image, though, another question arises – whether the fit always has a positive impact on the brand. These instances demonstrate that
different companies approach the endorsement strategy in different ways. The question is –
which one is the best and what companies should take into account so that their brand can
fully benefit from the endorsement deal.
Thus, all these examples show that the branding strategy of celebrity endorsements in all their
forms have undoubtedly became part of our lives as consumers. Their diversity makes them
quite interesting in terms of observing in what combinations they will be further applied and,
especially, with what result – how successful they will be in capturing as much attention as
possible and influencing consumers’ brand image and associations.
According to B. Bonin Bough,Vice President of Global Digital and Consumer Engagement at
9 | P a g e intense” (http://blogs.forbes.com/boninbough/). So, given the fierce competition about every single client in today`s world of abundant offerings, companies really need to focus on their
awareness and image building strategies. In order to differentiate from their competitors and
to attract and retain more customers living in a world of increasing commercial supply,
companies need to build strong brands via increasing consumers` brand knowledge and thus
establishing favourable associations about the brand in the minds of the consumers (Keller,
2012, p.58).
Therefore, it would not be surprising that companies, following these recommendations,
would continue applying the endorsement strategy intensively in future. What they should be
aware of in advance, though, is in which situations to apply which endorsement – to focus on
an exclusive celebrity or to bet on a widely used one; to keep in mind the level of fit between
the celebrity and the brand, and what the outcomes of the final endorsement deal would be.
These are the topics the following study will address.
1.1 Research framework
In today's celebrity-driven culture, it's tough to find any new product launch without a big
name attached to it. (http://www.forbes.com/). This is why, as the economy gets increasingly
networked, it is of vital importance for companies to understand how consumers are affected
from the linkage of their brand to other entities such as another person, place, thing, etc.
(Keller, 2003). This practice of linking a particular brand to other partners, also called brand
leverage, has been widely used, in order to induce secondary associations about the brand in
the minds of consumers’ (Keller, 2012). One of the questions that Keller (2001) himself poses when discussing his five brand imperatives is namely “In what ways do the images of
country of origin or country of brand, celebrity spokespeople, retail store, etc change or
10 | P a g e changes consumers’ brand perceptions and indicates that further research should be made regarding the way consumers` brand evaluations are altered when celebrity spokespersons are
used as endorsers. Furthermore, in the context of multiple entities attached to a brand, Uggla
(2004) suggests separating the contributions of each partner to the brand equity and
researching how much should be attributed to each one.
Thus, the aim of this study would be to explore how secondary associations created by
celebrity endorsers affect consumers` brand associations and evaluations1, and in particular,
whether the exclusivity of the celebrity and the celebrity-brand fit make a difference in terms
of the level of brand image, perceived quality and purchase intentions.
There are several reasons for the focus on the role of celebrities. First, celebrity endorsement
has been considered “a ubiquitous feature of modern marketing” (McCracken, 1989). This means that addressing this particular topic will be in line with the current advertising trends,
making the present study relevant and up-to-date.
Second, this branding strategy has been an object of thorough research interest, which can be
seen in the variety of studies in the last years (e.g. Halonen-Knight & Hurmerinta, 2010;
Amos et al., 2008; Uggla, 2008; Seno & Lukas, 2007; McDonald et al., 2002; Erdogan, 1999;
Till, 1998; etc.). This will be a fruitful academic foundation for the present study to step on.
Third, as Keller (2003) discussed, there is limited research about the role of the multiple
entities to which a brand can be linked. A “gap” in the literature has been discovered,
following Keller’s (2003) suggestions, namely the impact of the uniqueness of the linkages - i.e. are the other entities already linked to different (single or multiple) partners. In the
context of the present study this would mean researching the effect of exclusivity vs.
multiplicity of celebrity endorsers. Moreover, McDonald et al (2002) also indicate the lack of
1
11 | P a g e research focused on the phenomenon of multiple celebrity endorsements. Although this
strategy has been considered most frequent form of celebrity endorsement practice in the
advertising industry, research about it has been scarce (Um, 2008). Even the limited findings
about multiple endorsement effectiveness are controversial. On the one hand, having linkages
to too many entities would lead to image dilution through overexposure (Uggla, 2004). On
the other hand, from an equity perspective it is important for brands to have a large number of
associations (Krishnan, 1996). Thus, from the perspective of the trend of multiple
endorsements, the study will try to observe the effect of more linkages to more brands on the
consumers’ brand evaluations, in comparison with exclusive endorsements.
Fourth, focusing namely on celebrities suggests using examples of well-known brands in
combination with prominent faces from the media, which would make the study even more
engaging for the readers, as well as for the participants in the research.
Furthermore, an emphasis will be put on the congruence between the celebrity and the
endorsed brand(s). Previous research has found out that the level of fit between both parties
has a strong impact on consumers` brand evaluations (Kamins, 1989, 1990; Misra & Beatty,
1990, Kamins & Gupta, 1994, Till & Busler, 2000) therefore its effect should be taken into
account. But to the author’s knowledge there is very limited research incorporating the
congruence in relation to multiple celebrity endorsements and comparing the effects with
exclusive ones. Therefore, Tripp`s et al (1994) suggestion for further investigation will be
followed, specifically: “including of the congruence hypotheses in a multiple endorsement
context”.
In addition to the context of celebrity-brand congruence, it has not yet been investigated
which celebrity-brand combination would have more favourable effect on consumer`s brand
12 | P a g e less-matching brand, or would a multiple endorser promoting a high-matching brand be
actually more effective. Before singing a celebrity as a spokesperson for their brand,
companies are facing, on the one hand, a big pool of endorsers to choose from, and on the
other hand – a lot of factors to consider, such as budget, image, brand characteristics, target
audience, among other (Erdogan, 1999). Therefore, the learnings from an investigation about
the effects of the different combinations of endorsers and congruence levels would provide a
clearer insight of the most beneficial combination for the specific brand.
Moreover, authors have been focusing on the models of Source Credibility (Hovland &
Weiss, 1951) and Source Attractiveness (McGuire, 1985) in order to observe celebrities’
effect on consumers’ persuasion and, following, brand evaluations, including purchase intentions (Friedman et al., 1976; Petty & Cacioppo, 1983; Ohanian, 1991; Till & Busler,
1998; etc.). These findings demonstrated that celebrities` perceived attractiveness,
trustworthiness and expertise are an indicator for celebrity endorsement effectiveness and
have a significant impact on consumers` brand perceptions. Therefore, to research this effect
in a multiple endorsement context, as suggested by Amos et al, (2008) the attractiveness
construct will be added to the Source Credibility model and will be considered following
Ohanian`s (1991) Source Credibility scale.
In addition, the present study will follow Low & Lamb`s (2000) approach of measuring brand
associations in terms of brand attitude and perceived quality, as the authors suggest that
assessing a brand`s attitude and perceived quality is “the best way to begin measuring brand associations”. Their argumentation of choosing these particular constructs are that they reflect Aaker`s (1991) conceptualization of brand associations (“anything linked in memory to a
brand”); they are “most commonly cited consumer brand perceptions in the empirical marketing literature”; “have reliable, established and published measures and “are discussed frequently in prior conceptual research”. Also, as Lafferty & Goldsmith (2004) point out, the
13 | P a g e brand attitude is an important construct, because it precedes consumers’ purchases and “thus
is a major concern of marketing management”. This implies that consumers` purchase
intentions are also of high importance, which is supported by Esch et al (2006), who discuss
that ensuring strong current and future purchases is the “ultimate goal of brands”. Moreover, the authors comment that brand knowledge influences consumers’ response to the brand, including perceptions, preferences and behaviours, i.e. consumers brand associations affect
their brand evaluations and purchase intentions. For these reasons, the construct of Purchase
intentions will also be incorporated in the research as a means to assessing brand
associations.
1.2 Problem definition:
The specific problem which will be in the core of this study is:
In terms of influencing customers` brand knowledge via inducing secondary brand associations, how do brand associations and evaluations (considered as brand attitude, perceived quality and purchase intentions) change when using Exclusive celebrity endorsement, as compared to Multiple celebrity endorsement.
Or in other words: Is “all publicity a good publicity” for the brand?
In order for the problem statement of this study to be resolved, the following sub-questions
need to be answered:
What is Brand knowledge and how does it relate to Brand associations What are secondary brand associations and how can they get evoked?
What are Celebrity endorsements; what is the difference between the effects of Exclusive vs. Multiple celebrity endorsement?
What is the difference between low and high level of celebrity-brand congruence; which is the most beneficial combination of endorser type and level of celebrity-brand fit when it comes to enhancing brand evaluations?
14 | P a g e
What effect does Source Credibility have on the evaluations of the brand-celebrity relationship?
1.3 Delimitations of the study
Although a wide range of topics will be addressed, there are some concepts which will not be
considered. As discussed in chapter 1.2, from the perspective of sources of secondary brand
associations, the focus will not be put on all sources (Companies, Country of origin,
Co-branding, Channel of distribution, Licensing, Events, 3rd party sources, Celebrity
endorsement), but rather only on the latter, i.e. Celebrity endorsement. However, future
research could address the inclusion of other sources to the same research context.
Furthermore, in order to ensure a realistic background of the research, the celebrities used in
the survey will be selected based on their true exclusive and multiple endorser statuses. In
addition, in the research set-up, those celebrities will be combined with fictitious brands, as
prior research suggests this approach to be more unbiased (Mowen & Brown, 1981; Misra &
Beatty 1990; Tripp & Jensen, 1994; Till & Busler 1998; Low & Lamb, 2000). Of course,
other stimuli would have potentially yielded different results, but this could be found out in a
follow-up future research.
In addition, the effect of exclusive vs. multiple endorsement could potentially be assessed in
different ways, however, as discussed, in the following study consumers` brand associations
will be assessed using Low & Lamb`s (2000) approach, focusing on the brand attitude,
perceived quality and purchase intentions, and not specifically on other constructs from the
literature (such as brand image, ad attitude, etc.). This is based on the aforementioned
argument of the authors that the former constructs are the best starting point for assessing
15 | P a g e
1.4 Theoretical contributions
The purpose of the study is to contribute to the body of research on celebrity endorsements.
First, the findings will enrich the existing knowledge about the effects of linking more
entities (i.e. multiple endorsers) to the same brand and how these linkages will affect
consumers’ brand evaluations. Thus, the study will address one of Keller`s five “Brand research imperatives” for the future (2000), which will contribute to the increase of the existing theoretical base. Importantly, the study will address the “gap” in the literature,
regarding the effectiveness of multiple celebrity endorsements, as compared with exclusive
endorsers.
Furthermore, following Um (2008), the concept of brand-endorser congruency will be
considered, but also from the uncommon perspective of a multiple endorsement. Thus, it
may be shown whether the current findings about congruency hold, with regard to brand
associations and evaluations, in the case of multiple endorsers. In addition, investigating and
comparing the effect of the different combinations of celebrity endorsers and varying level
of celebrity-brand fit will extend the existing pool of knowledge about celebrity
endorsement and will provide a new insight into the subject, which could also serve as a
basis for more detailed research in future.
Moreover, the Source Credibility model will be tested in a new setting, namely with
evaluating the effect of multiple endorser, in comparison with exclusive endorser, and
incorporating the congruency factor, thus making it possible for the findings to contribute to
the widely researched field of celebrities` attractiveness, trustworthiness and expertise.
In summary, if confirmed, the findings will demonstrate whether multiple endorsers are at
par with their counterparts – the exclusive ones, when it comes to leveraging brand
16 | P a g e
1.5 Managerial contributions
From a managerial perspective, the study would contribute to the extension of the current
managerial experience regarding the application of the branding strategy of celebrity
endorsement. Since the study aims at gaining new insights into how exclusive endorsements
are different than multiple ones in terms of brand attitude, purchase intentions and perceived
quality, companies would be able to choose the more appropriate branding strategy,
according to their marketing goals. This could further contribute to the increase of the
effectiveness of their marketing campaigns.
Looking from the perspective of a multiple endorsement, the study will try to answer the
question whether “all publicity is a good publicity” for the brand, i.e. is it always beneficial
for the brand to be linked to a famous public figure, or could there be cases when using a
celebrity spokesperson doesn`t necessary lead to more favourable brand attitude, perceived
quality and purchase intentions. If the second case holds true, there could be financial
implications linked to it. More specifically, companies could avoid investing considerable
amounts of money in an unsuitable branding strategy (i.e. endorser-brand combination) for
the corresponding marketing goal, and could allocate these resources in developing another
more beneficial one. This means that the research outcomes could be important prerequisite
for more successful future branding strategy.
Finally, depending on the findings, there is a potential for companies launching a brand
similar to the ones used in the study to consider contacting the corresponding celebrity
applied as an endorser in the research, or a similar one in terms of previous endorsements
and level of fit with the brand. Thus, companies will able to benefit from preliminary
17 | P a g e
1.6 Research outline
Having celebrity endorsement in its core, the study focuses first on examining the current
literature for theories underlying the concept, as well as explanations of the endorsement
strategy, the concept of congruence and source credibility, and provides a summary of the
relevant findings. Based on the available knowledge base and on the identified ‘gaps’, the conceptual framework is created and hypotheses are formulated. Next the methodology is
described, including research design, pre-test, stimuli and manipulations, procedure and
measurements. After obtaining the data, it has been prepared for analysis and the hypotheses
are tested, followed by a discussion of the results, theoretical and managerial implications.
Finally, a conclusion is summarized, limitations of the study are mentioned and suggestions
for further research are proposed.
2. Literature review
2.1 Brand knowledge and customer-based brand equity
One of the main priorities of brand management nowadays is building strong brands (Low &
Lamb, 2000). Strong brands are quite advantageous in terms of high revenues, both in
short-term and long-short-term aspects (Keller 2012; Kapferer, 2004). For this reason, strategic brand
management has in its core the building of such brands that endure in time and can be
leveraged in various market, as well as product categories (Aaker, 1996).
Keller’s (2012) customer-based brand equity model (CBBE), considered as one of the most widely used and so far “most parsimonious” models (Esch et al, 2006), sheds light on the ways to building a strong brand. Namely, after gaining a positive experience with a certain
product/service and its marketing, it should be ensured that consumers will link “what resides
18 | P a g e (Keller, 2012), thus forming specific brand knowledge. This brand knowledge will contribute
to differentiating the brand when consumers react to the brand marketing. As a result from
consumers’ familiarity of the brand and their positive brand associations stored in memory, a
CBBE will occur (Esch et al., 2006), ultimately leading to the establishment of a strong brand
(Keller, 2012). Thus, the important role that brand marketing (i.e. advertising, promotional
deals, other entities linked to the brand – like celebrity endorsers, etc.) plays in the process is
quite notable, as it appears to be a prerequisite for the formation of strong brands.
2.2 Dimensions of Brand knowledge and Brand associations
In order to exemplify how brand knowledge is situated in consumer`s memory, the brand
knowledge concept can be analysed parallel to the Associative network model (Keller, 2012).
According to the model, memory is represented as a bundle of nodes and links, where nodes
are considered stored information of any kind and links express the association strength
between the nodes (Anderson, 1989). Following the Associative network model, Keller
(2012) conceptualizes brand knowledge as a unifying entity of a brand node in memory and a
wide range of associations attached to it. Krishnan (1996) discussed the importance of the
number of brand associations. Over time, consumers form a considerable set of associations
about a wide range of brands. Consistent with the Associative network model, the higher the
number of associations about a brand is, the easier it is to access the particular brand node in
memory, because of the multiple linkages leading to this node. Therefore, from CBBE
perspective, a large number of brand associations are desirable. Based on the available set of
association, consumers form brand attitudes. Brand attitudes are considered consumers`
overall (positive or negative) evaluations of a brand (Mitchell and Olson, 1981). Brand
attitudes are of high importance, because they often serve as a foundation for behaviours that
consumers can perform towards the brand (e.g. purchasing). According to Keller (2012), the
19 | P a g e of the brand. It is defined as an overall evaluation of a brand's excellent or superior
performance, regarding its purpose, as compared with alternative brands (Dean, 1999).
This conceptualization demonstrates how intertwined brand attitude, purchase intentions and
perceived quality are, when it comes to assessing brand associations and evaluation, which
the following study is focused on.
2.3 Secondary brand knowledge and associations and their sources
As discussed, brand knowledge can be
formed in many ways, e.g. via marketing or
non-marketing actions (Keller, 2003). In
general, any exposure to the brand can alter
the existing brand status in the memory of
consumers. Sometimes, there may be lack or
insufficiency of brand associations in
consumer memory. In this case, marketers
can link the brand to another entity that has
its own brand knowledge in consumers’ minds. The purpose of this linkage is the creation of secondary brand knowledge. This can be achieved with forming secondary associations
about the brand by means of making the consumers infer that what they already know about
the second entity should be also valid for the brand, because of the relationship between
them. In this way, brand associations are transferred or “borrowed” from the other entity,
ultimately resulting in either altering existing brand knowledge (e.g. through strengthening
brand associations) or in building new brand knowledge (e.g. through adding new
associations) in consumers’ minds. (Keller, 2012).Thus, in effect, brand equity is leveraged indirectly from the other entity to the brand.
20 | P a g e Several entities can be identified as most common sources of secondary brand knowledge,
Endorsers being one of them (Figure 1 Secondary sources of brand knowledge, Keller, 2003).
Secondary associations are mainly created through primary associations relating to these
sources. Following the discussion in chapter 1.3, as Celebrity endorsement is in the core of
the study, no further attention will be paid to the rest of the sources of Secondary brand
associations.
2.4.1 Celebrity endorsement: Characteristics, Advantages & Disadvantages
Product endorsement is not a new phenomenon (Erdogan, 1999). Celebrities have been
endorsing products since the XIX century. Early examples of this practice are: one of the
lines of Cadbury’s chocolates, featuring the images of Queen Victoria and Prince Albert (Sherman 1985; www.cadbury.co.uk ); Pope Leo XIII and Thomas Edison endorsing “Vin
Mariani”, a potent mix of Bordeaux wine and cocaine (Feldman, 2007); actress Sarah
Bernhard appearing in 1890 on posters for La Diaphane, a famous French rice powder brand
(Lehu, 1993). Interestingly, even former U.S. president Ronald Reagan was a celebrity
endorser during his career as an actor, pitching for several products, including cigarettes
(Keller, 2012). In recent times, the use of celebrities in advertising has become increasingly
widespread, with every fourth commercial featuring a celebrity (Shimp, 2003), compared to
the late 1970s, when famous spokespersons were showing up in 17% of advertisements
(Howard, 1979). What underlines this increase is the fact that celebrities are highly successful
advertising tool for creating and enhancing brand image and equity (Choi & Rifon, 2012).
McCracken (1989) defines a celebrity endorser 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”. Keller (2012) elaborates further, that the mission of using a famous person is drawing attention to a brand and shaping the perceptions of the brand. This can be
21 | P a g e achieved by means of the inferences that consumers make about the brand, based on their
existing knowledge about the famous person.
Research has demonstrated that there are always benefits and harms associated with the
endorsement strategy. In the celebrity context, multiple studies have shown that using
celebrity endorsement as a means to support the brand strategy can be quite advantageous.
Erdogan (1999) found out that such benefits can include increased attention, image polishing,
brand introduction, brand repositioning and the possibility to underpin global strategies. In
particular, the influence of celebrities could be found in their potential to “attract attention to
the ad in the cluttered stream of messages”. Moreover, celebrities are traditionally perceived as being highly dynamic persons, possessing attractive and likable personal qualities (Atkin
and Block, 1983). Furthermore, celebrity endorsement has the potential to get a higher
memorability for the advertisement in which it is applied and may cause a positive affect that
could be generalized to the brand (Cooper, 1984; Plapler, 1974). According Pringle (2004),
there are also financial benefits to be gained from successful brand-celebrity partnerships.
Additionally, Friedman and Friedman (1979) have found out that celebrities are more
effective than other types of endorsers, such as “the professional expert”, “the company manager” or “the typical consumer”. In general, according Erdogan (1999), “it is safe to argue that celebrity endorsers are more effective than non-celebrity endorsers in generating
all desirable outcomes (attitudes towards the advertising and endorsed brand, intentions to
purchase, and in fact actual sales)”, under certain conditions.
Nonetheless, celebrity endorsement strategy can be a “two-edged sword” (Erdogan, 1999), because apart from the aforementioned benefits, the strategy could entail potential hazards,
such as overshadowing the brand, public controversy, image change, over exposure, loss of
public recognition, high costs. Moreover, Till & Shimp (1995) found that negative
22 | P a g e the celebrity, but also towards the endorsed product. Also, sometimes brands can enjoy
higher consumer awareness and liking than the celebrities who endorse them (Tellis, 2004),
which in this case would render the endorsement unnecessary. Furthermore, due to
overexposure, consumers may get the impression that celebrities don’t believe in the product they endorse, but do it merely for financial reasons (Keller, 2012).
Therefore, considering the risks from selection of inappropriate celebrities and the significant
financial investments in this advertising strategy, it is of high importance for advertisers to
understand consumer responses to celebrity endorsement and develop an effective approach
for proper endorser selection. (Choi & Rifon, 2012). Although endorsement effectiveness
also depends on factors such as ad execution quality, media selection, etc., it is highly
dependent on the appropriate choice of celebrity, which can be explained by the following
models (Fleck et al., 2012).
2.4.2. Models explaining the Celebrity endorsement process
Several models have been developed to explain what underlies the celebrity choice and the
process of celebrity endorsement as a whole. First, the Source Attractiveness model
(McGuire, 1985) suggests that the effectiveness of the message depends on physical characteristics of the source, which make the source attractive (Chao et al, 2005; Till & Busler, 2000). Previous research suggests that the physical attractiveness of a celebrity endorser positively impacts brand attitude, brand recall and purchase intentions (Joseph, 1982; Kahle & Homer, 1985) Second, the Source Credibility model (Hovland & Weiss,
1951) describes the effectiveness of the message as depending on the degree of expertise and trustworthiness of the source (celebrity). In line with the model of this study, past findings demonstrate that the effectiveness of celebrity endorsement can be mediated by characteristics from both Source Models, as well as by various variables, such as product
23 | P a g e
type, number of endorsements, characteristics of target consumers, and overall meanings attached to celebrities. Theses ‘meanings’ are conceptualized by McCracken in his Meaning Transfer Model (1989) following three steps: formation of celebrity image and meaning, transfer of these meaning and associations first from the celebrity to the endorsed product and then from the product to the consumers. This flow explains how celebrities transmit a vast set of associations to the brands they endorse, which is in line with the Associative Network model and CBBE. Therefore, before companies choose a well-known person to pitch for their brands, they need to make sure that this endorser conveys the right meaning (Fleck et al., 2012). And it is exactly this congruence between spokesperson and brand that facilitates meaning transfer from one to the other (Misra & Beatty, 1990; Lynch & Schuler, 1994; Kamins & Gupta, 1994).
2.4.3 Multiple endorsement
Celebrities not only sell products and promote brands, but in fact “exert powerful influence across all facets of popular culture and public life”, also shaping consumer perceptions and attitudes (Time Inc., 2006). As such, celebrities are increasingly engaged as spokespersons
for a range of brands in different product categories. For example, Tiger Woods endorses
Nike, American Express, Tag-Heuer, Buick, Accenture, etc. (Um, 2008).
The discussion so far has been referring to the practice of single endorsement – “when the spokesperson endorses one brand and no other brands or products” (Tripp et al., 1994), also
referred to as “uniqueness” (Keller, 2003). Since overtime celebrities are expected to pitch for more than one product, single celebrity endorsement is rare (Um, 2008). Indeed, a range of celebrities are “shared” by different advertising firms, i.e. they endorse more than one brand (Tripp et al., 1994). An example is Peyton Manning, who in 2009 endorsed eight
brands: Mastercard, Nerf, Oreo, Wheaties, Reebok, Gatorade, Sony and DirectTV
24 | P a g e and Lindt chocolates ((Krosnick & Alwin, 1987). This approach is defined as multiple
endorsement (Tripp et al., 1994; Mowen & Brown, 1981). As suggested by advertising
professionals (Foote & Belding, 1978), endorser selection should be made with precision, in
order to be most appropriate. Similarly to research findings discussed below, the industry
practice of ensuring exclusivity clauses by offering higher financial rewards for exclusive
endorsers (ReedSmith, 2006) also suggests that an exclusive endorser is indeed considered
more beneficial to a brand compared with an endorser who is already tied to multiple brands.
However, in the literature, limited research explores celebrity effectiveness depending on the
number of brands endorsed.
In the light of the Attribution theory (Kelly, 1967), Mowen & Brown (1981) discussed
whether the endorsement of multiple products tarnish a celebrity`s effectiveness. The authors
found out that a single endorsement resulted in more favourable product attitude, increased
purchase interest, and more positive response to the message of the ad, because engaging
with multiple brands indeed tarnished the endorser`s image.
Furthermore, Tripp, Jensen and Carlson (1994) manipulated the number of endorsers and the
number of exposures to the endorsements, utilising the Source Credibility model. The authors
found out that the celebrity is perceived as more trustworthy and as a better expert when
endorsing only one brand, as compared with several ones. Moreover, with the increase in the
number of endorsed brands, people were reacting less favourably toward the ad. Interestingly,
in contrast to Mowen and Brown’s (1981) findings, that the number of endorsed products strongly influences consumers’ brand evaluation and purchase interest, the study of Tripp et
al. (1994) did not find evidence to support this. According to the latter authors, this
divergence may be due to the different nature of the products endorsed. Whereas the former
use a fictitious brand, the latter test existing well-known brands, whose evaluations were
25 | P a g e multiple endorsements. This implies that further research is needed, as the current findings
are inconclusive.
2.4.4 Models explaining Multiple endorsement
As any other endorsement, the multiple celebrity endorsement might not always lead to the
desired results. In the psychology literature, a phenomenon called blocking is discussed,
which could be used for explanation of possible low effectiveness of multiple celebrity
endorsement. If a particular cue has established a strong association with occurrence of an
outcome, this association will not allow for additional cues to develop strong associations
with the same outcome. Thus, the first cue blocks the successful attachment of new cues
(Oxoby & Finnigan, 2007; Osselaer et al., 2001). According to Oxoby & Finnigan (2007),
blocking happens often when people act in multi attribute environment, being confronted by
several observable cues. Therefore, blocking may arise in consumers’ attempts to detect the particular outcome.
In the context of a multiple endorsement, several brands can be considered the cues, and the
endorser can be considered the outcome. Presumably, there will be a strong association
between the initial brand and the endorser, so subsequent brands “attached” to the same endorser would potentially cause blocking. (Kruschke, 1996, 2001).This implies that the
more brands a celebrity advocates, the higher the blocking effect would be, because the
associative link between the spokesperson and each following brand will be even lower. On
the other hand, if there is an exclusive relationship between a celebrity and a brand, no other
brands (cues) will be presented to consumers, leading to avoidance of blocking effects. Thus,
one can assume that the linkages between an endorser and more brands will become less
strong as the number of endorsed brands increases. This was also verified by Krosnick &
Alwin (1987), who posit that the brand endorsed as first is likely to be subjected to deeper
26 | P a g e The same blocking effect can be explained also on the basis of the Classical Conditioning
principle. Classical conditioning (CC) is a type of an associative learning process in which a
conditioned stimulus (CS) that had not previously produced a particular response becomes
associated with an unconditioned stimulus (US) that produces the response. As a result, the
CS will elicit the response that the US produces (Pavlov, 1927). In this context, the findings
in the area of associative learning may explain the positive effects of celebrity endorsement,
such as improved consumer perceptions of the brand derived from the pairing of a product
and a well-regarded celebrity figure (Kamins & Gupta (1994).
In principle, the endorser can be considered the US, which after pairing with a brand (CS)
evokes the desired conditioned response, such as brand-beneficial attitude and behavior.
Although many advertisements are created with the final goal of achieving CC (Nord &
Peter, 1980), multiple celebrity endorsement can be considered one of the situations in which
CC fails to occur (McSweeny & Bierley, 1984). Traditional CC postulates that conditioning
will occur whenever a CS is followed by a US, but CC will not occur if blocking occurs
(Kamin, 1969). Contrary to the CC, the case of blocking considers pairing the US with more
than one CS, where the desired conditioned response is elicited only to the first (oldest) CS,
which ultimately leads to unsuccessful development of further associative links to all other
CSs. Thus, the prior experience with the first CS blocks conditioning to the next CSs
(McSweeny & Bierley, 1984). Applied to the multiple celebrity endorsement context, this
means that consumers` conditioning will be most successful to the first brand (CS) that the
celebrity (US) endorses and their established experience with the particular celebrity-brand
pairing (CC) will block further associative links to subsequently endorsed brands (other CSs).
Therefore, the blocking effect could be an explanation of possible occurance of unfavourable
27 | P a g e celebrity endorsement might not be the most beneficial strategy for achieving desired brand
attitudes.
2.4.5 The concept of congruence2
In general, consumers consider celebrities being dynamic, likable and appealing, and it is
these characteristics that attract their attention to the endorsed products (Atkin & Block,
1983). Notably, choosing the right celebrity has turned into a task of high importance for
marketers and practitioners. Therefore, in the marketing literature a big body of research has
been devoted to determining those characteristics of celebrity endorsers, which will enhance
marketing communications or will facilitate the achievement of additional advertising goals
(Lynch and Schuler, 1994). A considerable attention in current research has been paid to the
concept of congruence between the spokesperson and the advertised product or brand. Based
on the literature, the overall conclusion that the level of perceived fit between both parties has
a strong impact on consumers` brand evaluations (Kamins, 1989, 1990; Misra & Beatty,
1990; Kamins & Gupta, 1994; Till and Busler, 2000), thus pointing out to the importance of the “matchup” factor. According to Khale & Homer (1985), advertising effectiveness is driven by the match between “the message conveyed by the image of the celebrity and the message about the product”. Misra & Beatty (1990) further elaborate that match-up or congruence (i.e. or fit) refers to such a celebrity-brand combination, where “highly relevant characteristics of the spokesperson are consistent with the highly relevant attributes of the brand”. In order to provide clarity on the concept of fit, an overview of the existing literature findings concerning the matchup concept is presented below.
According to Kanungo & Pang (1973), there should be “fittingness” between the spokesperson and the advertised product. In their two experiments, the authors examined
2 A variety of terms have been applied to refer to the concept (congruence, fit, link, match-up effect), therefore
28 | P a g e
different combinations of products, (perceived as masculine, feminine, equally masculine and feminine, and neither masculine, nor feminine) and their corresponding male or female endorser. After combining the results, it was evident that when the gender of the spokesperson matched the perceived femininity or masculinity of the products, the effectiveness of the pitcher was higher, resulting in more favourable product evaluations. This can be explained by means of the “perceptual and attitudinal congruence” stemming from such fitting product-endorser combinations and because a congruous experience is psychologically comfortable for the individual (Zajong, 1990), it leads to more positive product attitudes. Hence, the authors concluded that a fit is present, when the image of the product, in terms of its characteristics and associations, is consistent with the type of the endorser utilized in the ad. Thus, support was found for the fittingness hypothesis. In addition, regarding the gender type of the endorser and the perceived product image, Caballero & Solomon (1984) found further support for the concept of fit by showing that when a male spokesperson was matched with a product having a “predominantly male image” (beer), product sales increased. On the other hand, Debevec and Iyer (1986) researched the effects of mismatching spokespersons and product types. In comparison to Kanungo & Pang (1973), the authors found the contrary result, namely that the female endorser evoked more favourable attitudes and buying intentions for the masculine product than did the male endorser and vice versa. However, it should be noted, the difference in the outcomes may be due to the long time span (13 years) separating both studies, which could explain the change in gender-role perceptions and cultural stereotypes (Lynch & Schuler, 1994). What is notable is Debevec & Iyer’s (1986) suggestion that intentional mismatching of spokesperson’s gender and endorsed product can be used for changing the product image, which further implies that the product image could be altered also by means of unintentional mismatch. Hence, marketers need to be alert when selecting a spokesperson to pitch their products.
29 | P a g e
Furthermore, a widely researched topic in the literature has been the fit between a product and an endorser, based on his/her characteristics, such as physical appearance and attractiveness. Peterson & Kerin (1977) found support for the congruence hypothesis by showing that the pairing of a seductive spokesperson and a body oil (i.e. sensual product), as compared with the other treatment conditions, resulted in more positive ad evaluations and increased manufacturer reputation. Thus, the authors suggested that the fit was based namely on the physical appearance of the model combined with the sensual nature of the product, which matched together. Furthermore, partial support for the fit hypothesis was provided by Baker & Churchill (1977), who manipulated spokespersons’ attractiveness, gender and product type (romantic vs. unromantic) and found a significant interaction only between the female attractive endorser and romantic product with regard to the behavioural intentions. However, the authors discussed that the outcome could be due to the stimuli execution, thus still supporting the concept of fit.
Moreover, an additional research direction has been exploring the effects of manipulating and matching endorser types and product types. Friedman & Friedman (1979) utilized three spokesperson types (celebrity, endorser and typical consumer) and three product categories (costume jewellery, vacuum cleaner and cookies) differing in the perceived risks regarding their purchase (psychological and social risk, and financial, performance and physical risk). The concept of fit was once again confirmed, when the authors found out that most positive overall ad evaluations and attitudes towards the products, as well as the strongest purchase intentions and the highest endorser’s credibility were achieved when, as expected, the celebrity was matched with the costume jewellery, the expert was paired with the vacuum cleaner and the consumer was combined with the cookies. Thus, the authors demonstrated that the effectiveness of fit can be contributed not only to endorsers’ physical characteristics, but also to their social or cultural role. Additionally, a range of authors also utilized the
30 | P a g e
celebrity endorser as a type of spokesperson in their research, putting an emphasis also on the attractiveness concept. As a foundation of the so called matchup hypothesis, Kahle & Homer (1985) and Kamins (1990) suggested that the image of the celebrity and the product message and image should be congruent (i.e. fit together). Incorporating the celebrity attractiveness in their research, the authors suggested that attractive celebrity spokespersons should be more effective when pitching products whose main purpose is namely enhancing one’s attractiveness. In line with this reasoning, Kahle & Homer (1985) demonstrated that brand attitude was more positive and purchase intentions were stronger when the attractive celebrity was paired with an attractiveness-related product (razor blade). On the other hand, they couldn`t show that the attractive celebrity is less effective when endorsing a product not stimulating attractiveness, because they didn`t include such type of product. In order to complete Kahle & Homer’s (1985) experiment, Kamins (1990) performed a study, incorporating both attractiveness related and non-related products (luxury car and home computer), as well as both attractive and unattractive endorsers (Tom Selleck and Telly Savalas). Consistent with Friedman & Friedman’s findings (1979), Kamins (1990) demonstrated that endorsers’ characteristics interact with the nature of the endorsed products. Specifically, the author showed that the combination of Tom Selleck & the luxury car (i.e. attractive endorser and attractiveness-enhancing product) led to a significant increase in positive evaluations of celebrity`s credibility and attitude towards the ad, as opposed to the contrary celebrity-product combinations, thus making the author conclude that “the use of an attractive celebrity spokesperson appears to be effective for a particular category of products, those which are attractiveness related”. However, Kamins (1990) failed to support the expected matchup interaction for brand attitude and purchase intentions. This was done by Till & Busler (2000), who showed that brand attitude and purchase intentions were indeed increased when the attractive endorser was paired with an attractiveness-enhancing product
31 | P a g e
(cologne), but also demonstrated that a stronger matchup effect could be found based on celebrity`s perceived expertise, rather than physical characteristics (i.e. attractiveness). Similarly, regarding brand attitude, Misra & Beatty (1990) observed a transfer of affect, only when the brand was congruent with the celebrity endorser (Unitough jeans with Clint Eastwood, but not with Carol Burnett), as opposed to the non-congruence condition (Funnybone with Carol Burnett, but not with Clint Eastwood), resulting in more favourable brand attitude. Likewise, product attitude was found by Kamins & Gupta (1994) to increase in favourability when spokesperson-product congruence was correspondingly higher. In addition, the authors extended Kahle & Homer’s (1985) findings with regard to exploring the matchup effect from attractiveness perspective, by showing that the higher product-celebrity fit also enhanced endorser’s believability. Thus Kamins and Gupta (1994) suggested that the issue of the matchup hypothesis can be referred to as more “global”.
Moreover, an additional perspective of the level of fit can be derived by looking at the celebrity-brand matchup from the perspective of a co-branding partnership. This suggestion is based on the reasoning that due to the fact that celebrity endorsers can be managed in a
professional manner, as well as that they possess various associations and features of a brand,
one could argue that celebrities can be actually considered brands (Thomson, 2006). Thus,
looked from this point of view a celebrity endorsement of a brand can be regarded a
co-branding partnership. Importantly, previous studies empirically confirmed this particular
assumption. It was shown that celebrity endorsements and brand alliances were both based on
the same processes, namely a meaning transfer process (i.e. in line with McCracken, 1989)
and associative learning (Halonen-Knight & Hurmerinta, 2010). Following, support was
found that celebrities could be considered co-brands (Halonen-Knight & Hurmerinta, 2010;
32 | P a g e to which the two brands (i.e. celebrity and endorsed brand) are “complementary” and “consistent” with each other (Simonin & Ruth, 1998).
In summary, the findings in the existing literature demonstrate that there are various bases of celebrity-product fit, such as perceived characteristics of the spokesperson (e.g. gender) and those of the endorsed product (e.g. perceived femininity/masculinity), the social and/or cultural role of the spokesperson (i.e. spokesperson type, such as celebrity, endorser or consumer), physical characteristics of the endorser (e.g. attractiveness) and product type (attractiveness-enhancing and non-enhancing product), as well as other characteristics like endroser’s credibility (Ohanian, 1991), (e.g. celebrity’s expertise and trustworthiness). Following the discussed positive effects that can be derived from a successful matchup, such as more favourable product and ad evaluations, higher product sales, increased manufacturer reputation, stronger purchase intentions, more positive brand attitude and higher brand recall, one general conclusion could be drawn, namely that the endorser-product fit is a very important factor for achieving advertising effectiveness (Amos et al., 2008). Finally, in order to exemplify the importance of a proper matchup, it is worth quoting the senior vice-president of Seven-Up managing the marketing activity of the company claiming that “celebrities are an unnecessary risk unless they are very logically related to the products” (Watkins, 1989).
2.4.6 Models explaining the concept of Congruence
In the years of research regarding the phenomenon of congruence, a number of mechanisms
such as the Elaboration Likelihood Model (Petty et al. 1983), associative learning (Till and
Schimp, 1998), social adaptation theory (Kahle and Homer, 1985), schema theory (Lynch and Schuler, 1994), attribution theory (Mowen and Brown, 1981; Tripp et al, 1994), cultural meaning transfer (McCracken, 1989), and source characteristics (Ohanian, 1991) have been applied in order to explain the effects of a celebrity-product fit and hence, the effectiveness of the celebrity endorsement (Till and Busler, 2000). As most relevant to the purposes of the
33 | P a g e
present study and in line with the previously discussed models of Associative network and Meaning transfer, the concepts of Associative learning will be explored below and will be applied to the celebrity endorsement context.
Furthermore, the celebrity-product fit can be explained by means of the associative learning theory. According Klein (1991), associative learning is a mechanism that facilitates the establishment of relationships and links between concepts. As previously mentioned, an associative network consists of nodes in memory linked together, storing particular pieces of information (Anderson, 1989). In the context of a celebrity endorsement, both the brand and the celebrity represent units in memory which, based on the individual`s personal experiences with them, are connected to other units and these connections represent the association set of both parties. Importantly, it is the fit that determines how easily an associative link between two nodes will be established (Till & Busler, 2000). In general, the more similar two concepts are, the easier and the more likely these concepts are to become linked to each other and thus connected in a common associative network (Garcia & Koelling, 1966). Hence, a proper matchup between a celebrity and a product will allow for a stronger associative link between them, leading to an increase in the celebrity endorsement effectiveness.
3. Hypotheses
In this section the relationships lying in the core of the study are being presented and a set of
hypotheses is being established. Following the detailed overview of the theoretic findings, the
present study will address Endorser type (exclusive vs. multiple), Celebrity-brand fit, Source
credibility Evaluations of brand associations (Brand attitude, Purchase intentions, Perceived
quality). Figure 2 exemplifies the conceptual framework of the study, which serves as a
visualization of the relationships that will be tested, as well as shows the formulated
34 | P a g e Figure 2 Conceptual Framework
The focus on these particular elements will allow for the answering of the central research
questions of the study, namely in what ways is the Exclusive endorser different than the
Multiple one with regards to consumers` brand associations and evaluations, as well as how
the level of Celebrity-brand congruence affects them, depending on the endorser`s type. More
specifically, the study tries to identify which of the four endorser-fit conditions (Exclusive
endorser*High fit; Exclusive endorser*Low fit; Multiple endorser*High fit; and Multiple endorser*Low fit) are most beneficial for inducing more favourable brand evaluations.
In line with the Associative learning theory discussed above (Klein, 1991), one of the most
comprehensive models for explaining celebrity endorsement effectiveness is the Source
credibility model (Erdogan, 1999), categorized by the level of trustworthiness, expertise and
attractiveness of the endorser. The more honest and believable, and knowledgeable and experienced, the more trustworthy and expert the endorser will be in consumers’ perceptions. (Erdogan,1999). Tripp, Jensen & Carlson (1994) demonstrated empirically that the celebrity
is perceived as more trustworthy and as a better expert namely when endorsing only one
brand, in comparison with several ones. Therefore the following is hypothesised:
35 | P a g e
Multiple authors have experimented with the level of endorser’s credibility and observed that it can vary not only depending on the type of endorser, but also due to the nature of the endorsed product or brand, and more specifically, depending on how congruent the endorser and the product/brand are. For example, Friedman & Friedman (1979) showed that attractiveness related product combined with an attractive endorser resulted in an increase of overall brand responses and attitudes towards the endorser’s characteristics (part of the Source credibility model). Till & Busler (2000) also demonstrated that a strong matchup effect could be found based on celebrity`s perceived expertise and the endorsed product, which means that the higher product-celebrity fit, the more enhanced endorser’s believability. In other words, Source Credibility is expected to be higher not only in the case of Exclusive endorsement, but also when there is a higher level of fit. Also, even in the case of Multiple endorsement, the high level of fit is expected to be more beneficial, as compared with low level of fit, and is therefore expected to lead to higher Credibility too (Tripp, Jensen and
Carlson, 1994). This can be explained by means of the Filtering model, which suggests that
any brand information which is incongruent with the existing celebrity’s characteristics, will be filtered out, thus allowing the congruent information to be superiorly encoded and more favourably evaluated (Misra & Beatty, 1990). Therefore, based on the existing knowledge of the importance of fit, the following hypotheses are proposed:
H2: Source credibility will be higher in the case of high fit vs. low fit.
The next set of hypothesis has not yet been addressed in the existing branding literature, even though it directly concerns answering the main research questions, more specifically – which endorser-fit combination is of biggest importance for achieving highest brand evaluations. So far, following the academic research, this study has proposed that in terms of increasing endorser`s credibility, the effect of the exclusive endorser will be stronger than the effect of the multiple endorser, as well as that the effect of the high fit will be stronger than the effect