The effect of celebrity endorsement on customer-based brand equity concerning fast moving consumer goods : an experimental investigation

(1)

The Effect of Celebrity Endorsement on Customer-based Brand

Equity concerning Fast Moving Consumer Goods

An Experimental Investigation

Master Thesis

Student: Anil Ünlü Student number: 11387254

E-Mail: uenlue.anil@gmail.com

Study program: MSc. in Business Administration – Track Digital Business Supervisor: Dhr. Kristopher O. Keller MSc

Submission Date: 23.06.2017

(2)

Statement of originality

This document is written by Anil Ünlü who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is 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 is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

Advertising is highest in the FMCG industry, with recipients flooded with numerous ads every day (Nielsen, 2016). To get access to consumer mind-set, celebrity endorsement is one marketing investment that especially FMCG manufacturers make use of, not only to increase awareness, but to elicit more positive brand evaluation with consumers (Millward Brown, 2016, Till, 1998). But do celebrities help in achieving higher customer-based brand equity (CBBE) considering different FMCG categories, and under which level of involvement do celebrities work best? By means of a 2 × 3 between-subjects experimental design, the present research examines the effects of celebrity endorsements on CBBE of three FMCG. Using Aaker’s brand equity ten (1996), the author found that ads featuring the celebrity lead to more favourable outcomes in terms of CBBE. In addition, the moderating effect of low, moderate and high involvement is studied. Under low and moderate involvement, there is an interaction effect between celebrity endorsement and involvement, so that the positive effect of celebrity endorsement is higher under moderate involvement in comparison to low and high-involvement FMCG. This leads to novel insights in form of an inverted U-shape. Implications are discussed for academic researchers who predict the impact of celebrity endorsement and product involvement on CBBE and for marketers who intend to translate this investment into brand equity success.

Keywords: Celebrity endorsement, customer-based brand equity, involvement, Elaboration Likelihood Model, FMCG.

(4)

1 Introduction

he consumer goods industry exhibits continuous growth in revenue due to a boost in consumption and rise of the middle class through large parts of the world, and is expected to increase globally by 43% between 2010 and 2020 (PWC 2016, Thain & Bradley, 2014). The fast-changing and evolving fast moving consumer goods (FMCG) industry, which includes four major types of products: food and beverage, home and personal care, tobacco, and alcoholic drinks, is particularly characterized by increased competition among manufacturing and private-label brands, leaving room for only the strongest brands (Thain & Bradley, 2014, Bain 2015).

In order to differentiate themselves from direct competitors, companies producing FMCG aim to achieve high brand equity from a consumer’s point of view, known as customer-based brand equity (CBBE) (Yoo, Donthu & Lee, 2000). A high CBBE indicates that consumers recognize and recall the brand; show strong, favorable and unique associations; perceive the brand as higher quality; and thus show more loyalty (Aaker, 1991, Keller, 1993).

However, due to high competition among companies, advertising in the FMCG industry is at its highest with a share of 20.7% of all industries (Nielsen, 2016), with consumers flooded with numerous ads through multiple channels every day (Thorson, 1990). Therefore, companies struggle in receiving satisfactory attention from consumers. To gain access to the customer mind-set, celebrity endorsement has been emphasized by previous articles as one marketing investment that companies can employ (Keller & Lehmann, 2003). Celebrity endorsement can serve as an external cue, and if well-chosen, ads featuring celebrities are able to draw attention to a product or brand more easily than ads without celebrities; marketers hope that the right celebrity can lend his or her image to their brand, so that people who consume celebrity images will also consume products associated with them (McCracken, 1989, Erdogan, 1991, Forbes 2016, Kotler & Keller, 2012).

(6)

Introduction

Using celebrities to endorse products or services is prevalent in advertising, done to elicit more positive brand evaluations with consumers (Kamins & Gupta, 1994, Till, 1998), and indeed used in advertising around the world (Millward Brown, 2016). While 15% of all advertisements in the U.S. involve celebrities, the number in Asian markets such as India or Taiwan is even higher with 24% and 45% respectively (Crutchfield, 2010). Since consumers are more likely to choose goods endorsed by celebrities (Dyson & Turco, 1998), companies utilize celebrity endorsement in 25% of all TV- commercials (Erdogan et al., 2001) and invest large amounts of money to align their brands with endorsers. In 2006, for example, U.S. retailers spent up to three billion dollars on celebrity advertising (White, Goddard & Wilbur, 2009). FMCG manufacturers are enterprises which use this type of advertising more frequently and most commonly promote product categories such as personal care, food and beverages, and hair care (Millward brown, 2016).

As the cost involved is very high, the choice of a credible celebrity is crucial in terms of sales and return on investment (Erdogan, 1999). Thus, the literature recommends endorsers that show high (1) credibility and attractiveness (Hovland et al. 1953, Kotler & Keller, 2012, Ohanian, 1990), (2) celebrity-product congruence (Kamins 1990), and (3) ability to transfer a set of meanings towards the endorsed brand (McCracken, 1989). To reinforce its high-status image, Spar Austria for example, has used James Bond legend Pierce Brosnan in its ad for its own luxury brand label Spar Premium (Spar international, 2017). Nespresso, in turn, employs George Clooney, who is consistent and in line with the image Nestlé wants to convey for its product: “dark, rich, refined and mature” (Tabet et al., 2013). Pierce Brosnan and George Clooney are examples of well-executed celebrity endorsements, but there are also celebrities who may transfer unfavorable associations towards a brand. Celebrities like Donald Trump might have high recognizability but elicit a negative effect among many groups (Kotler & Keller, 2012). Examples of incidents that led to failed endorsements include Tiger Woods’ extra-marital affairs and Kate Moss’ drug incident. Initially considered as optimal endorser,

(7)

Introduction

misbehavior cost Tiger Woods endorsement deals with Accenture, Gillette, and Gatorade (Kanter, 2009).

Past research has underpinned that endorsement effectiveness is moderated by level of involvement (Petty et al., 1983), which varies between product categories in the FMCG industry. These results were interpreted as follows: under conditions of high involvement, true product qualities rather than celebrities influenced attitudes, whereas under conditions of low involvement, celebrities but not arguments influenced attitudes (Petty et al., 1983). Multiple studies in the past concluded that celebrities can lead to favorable attitudes towards the endorsed brand and advertising (Petty et al. 1983, Till et al., 2008). This makes them more effective than non-celebrity endorsers in generating all desirable outcomes (Erdogan, 1999). Marketing communications can be considered a tool in building CBBE (Keller, 2005, Yoo et al., 2000), but Spry (2009) questions how celebrity endorsement can lead to improvements in brand equity. Previous marketing research addressed the impact of celebrity endorsement on CBBE in an insufficient manner and did not take the impact on different FMCG product categories into account.

Until now, there has been limited understanding of how the level of involvement in endorsed advertising campaigns might moderate the endorsed brand’s equity. Particularly in product categories where few differences are perceived among brands, celebrities might help to differentiate between the brands at an emotional level (Callcoat & Phillips, 1996, Das & Padhy, 2012)

Despite the importance of brand equity in marketing theory and practice, there is a significant gap in the literature regarding the impact of celebrity advertising influencing the CBBE of low-, moderate-, and high-involvement FMCG product categories. Therefore, it is of great importance to investigate under which scenarios it is useful to use celebrities as an advertisement tool to drive CBBE. This allows obtaining useful insights into the concept of CBBE and the various ways that marketing communications can build brand equity (Keller,

(8)

Introduction

2009). By means of quantitative research and a 2 × 3 between-subjects experimental design, the aim is to discover whether brands are better able to reach their objective of garnering high CBBE when using celebrity-endorsed advertising, under conditions of low, moderate, or high involvement. Since the FMCG industry is characterized by low-, moderate-, and high-involvement products, the study will investigate all scenarios.

Each driver of CBBE can be analyzed by using certain questions in the form of a questionnaire. This method enables the estimation of brand equity for multi-category brands (Christodoulides & de Chernatony, 2010). In this context, it means measurement of cognitive and behavioral brand equity at the consumer level through a survey in order to answer the following research question: “How does celebrity endorsement of fast-moving consumer goods, moderated by level of involvement, influence customer-based brand equity compared to advertising without celebrities?”

The study addresses the research question in measuring CBBE with the construct of Aaker’s Brand Equity Ten (Aaker, 1996).

The analysis is important for both research and practitioners. Research on celebrity endorsement has mainly focused on celebrity characteristics (Ohanian, 1990), transfer of meanings (McCracken, 1989), and congruence (Kamins et al., 1994). Atkin and Block (1983) examined the hypothesis that celebrity endorsement generates greater recall for the endorsed brand and the advertisement, while Dean (1999) determined that celebrity endorsement can influence perceived product quality. There is, however, limited research that measures the total construct of CBBE. Extant research does not answer the questions of whether celebrities have value in terms of developing CBBE (Spry, 2011) and how the level of involvement moderates the endorsement effect.

Substantively and empirically, the present study is one of the first experimental investigations that focuses on the impact of celebrity endorsement on CBBE. By doing so, the research aims at adding to the extant literature by demonstrating that (1) celebrity

(9)

Theoretical Framework

endorsement (vs. non-celebrity endorsement) leads to an increase in CBBE and (2) level of involvement moderates the positive relationship. The contributions are twofold. First, it is proven that, on average, celerity endorsement is uniformly preferred in terms of CBBE of FMCG. Second, the present study investigated the moderating impact of involvement. In adding one involvement level, namely moderate involvement, this study extends empirical knowledge. It is found that under low and moderate involvement the relationship of celebrity endorsement is indeed moderated by involvement. Data illustrated that although highest total effect under low-involvement, celebrity effectiveness increases with moderate-involvement, but declines with higher levels of involvement. The trend follows an inverted U-shape, with celebrity effectiveness at moderate involvement representing the peak. This implication stands in contrast to the empirical work of Petty et al. (1983), even though they only examined celebrity effectiveness under low and high involvement.

Managerially, the implications help marketing practitioners of consumer packaged goods to investigate under which conditions and product categories a celebrity endorser pays off. Depending on the different involvement level, the effectiveness of celebrity endorsement may vary. By feeling pressured by budget restrictions and ROI, these insights reduce the risk of spending large amounts of marketing budgets on inefficient endorsements. As more and more money is spent on celebrity advertisement, and considering the number of failed endorsements in the past, these insights have clear managerial relevance. In understanding the potential brand implications, marketers can choose a dedicated celebrity in this environment.

2 Theoretical Framework

In order to build a research foundation, the following section provides a deep inside into the theoretical constructs of celebrity endorsement, level of involvement, as well as CBBE. This aims in leading the way towards specific formulated hypotheses. In addition, it is necessary to illustrate how the constructs are linked with another.

(10)

2.1 Celebrity Endorsement

Celebrity endorsement is considered to be the most prevalent and successful form of advertising and is frequently chosen by marketers as a marketing mix variable even though the costs involved are very high (Sridevi, 2014). 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.” In considering different endorsement types, the present research will examine the most frequently used type, the co-present mode, in which the celebrity physically appears with the product (McCracken, 1989). Celebrities are perceived to be more credible endorsers of products compared to non-celebrities (Muda et al, 2014). In comparison to anonymous models, who only transfer meanings towards the ad, celebrities are able to deliver powerful meanings and transfer them towards endorsed products (McCracken, 1989). This can result in higher brand attention and recall (Kotler & Keller, 2012). However, to select an appropriate celebrity to test the hypotheses, the factors that influence endorsement effectiveness need to be illustrated further. 2.1.1 Celebrity Effectiveness

Given the large sums of money spent on celebrity advertising, the selection of the right celebrity endorser can be crucial in terms of ROI. Previous studies provide evidence that credible sources are more persuasive than sources of low credibility (Ohanian, 1990, Kotler & Keller, 2012). Moreover, research examined that endorser credibility plays a significant role in the attitude of the consumer towards the communicator (Ohanian, 1990).

According to the Source Credibility Model, the effectiveness of a message depends on the level of expertise, trustworthiness and attractiveness of a chosen endorser (Hovland & Weiss, 1951, Ohanian 1990). Thus advertisers should hire endorsers who internalize these attributes (Shimp, 1997). The more pronounced the three aspects are, the more persuasive an ad’s message will be (Kahle & Homer, 1985, Kotler & Keller, 2012), and the higher purchase intentions will be (Erdogan, 1999, Friedman et al., 1976). Due to the importance and

(11)

acceptance of this concept, Ohanian’s credibility scale (1990) will be used to find an appropriate celebrity.

Selecting suitable endorsers, however, consists of more than just hiring credible individuals (McCracken, 1989). Celebrity endorsers and brands are paired together and transfer associations to each other (Till et al., 2008). Moreover, the effectiveness of a chosen endorser depends upon the meanings he or she brings to the endorsement (McCracken, 1989). These meanings can include a large variety including, “status, gender, and age, as well as personality and lifestyle types” (McCracken, 1989). Pierce Brosnan, later used in the experiment, delivers an upper-class status and is associated with a strong personality and an extraordinary lifestyle. This set of meanings enters into the endorsement process when he advertises, for instance, Spar Austria’s own private label, Spar Premium. In line with the James Bond manner he boards a luxury yacht and catches a beautiful woman in the act of stealing his briefcase containing a Spar Premium delicacy. In inviting her to an extravagant meal of Spar Premium dishes the dinner reflects luxury, the same value that Spar Premium stands for (European Supermarket Magazine, 2017).1

However, the transfer of meanings from celebrities towards products is accomplished by advertising, which serves as a tool of meaning transfer (McCracken, 1989). Furthermore, celebrity endorsement is more effective in increasing brand attitudes and purchase intentions when the celebrity image and product message is congruent (Forkan, 1980, Till 2000). Hence, celebrity choice should be based on the presence of a high fit between the celebrity’s image and the product (Kotler & Keller, 2012, Misra & Beatty, 1990). From the consumer’s perspective, high congruity leads to higher believability (Erdogan, 1999, Kamins & Gupta 1994). If there is less congruity, the endorser will overshadow the brand and the target

1

See commercial at: https://www.spar.at/content/dam/website/website_at/Aktionen/TV-Spots/SPAR_Premium_2016_50s_SparAT.mp4

(12)

audience will remember the celebrity but not the brand (Evans, 1988). The study will acquire a celebrity who is believable, attractive, credible, and reliable to the target audience.

2.2 Level of Involvement

According to Wu & Wang (2011), “Involvement is the level of relevance an individual has towards an event or object […], and is the level of attention a consumer has for a product or the level of motive and interest a product triggers” (Batra & Ray, 1985, Zaichkowsky, 1985). The current research refers to product involvement as the “consumer’s concern or reaction towards products or brands” (Wu & Wang, 2011). Past studies underpinned that endorsement effectiveness is moderated by the level of involvement (Petty et al., 1983).

The Elaboration Likelihood Model (ELM) developed by Petty et al. (1983) is widely accepted and cited in a high number of studies, can be applied to nearly any context, and is commonly used in advertising research (Bart et al., 2014). According to the ELM, a person can be coaxed by a stimulus and thus two possible routes can be taken. Under conditions of high involvement, consumers take the central route for persuasion. The central route is characterized by a thoughtful consideration of arguments, whereas under conditions of low involvement, consumers take the peripheral route to persuasion (Freeman & Chen, 2015). This results that in the case of high-involvement products, consumers are more likely to be influenced by a particular product’s true qualities (Freeman & Chen, 2015). Consumers who receive messages via the peripheral route are more likely to be influenced by peripheral cues such as celebrity endorsement since the messages are not processed cognitively (Kotler & Keller 2012, Freeman & Chen, 2015). Building strong brand equity is of great importance especially for low-involvement products, which are most likely to appear in the FMCG industry (Rust et al., 2004), but indeed varies across product categories. This includes products that are typically inexpensive and purchased frequently (Kotler & Keller, 2012). The following research aims to extend the actual state of knowledge by focusing on FMCG, which includes all, products that require low, moderate, and high involvement.

(13)

2.3 Customer-based Brand Equity

A Company’s brand is one of the most valuable intangible assets in the creation of a firm’s success; it differentiates its products from those of competitors and it is incumbent on marketing to properly manage its value (Ailawadi, 2004, Kotler & Keller, 2012).

The concept of brand equity and the important intangible value that it brings has gained remarkable attention in the marketing literature and is therefore one of the most popular marketing topics in the recent years (Keller, 2009). Firstly, brand equity can be measured based on choice or share in the marketplace, which is termed as sales-based brand equity, and secondly from the consumers’ perspective, which relates to what consumers think and feel about the brand, hereafter CBBE (Datta et al., 2017). Building on the current research, in which the focus of organizations has moved from products or services to the consumer, is crucial, because CBBE is a strong predictor of a firm’s performance, and putting consumers central in terms of marketing efforts has been proven to be successful in previous examinations (Rust et al., 2004, Kotler & Armstrong, 2009). As this shift implies a consumer-based strategy, the present research will analyze the effects from a consumer’s point of view and thereby assume that the brand’s value resides in the hearts and minds of consumers (Datta et al. 2017, Keller, 2003).

The lack of an agreed definition has originated various methodologies for measuring CBBE. The following study will take the construct of Aaker’s Brand Equity Ten (Aaker, 1996) into consideration, which has been widely accepted as relevant and most notable (Datta et al., 2017). According to this concept, CBBE is grounded in cognitive psychology and consists of four dimensions: brand loyalty, brand awareness, perceived quality, and brand associations (Aaker, 1991). A high CBBE implies that consumers are able to recognize and recall the brand; show strong, favorable and unique associations; perceive the brand as higher quality; and thus exhibit more loyalty (Aaker, 1991, Keller 1993). Although marketing activities are known to have a significant influence on CBBE (Ailawadi, 2003, Keller, 2007),

(14)

there is a lack of knowledge about the drivers of CBBE. Thus, the literature calls for further research identifying marketing activities that can build CBBE (Keller & Lehmann, 2006; Yoo et al., 2000). As marketing actions are known to have the potential to affect CBBE, the context of the present research is to address the impact of celebrity endorsement on CBBE. 2.4 Conceptual Framework and Hypotheses

Figure 1 summarizes the conceptual model and the hypotheses, as discussed next. FIGURE 1

Conceptual Framework

2.4.1 The Effect of Celebrity Endorsement on Customer-based Brand Equity

Past research has demonstrated evidence that celebrity endorsements can lead to favorable attitudes towards the endorsed brand and the ad (Petty et al. 1983, Till et al., 2008). Moreover, Erdogan (1999) argues that celebrities (vs. non-celebrities) are more effective in generating desirable outcomes such as purchase intentions, as they are capable of transferring associations from themselves towards the brands they endorse (McCracken, 1989). Due to their public fame, celebrities play a significant role in the attitude of the consumer and thus are able to demand attention, and are also able to create higher recall rates for messages in today’s environment in which consumers are flooded with massages (Atkin & Block, 1983, Erdogan, 1999, Rotfeld 2002).

Nonetheless, Keller (2003) claims that communication activities have a significant influence on CBBE and thus should be considered a brand- building tool. Although marketers

(15)

are able to call upon an increasing number of tools to develop appropriate associations, celebrity endorsement represents one suitable method in which meanings and associations can be transferred from celebrity towards brands (McCracken, 1989). Thus, when using celebrities to support an ad’s idea, celebrity endorsement is seen as a potential and powerful tool for leveraging and building CBBE (Till, 1998, Garfield 1999). In their research, Walker et al. (1992) demonstrated that matching celebrities with products positively affected subjects’ perceptions of the examined products consistent with the image of the celebrity (Till, 1998). Based on this argumentation it is expected that celebrity endorsement has a significant positive effect on CBBE:

H1: Celebrity endorsement positively affects based brand equity, so that customer-based brand equity shows higher values with the celebrity (vs. non-celebrity).

2.4.2 The Moderating Role of Involvement

Ads featuring celebrity endorsement are expected to have a more positive effect on CBBE in comparison to ads without celebrities, but effectiveness might vary under different conditions. Notably, the literature emphasized that endorsement effectiveness is moderated by the level of involvement (Petty et al., 1983). Individual level of involvement is a significant factor in determining differences among celebrity endorsement effectiveness (Kotler & Keller, 2012).

Based on the findings of the ELM, a person can be coaxed by a stimulus, and there are two routes that are able to affect consumer judgement (Petty et al. 1983). First, the central route is characterized by consumers that are rationally considering the alternatives. In the second peripheral route, consumers’ attitudes change because of external positive or negative cues rather than considering product-related features. This finding’s result is that in the case of high-involvement products, consumers are more likely to be influenced by a particular product’s true qualities and thereby apply central route information processing as they show higher motivation and the ability to process the product information (Freeman & Chen, 2015, Petty et al., 1983, Wu & Wang 2011). In turn, consumers who have low involvement are

(16)

Method

receiving messages via the peripheral route and are more likely to be influenced by non-product-related cues such as celebrity endorsement, since the messages are not processed cognitively and consumers are evaluating the credibility of the source rather than the functional attributes of the product (Kotler & Keller 2012, Wu & Wang, 2011).

Consumers of low-involving FMCG products for example, choose a brand without much evaluation at the point of sale and may switch the brand in the next purchase out of a desire for a different taste or the like (Kotler & Keller, 2012). In product categories where consumers do not spend much time on purchase decision, celebrity endorsement might help to differentiate between the brands at an emotional level (Callcoat & Phillips, 1996, Das & Padhy, 2012). Literature emphasizes that secondary associations such as celebrity endorsements affect consumer evaluation of a product more effectively when consumers lack motivation and the ability to judge on a deeper level (Keller, 2005). Thus,

H2: The positive relationship between celebrity endorsement and customer-based brand equity is moderated by level of involvement, so that this relationship is stronger for low involvement compared to moderate and high involvement.

3 Method

In order to test the study’s hypotheses, a 2 (celebrity vs. non-celebrity) × 3 (low vs. moderate vs. high involvement) between-subjects experiment was designed (see table 1). The hypotheses were tested by using a self-administered online questionnaire. The main study is preceded by two pre-tests that first determined the low, moderate and high product- involvement manipulation in the experiment, and second the celebrity to be used under each condition. The 2 × 3 factorial design led to six different questionnaires as presented in Table 1. Each of the questionnaires is reported in Appendix A.

(17)

Method

TABLE 1

Questionnaire Legend for 2 × 3 Factorial Design

Survey Celebrity Level of Involvement

Yes / No Low / Moderate / High

1 Yes High 2 Yes Moderate 3 Yes Low 4 No High 5 No Moderate 6 No Low 3.1 Pre-test

The first pre-test was performed to check the manipulation of product involvement. Seventeen respondents participated in an online pre-test developed on Qualtrics. After being asked for their consent to participate in a study about personal involvement, participants then evaluated nine FMCG, using the 10-item Revised Personal Involvement Inventory scale (Zaichkowsky, 1994). This scale measures product involvement on affective and the cognitive components of consumers’ product involvement. The pre-test is reported in Appendix B. Product involvement for the nine products were calculated using the means for all 10 items (Zaichkowsky, 1994). The average product involvement ratings for the nine products are displayed in the Table 2. The difference between perfume and frozen pizza were the largest among all mean differences. These two products were chosen as the high vs. low manipulation of consumer involvement in the experiment. As being in between, water is chosen as moderate-involvement product.

(18)

Method

TABLE 2

Pre-test Product Involvement (n = 17)

Table 2 provides descriptive statistics on product involvement for nine different FMCG. Based on means, perfume and frozen pizza were chosen as “high vs. low” manipulation of consumer involvement in the experiment. As being in between, water is chosen as moderate-involvement product.

Variable Minimum Mean Maximum Median Std. Deviation

Shampoo 3.7 4.63 5.8 4.40 0.74 Perfume 4.4 5.46 6.1 5.70 0.57 Soap 2.2 3.99 5.6 4.30 1.03 Frozen Pizza 2.6 3.33 4 3.40 0.51 Chocolate 1.4 4.41 6 4.60 1.27 Toothpaste 2.4 4.86 6.5 5.00 0.96 Water 1.4 4.61 6.9 5.00 1.28 Energy Drink 1.9 3.34 5.3 3.30 1.04 Face Cream 1 3.95 6.3 4.10 1.56

McCracken (1989) emphasizes the importance of congruence between the endorsed product and the celebrity. The celebrities chosen for the pre-test is based on Nielsen’s (2015) research, which listed the top 10 most liked celebrities in Q1 2015. However, their research is based on the US market. Since the sample in the study mainly consists of Dutch, German and other nationalities the use of an appropriate celebrity must be tested before conducting the main experiment. Three movie actors, Pierce Brosnan, Matthew McConaughey and Jennifer Garner were chosen from Nielsen (2015) to carry the low, moderate and high-involvement products. The scale is used on a 7-point Likert scale and for each product, respondents were asked if they perceive a good match between celebrity and product (Kamins, 1990) and if the endorsement is credible (Ohanian, 1990). The items used were: bad – good, unattractive – attractive, unreliable – reliable, not credible – credible and unbelievable – believable (Ohanian, 1990). The second pre-test is reported in Appendix C. The means indicated that endorsement between: (1) Pierce Brosnan and perfume (2) Pierce Brosnan and water (3) Pierce Brosnan and frozen pizza were perceived as most relevant (see Table 3).

(19)

Method

TABLE 3

Pre-test Celebrity-product Congruence (n = 19)

Table 3 provides descriptive statistics about three celebrities that were chosen for carrying three different FMCG products. A higher mean indicates better celebrity-product congruence. The celebrity selection is based on Nielsen (2015).

Variable Perfume Water Frozen Pizza

M SD M SD M SD Celebrity Pierce Brosnan 5.53 1.51 4.67 1.50 4.14 2.03 Matthew aMcConaughey 4.93 1.26 4.54 1.67 3.31 1.71 Jennifer Garner 4.64 1.74 4.33 1.71 3.79 1.90 3.2 Sample

The study focusses on consumers of FMCG. This results in that these consumers represent the population of interest in the empirical study. The research will use non-probability sampling techniques and sampling is therefore done in a non-random manner. Furthermore, the selection of the sample is based on convenience sampling, where respondents were personally contacted through sending an anonymous link via e-mail, Facebook and Whatsapp. After collecting the data, the main study reached 504 respondents, so that after randomization each survey were filled in by (1) 86, (2) 85, (3) 83, (4) 79, (5) 86, and (6) 85 participants. Germans represent the majority of the experiment with 59.3% while the Dutch come second with 8.5%. Moreover, the sample consists of 63.3% Male and 36.7% Female. Figure 1 summarizes the descriptive statistics under all conditions:

(20)

Method

FIGURE 1

Main Experiment Descriptive Statistics (n = 504)

Figure 1 shows means across all experimental conditions. In terms of CBBE, celebrity endorsement always scores higher than non-celebrities. After randomization, each survey were filled in by (1) 86, (2) 85, (3) 83, (4) 79, (5) 86, and (6) 85 participants. Germans represent the majority of the sample with 59.3% followed by the Dutch who come second with 8.5%.

3.3 Measurement

The pre-test determined the stimuli used for the main experiment. For the independent variable, endorsers were chosen by the criteria trustworthiness, expertise and attractiveness on a 7-point Likert scale (e.g. the endorser is reliable) interval level (Ohanian, 1990). The Personal Involvement Inventory scale (Zaichkowsky, 1994) was used to measure the moderator product involvement (e.g. imagine you are in a drug store to buy a shampoo, to you a shampoo is boring – interesting). The average means were used to examine high, moderate and low involvement (Wu & Wang, 2011).

In measuring CBBE the construct of Aaker’s Brand Equity Ten (Aaker, 1996) is used. As illustrated in table 4, the theoretical construct of CBBE is split into four dimensions brand awareness, brand associations, perceived quality and brand loyalty. Based on the dimensions

4.628 4.517 4.479 4.561 4.084 4.304 3.7 3.9 4.1 4.3 4.5 4.7

Low Moderate High

C B B E Level of involvement Celebrity Non-Celebrity

(21)

Method

defined by Aaker (1996), the next step derived measurable indicators, which were then transferred to survey items (e.g. “I can easily remember this brand”). The items are slightly rephrased, since CBBE will be based on six self-issued campaigns. The revised construct consists of sixteen items and has a high reliability, with Cronbach’s alpha = .885:

TABLE 4

Operationalization of CBBE

Variablea Items

Brand awareness

Brand recognition I can recognize the brand among other competing brands Brand recall I can easily remember this brand

I can quickly recall the symbol or logo of this brand Brand knowledge I know what this brand stands for

Brand opinion I have an opinion about this brand

Brand associations

Perceived value The brand provides good value for the money Personality This brand has personality

I have a clear image of the type of person using this brand Organization I consider the company who stands behind this brand as trustworthy Differentiation I can easily differentiate this brand from competing brands

Perceived quality

I perceive good quality of this brand

In comparison to alternative brands, this brand has consistent quality

Brand loyalty

Price premium I would buy this brand although it is more expensive than other brands Loyalty I consider myself to be loyal to this brand

I would not buy any other brands of equal quality, when this brand is available

_{I would recommend the brand to others}

a. Total construct of CBBE based on Aaker (1996), Yoo & Donthu (2001), Girard et al. (2017)

All items have a good correlation with total score of the scale > .30. The fifth dimension market behavior was excluded, as it relies on market data and is not measurable in a consumer survey (Aaker, 1996). Later on, the survey consisted of two manipulation checks. The first check examines if people can recognize the person on the campaign, respondents chose between two options (a) text field or (b) “I don’t know”. The second check included the

(22)

Method

product involvement. At the final part of the survey, the questionnaire asked respondents for their demographics.

3.4 Procedure

The main experiment was set up using the online survey platform Qualtrics and lasted from 11th of April until the 28th of April 2017. After receiving and clicking on the link, participants were provided with a short introduction of the research. In order not to affect the results, the purpose of this study was not mentioned to sustain internal validity (Field, 2003). Furthermore, to incentivize participants, six €15 Amazon gift cards were randomly distributed. After the welcome message, one of the six conditions (celebrity-product combination) randomly appeared on the screen of each participant. Followed by seeing the campaign, participants were asked to fill in the questionnaire based on the stimulus that they had just received. Subsequently, the questionnaire closed by a last message thanking respondents for their participation.

3.5 Data analysis

A Multiple regression analysis was conducted to investigate the relationship between celebrity endorsement and level of involvement to predict CBBE. Level of involvement consists of three categories and need to be rephrased in two dummy variables (Field, 2013). In the current research, moderate level of involvement is chosen as baseline and builds the group in which low and high involvement are compared. The regression model for CBBE with the predictor celebrity and level of involvement, with moderate involvement as baseline is:

CBBEi,c = β0 + β1 Celebrity + β2 Moderate vs. Low + β3 Moderate vs. High + β4 (Celebrity × Low) + β5 (Celebrity × High)

+ ∑₌ β Control + ε,

where i indicates low, moderate and high-involvement and c celebrity and non-celebrity. The equation expresses CBBE as a function of celebrity endorsement, level of involvement, the interaction of celebrity and involvement and the control variables. Celebrity represents the

(23)

Results

absence (no = 0) or presence (yes = 1) of the celebrity. The dummy variable, Moderate vs. Low, depicts low involvement (yes = 1), while Moderate vs. High (yes = 1), constitutes high-involvement conditions.2 The interaction effect between celebrity and level of involvement, plus the control variables gender, age and education complete the model. The error term ε is normally distributed. All predictor variables are dummy variables and do not need to be mean-centered.

4 Results

4.1 Correlation

Table 5 shows correlations, means and standard deviations across the observations in the sample. As expected, it is found that celebrity endorsement correlates positively with CBBE (r = .12, p < .01). Level of involvement, in turn, correlates negatively (r = ‒.09, p < .05) with CBBE. As being the only significant control variable, gender positively correlated with CBBE (r = .10, p < .05). This implicates, that being a woman is correlated to a higher score on the CBBE scale. That makes sense, since the celebrity and non-celebrity chosen in the experiment were male and were among others chosen based on their attractiveness.

TABLE 5: Correlation Matrix (n = 504)

Table 5 shows correlations across the observed data. Celebrity endorsement is positively correlated with CBBE, while level of involvement is slightly negative correlated with CBBE. As measured construct, CBBE shows high reliability (Cronbachs α = .89)

Variables 1 2 3 4 5 6 1. Celebrity endorsement 1 2. Level of involvement -.02 1 3. CBBE .12** -.09* 1 (.89) 4. Gender .03 .06 .10* 1 5. Age .00 -.04 -.05 -.20** 1 6. Education level .02 -.03 -.07 -.00 .27** 1 Mean .50 2.01 4.39 1.37 2.67 4.07 Standard deviation .50 .81 .93 .48 .55 .91

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

2

If for example, celebrity is present and both involvement variables are zero, multiple regression analysis calculates the CBBEmoderate,celebrity scenario.

(24)

Results

4.2 Hypothesis Testing

Table 6 summarizes the elasticities obtained across the sample. Two regression models are conducted. The Main-Effects Model shows observation without the interaction effects, while the Full-Effects Model, with moderate involvement as baseline, takes also the interaction effects into account. Including the Full-Effects Model illustrates how the moderator strongly affects the existing causal mechanisms. Through the increase in slope, and by strengthening the positive impact, especially under moderate-involvement, the interaction effects show novel insights. Both regression models are statistically significant, the Main-Effect Model with F (6,497) = 4.03, p < .00 and the Full-Effects Model with F (8,495) = 3.77, p < .00.

In support of hypotheses 1, a significant main effect of celebrity endorsement on CBBE (b1 = .23, p < .01) is found. This means celebrity endorsements have a .23,

approximately 5.6%, higher CBBE than ads without celebrities.

In order to test hypothesis 2, the interaction effect between involvement and celebrity endorsement needs to be examined. This needs a turn to the Full-Effects model, where all parameters remain robust in sign and level of significance. When using moderate involvement as baseline, celebrity endorsements (b1 = .43, p < .01) have a .43, approximately 11%, higher

CBBE than ads without celebrities. Low-involvement manipulation (b2 = .48, p < .01) was

statistically significant, while high-involvement (b3 = .21, p < .20) led to marginally

significant results. In addition, it is found evidence that the regression coefficient for the interaction 1, low-involvement × celebrity (b4 = ‒.37, p < .10) is statistically different from

zero, t (498) = ‒1.99. This interaction effect shows, although the CBBE under low involvement might be higher, the link between celebrity endorsement and CBBE is smallest under low involvement with b1 + b4 = .06. The second interaction effect, high involvement ×

celebrity (b5 = ‒.26, p < .20) is not statistically different from zero with t (498) = ‒1.29, so

that there is no significant interaction effect between celebrity endorsement and high involvement. Although marginally significant, as seen from probing the interactions, it can be

(25)

Results

identified that under high involvement the link between celebrity endorsement and CBBE is higher than under low involvement, but lower than moderate-involvement conditions with b1

+ b5 = .17.

Re-specification with low-involvement as baseline supports the arguments. Both manipulations, moderate involvement (b2 = ‒.48, p < .01) and high involvement (b3 = ‒.26, p

< .10) are statistically significant. In line with the interaction effect examined before, moderate involvement × celebrity endorsement (b4 = .37, p < .10) is statistically different

from zero and positive with t (498) = 1.99. High involvement × celebrity (b5 = .11, p > .10)

however, is not statistically different from zero t (498) = .85, so that there is no significant, but slightly positive, interaction effect between celebrity endorsement and high involvement. As for the control variables, it is found that gender has a positive effect on CBBE (b6 = .16, p

< .10) in such a way that CBBE is higher under female consumers. In contrast, age (b7 = ‒.01, p > .10) and education level (b8 = ‒.06, p > .10) led to insignificant, but slightly negative

impact on CBBE.

However, a closer inspection of the conditional effects indicates the relationship between celebrity endorsement and CBBE only among respondents who perceive low- and moderate involvement. Thus, the effect of celebrity endorsement on CBBE partly depends on consumers’ level of involvement and the positive relationship between celebrity endorsement and CBBE is stronger among participants who perceive moderate involvement. H2 is partially supported. Figure 2 illustrates the interaction effects. Celebrity effectiveness however, shows a curvilinear distribution in form of an inverted U-shape. This finding extends beyond the formulated hypothesis and will be presented next.

(26)

Results

TABLE 6

Multiple Regression Model

Table 6 shows elasticities across the observed data. The Main-Effects Model examines only the impact of celebrity endorsement and involvement on CBBE. With moderate involvement as baseline, the Full-Effects Model takes also the interactions into account. All parameters stay robust in significance.

Modela Main-Effects Model Full-Effects Model

B SE β t-Value B SE β t-Value Constant 4.27*** .08 51.00 4.08*** .30 13.47 Celebrity .23*** .08 .12 2.81 .43*** .14 .23 3.10 Moderate as baseline vs. Low .29*** .08 .08 .08 .48*** .14 .24 3.38 Moderate as baseline vs. High .10 .08 .08 .08 .22 .14 .11 1.54 Interaction Effects Low × Celebrity -.37* .20 -.15 -1.84 High × Celebrity -.26 .20 -.10 -1.29 Control Variables Gender .17* .09 .09 1.91 .16* .09 .08 1.85 Age -.02 .08 -.01 -.20 -.01 .08 -.01 -.14 Education level -.06 .05 -.06 -1.22 -.06 .05 -.06 -1.22 Notes: *: p < .1; **: p < .05; ***: p < .01 a. Model Summary: R2 = .05

b. Dependent Variable: CBBE

FIGURE 2 Interaction Effects

Figure 2 shows an interaction. Although Multiple Regression calculated only the low involvement × celebrity endorsement (B = ‒.37, p < .10) and moderate involvement × celebrity endorsement (B = .37, p < .10) interaction as statistically significant the slope of all groups follow the same direction.

4 4.1 4.2 4.3 4.4 4.5 4.6 4.7 No Celebrity Celebrity CBB E Low-involvement High-involvement Moderate-involvement

(27)

Results

4.3 Extension, the Inverted U-shape

The purpose of the study was to provide evidence that (a) celebrity endorsement leads to a significant increase in CBBE and (b) the positive relationship is moderated under different levels of involvement. Closer inspection of interaction effects however, led to a moment of serendipity. While celebrities have lowest effects under low-involvement conditions (b1 + b4 =

.06, p < .10), moderate involvement represents the peak and turning point (b4 = .37, p < .10),

with a declining effect when turning towards high involvement (b1 + b5 = .06, p < .20). This

data finds clear nonlinearities in form of an inverted U-shape as illustrated in Figure 3.

One may ask why the interaction effects show a curvilinear distribution. The interaction effect is a product of celebrity endorsement, which has a positive impact on CBBE, and level of involvement. Through the interaction, an inverted U emerges in such that level of involvement shifts CBBE scores of the moderate-involvement product so intense that it exceeds the high-involvement product. Thus, the interaction effects steepen the gradient, with highest increase under moderate involvement. In terms of celebrity effectiveness, data shows that too much involvement can be as detrimental as too little.

(28)

Discussion

FIGURE 3 Interaction-strength

Figure 3 shows an inverted U-shape. With moderate involvement as baseline, low involvement shows the smallest interaction effects (B = ‒.37, p < .10). High involvement is marginally statistical significant (B = ‒.26, p < .20).Thus, celebrity endorsement is most effective at moderate involvement.

5 Discussion

Advertisement in the FMCG industry is highest compared to other industries (Nielsen, 2016), with consumers flooded with a vast number of ads every day. To get access to customer mind-set, celebrity endorsement is one brand-building tool that companies can make use of (Keller, 2005). Celebrities are used more frequently by FMCG manufacturers which most commonly promote product categories like personal care, and food and beverages (Millward brown, 2016). But do celebrities help in achieving a higher CBBE considering FMCG? It is found that, on average, celebrity endorsement indeed led to a higher CBBE than advertising without celebrities under all levels of involvement. In addition, to understand under which involvement celebrity endorsement works best, the moderating impact of the level of involvement is studied. When consumers are facing low- and moderate-involvement FMCG products, the positive effect of celebrity endorsement is strengthened, and in terms of CBBE, even more favorable in moderate-involvement conditions. In addition, the interaction effects

0.06 0.43 0.17 0 0.1 0.2 0.3 0.4 0.5

Low Moderate High

Int er act ion -ef fec t Level of involvement

(29)

Discussion

show nonlinearities in the form of an inverted U-shape and extend the formulated hypotheses, as reflected on the theoretical and managerial implications next.

5.1 Theoretical Implications

The current research aimed at adding to the extant literature on celebrity endorsement by demonstrating that (1) celebrity endorsement (vs. non-celebrity endorsement) leads to an increase in CBBE and (2) the level of involvement moderates the positive relationship.

First, previous research about celebrity endorsement traditionally paid attention to the endorser in the brand-endorser relationship and examined elaborately which type of endorsers are suitable in terms of endorser characteristics (Ohanian, 1990), meanings (McCracken, 1989, Walker et al. 1992) and congruence (Kamins et al. 1994). In their studies, Atkin & Block (1983) and Petty et al. (1983) found that celebrity endorsers increased positive attitudes towards the ad. But can celebrities have an impact on CBBE? By means of an experiment, the underlying research adds to the extant literature by addressing the brand-endorser relation. The results show that celebrity endorsement is uniformly preferred in terms of the CBBE of FMCG. This is in line with the theory-based expectations. Moreover, these findings are consistent with Keller (2005) and Till (1998), who argue that using celebrity endorsers as an executional device has the potential to impact CBBE in creating new associations or updating existing associations.

Second, the present study examined the moderating impact of involvement. Petty et al. (1983) report that the effectiveness of celebrity endorsement varies under low and high levels of involvement and implies that in terms of advertising effectiveness, celebrities are more effective under low-involvement conditions. The present research extends the work of Petty et al. (1983) in adding one involvement level, moderate involvement. It is found that under low and moderate involvement the relationship of celebrity endorsement on CBBE is indeed moderated by involvement. In probing the interactions, it is seen that, although it has the highest total effect, celebrity effectiveness under low involvement is smaller than under

(30)

Discussion

moderate involvement and even (marginally significantly) smaller than in high-involvement conditions. This implication stands in sharp contrast, but comparison has to be taken with caution since Petty et al. (1983) examined the impact of celebrity endorsement on advertising and not on CBBE. Moreover, the moderate-involvement product in the present research has the lowest CBBE without the celebrity, but overtakes the high-involvement product due to the interaction effect. Thus, the interaction effect shows a curvilinear distribution in the form of an inverted U, with moderate involvement representing the peak. This novel insight extends the hypotheses of the present study and warrants further investigation. The present data show the importance of peripheral cues in brand equity research.

5.2 Managerial Implications

In consequence of restricted marketing budgets, managers of FMCG brands should consider the customer’s perspective as well as a financial point of view and ask themselves if celebrities are worth the money. Thus, FMCG companies are advised to calculate the net present value (NPV)3 of their campaign to decide whether to undertake the investment or not. If positive, decision maker of FMCG companies are advised to undertake the celebrity campaign in question due to its favorable impact on CBBE.

Advertising with celebrities are strategic in that they can be crucial in transferring associations from themselves toward brands and help to sell products (Keller & Kotler, 2012), but managers need to take the pitfalls into consideration. Threatened by the risk of being involved in scandals (e.g. Tiger Woods and Accenture) or due to multiple endorsements (e.g. Beyoncé and Pepsi), acquisition of celebrities can be an expensive flop. One alternative to promote celebrities is to create a brand’s own celebrities with unique associations and longer usage time. Mexican beer brand Dos Equis grew in sales by over 25% by introducing their own celebrity in 2006, Jonathan Goldsmith, titled in their ads as “The Most Interesting Man in

3

NPV is used in financial decision making for analysing an investment or project. It is calculated as: NPV = Present value of benefits ‒ Present value of costs (Berk et al., 2013)

(31)

Discussion

the World” (Sager, 2016, Kotler & Keller, 2012). With thousands of social media followers, Jonathan Goldsmith is the perfect example that establishing an exclusive celebrity can lead to success. He is the target of several social media content like image macros and even the American sitcom Family Guy produced an episode with the title, “The Most Interesting Man in the World.”

The use of celebrities fluctuates enormously across geographical areas. However, using the right celebrity, in the study emphasized as one who (a) is likeable, believable, and attractive and (b) shows high endorser-product congruence, can be a powerful brand asset, no matter which country or product category (Millward Brown, 2016).

Finally, in terms of celebrity effectiveness, too much involvement can be as bad as too little. Because celebrity effectiveness is highest on moderate involvement, practitioners should raise low-involvement products into moderate-involvement products, like fruit juice makers did when including vitamins to reinforce their drinks (Kotler & Keller, 2012). This will lead to more effective celebrity campaigns.

5.3 Limitations and Further Research

The present research has several limitations that offer some suggestions for further studies. First, the high-involvement product used in the experiment, men’s perfume, also took female respondents into consideration, although they might neither be interested in the product nor have any purchase intentions. Further research might find a significant interaction effect in focusing on products that were consumed and purchased by both gender groups.

Second, generalizations must be taken with caution. The present research took only manufacturing brands into consideration. Private label brands also offer FMCG and gain importance, especially in Europe, as their presence grows on supermarket shelves (Ailawadi & Keller, 2004, Bain, 2015). Further research could compare the impact of celebrity endorsements on the brand equity of manufacturing brands vs. national brands, as retail stores need to offer A-brands and private labels together to maintain variety (McKinsey, 2013).

(32)

Discussion

Third, the information about Pierce Brosnan was only presented in the beginning of the survey, without cognitive disturbances in between. This does not represent real-life scenarios, where consumers process considerable information between ad exposure and product tension. According to Till et al. (2008), celebrities should have been presented multiple times over a period of time, resulting in associative learning. Thus, repetition can strengthen the brand-celebrity link. Further studies can first show the ad several times distributed among several days, but send the survey seven days after the last stimulus.

Fourth, the non-celebrity scenarios are created with Adobe Photoshop. In adding the face of an unknown endorser who looks similar to Pierce Brosnan, the study aimed to minimize validity threats. However, the non-celebrity in the study shows a strong smile, while the celebrity does not. Future studies need to take facial expression and emotional appeal into consideration, which might create bias in the results.

Finally, the inverted U-shape that emerges when level of involvement moderates the causal relationship between celebrity endorsement and CBBE offers avenues for further research. Driven by the idiom that too much of something can be as detrimental as too little, the present research calls for research investigating the inverted U-shape distribution of celebrity effectiveness under low, moderate and high involvement. The research by Haans et al. (2016) provides a checklist for theorizing and testing U-shaped relationships.

(33)

References

6 References

Aaker, D.A. (1991). Managing brand equity. New York: Free Press.

Aaker, D. A. (1996). Measuring brand equity across products and markets. California

Management Review, 38(3), 102-120.

Aaker, J. L. (1997). Dimensions of brand personality. Journal of Marketing Research, 34(3),

347-356.

Ailawadi, K. L., Lehmann, D. R., & Neslin, S. A. (2003). Revenue premium as an outcome measure of brand equity. Journal of Marketing, 67(4), 1-17.

Ailawadi, K. L., & Keller, K. L. (2004). Understanding retail branding: conceptual insights and research priorities. Journal of Retailing, 80(4), 331-342.

Agrawal, J., & Kamakura, W. A. (1995). The economic worth of celebrity endorsers: An event study analysis. Journal of Marketing, 56-62.

Atkin, C., & Block, M. (1983). Effectiveness of celebrity endorsers. Journal of Advertising

Research, 23, 57-61.

Bain & Company (2015, June 24). Winning with brands. Retrieved June 20, 2017, from http://www.bain.com/publications/articles/winning-with-brands.aspx.

Baron, S., Davies, B., & Swindley, D. (1991). Macmillan dictionary of retailing. London: Macmillan.

Bart, Y., Stephen, A. T., & Sarvary, M. (2014). Which products are best suited to mobile advertising? A field study of mobile display advertising effects on consumer attitudes and intentions. Journal of Marketing Research, 51(3), 270-285.

Berk, J., DeMarzo, P., Harford, J., Ford, G., Mollica, V., & Finch, N. (2013). Fundamentals

of Corporate Finance. Pearson Higher Education AU.

Callcott, M. F., & Phillips, B. J. (1996). Observations: Elves make good cookies: Creating likable spokes-character advertising. Journal of Advertising Research, 36(5), 73-73.

Cheng-Hsui Chen, A. (2001). Using free association to examine the relationship between the characteristics of brand associations and brand equity. Journal of Product & Brand

Management, 10(7), 439-451.

Christodoulides, G., & De Chernatony, L. (2010). Consumer-based brand equity

conceptualization and measurement: A literature review. International Journal of Research in

Marketing, 52(1), 43-66.

Cobb-Walgren, C. J., Ruble, C. A., & Donthu, N. (1995). Brand equity, brand preference, and purchase intent. Journal of Advertising, 24(3), 25-40.

(34)

References

Crutchfield., D. (2010, September 22). Celebrity Endorsements Still Push Product. Retrieved March 12, 2017, from

http://adage.com/article/cmo-strategy/marketing-celebrity-endorsements-push-product/146023/

Das, S., & Padhy, P. (2012). Brand Perception by Celebrity Endorsement. International

Journal of Advanced Research in Management and Social Sciences, 1(2).

Datta, H., Ailawadi, K. L., & van Heerde, H. J. (2017). How Well Does Consumer-Based Brand Equity Align with Sales-Based Brand Equity and Marketing-Mix Response?. Journal of Marketing, 81(3), 1-20.

Dean, D. H. (1999). Brand endorsement, popularity, and event sponsorship as advertising cues affecting consumer pre-purchase attitudes. Journal of Advertising, 28(3), 1-12. Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing

Management, 15(4), 291-314.

Erdogan, B. Z., Baker, M. J., & Tagg, S. (2001). Selecting celebrity endorsers: The practitioner's perspective. Journal of Advertising Research, 41(3), 39-48.

European Supermarket Magazine. (2017, January 08). Pierce Brosnan Shakes Up Spar Premium Ad. Retrieved March 12, 2017, from https://www.esmmagazine.com/pierce-brosnan-shakes-spar-premium-ad/36990

Evans, R. B. (1988), Production and Creativity in Advertising, London: Pitman.

Farquhar, P. H. (1989). Managing Brand Equity. Marketing Research, 1(3), 24-33. Field, A. (2013). Discovering statistics using IBM SPSS statistics. London: Sage.

Forkan, J. (1980). Product matchup key to effective star presentations. Advertising Age, 51(6), 42.

Fowles, J. (1996). Advertising and Popular Culture (Vol. 5). London: Sage.

Freeman, K. S. & Chen C.C. (2015). Wither the Impact of Celebrity Endorsement. Online Journal of Communication and Media Technologies. 213-237.

Friedman, H. H., Termini, S., & Washington, R. (1976). The effectiveness of advertisements utilizing four types of endorsers. Journal of Advertising, 5(3), 22-24.

Friedman, H. H., & Friedman, L. (1978). Does the celebrity endorser's image spill over the product. Journal of the Academy of Marketing Science, 6(4), 291-99.

Garfield, R. (1999). TD Waterhouse makes trade with celebs. Advertising Age, 70, 79. Girard, T., Trapp, P., Pinar, M., Gulsoy, T., & Boyt, T. E. (2017). Consumer-Based Brand Equity of a Private-Label Brand: Measuring and Examining Determinants. Journal of

(35)

References

Haans, R. F., Pieters, C., & He, Z. L. (2016). Thinking about U: Theorizing and testing U‐and inverted U‐shaped relationships in strategy research. Strategic Management Journal, 37, 1177-1195.

Hawkins, D. I., Best, R. J., & Coney, K. A. (1983). Consumer Behavior: Implication for Marketing, Strategy Revised ed. Business Publication. Inc, Plano, Texas, 448.

Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion, New Haven, CT: Yale University Press.

Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public opinion quarterly, 15(4), 635-650.

Jacoby, J. (1984). Perspectives on information overload. Journal of Consumer

Research, 10(4), 432-435.

Kahle, L. R., & Homer, P. M. (1985). Physical attractiveness of the celebrity endorser: A social adaptation perspective. Journal of consumer research, 11(4), 954-961.

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.

Kamins, M. A., & Gupta, K. (1994). Congruence between spokesperson and product type: A matchup hypothesis perspective. Psychology & Marketing, 11(6), 569-586.

Kanter, R. M. (2014, July 23). Tiger Woods and the Coming Decline of Celebrity

Endorsements. Retrieved April 18, 2017, from https://hbr.org/2009/12/tiger-woods-and-the-coming-dec.html

Keller, K.L. (1993). Conceptualizing, measuring, and managing customer-based equity. Journal of Marketing, 57(1), 1-22.

Keller, K.L. (2003), Strategic Brand Management: Building, Measuring and Managing Brand Equity, 2nd ed., Prentice Hall, Englewood Cliffs, NJ.

Keller, K. L., & Lehmann, D. R. (2003). How do brands create value?. Marketing

Management, 12(3), 26-31.

Keller, K. L. (2005). Branding shortcuts: Choosing the right brand elements and leveraging secondary associations will help marketers build brand equity. Marketing Management, 14(5),

19-23.

Keller, K. L., & Lehmann, D. R. (2006). Brands and branding: Research findings and future priorities. Marketing science, 25(6), 740-759.

Keller, K. L. (2007). Advertising and brand equity. GJ Tellis, & T. Ambler, Handbook of

Advertising, 54-70.

Keller, K. L., & Aaker, D. A. (1992). The effects of sequential introduction of brand extensions. Journal of Marketing Research, 35-50.

(36)

References

Keller, K. L. (2009). Building strong brands in a modern marketing communications environment. Journal of Marketing Communications, 15(2-3), 139-155.

Kelman, H. C. (1961). Processes of opinion change. Public Opinion Quarterly, 25(1), 57-78. Kotler, P., & Armstrong, G. (2009). Principles of Marketing 12th Edition: Pearson Prentice Hall. Upper saddle river New Jersey.

Kotler, P. & Keller, K.L. (2012). Marketing Management. New Jersey: Prentice Hall

McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement process. Journal of Consumer Research, 16(3), 310-321.

McGuire, W. J. (1968). Personality and susceptibility to social influence. Handbook of

Personality Theory and Research, 2, 1130-1187.

McKinsey (2013). ‘Better every day’: A conversation with the CEO of Ahold. Retrieved from http://www.mckinsey.com/client_service/retail/latest_thinking/perspectives_second_edition/~ /media/86880A22FD594CA6A45DFA6E23D6994E.ashx.

McSweeney, F. K., & Bierley, C. (1984). Recent developments in classical conditioning. Journal of Consumer Research, 11(2), 619-631.

Millward Brown (2016). Celebrity-based ad campaigns: What are the pros and cons? Retrieved from

http://www.millwardbrown.com/docs/default-source/china-downloads/newsletter/millwardbrown_kp_celebrity_based_ad_campaigns-pdf.pdf 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.

Moore, D. (2013, October 2016). TV Advertising Gold. Retrieved March 12, 2017, from http://www.nielsen.com/eu/en/insights/news/2013/tv-advertising-gold.html

Muda, M., Musa, R., Mohamed, R. N., & Borhan, H. (2014). Celebrity entrepreneur endorsement and advertising effectiveness. Procedia-Social and Behavioral Sciences, 130, 11-20.

Nielsen (2016). Germany 2016. Retail, Consumers, Advertising. Retrieved from http://www.nielsen.com/content/dam/nielsenglobal/de/docs/Nielsen_Booklet_Retail-Consumers-Advertising_2016.pdf

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. Ohanian, R. (1991). The impact of celebrity spokespersons' perceived image on consumers' intention to purchase. Journal of Advertising Research, 31(1), 46-52

Olenski, S. (2016, July 20). How Brands Should Use Celebrities For Endorsements. Retrieved March 12, 2017, from https://www.forbes.com/sites/steveolenski/2016/07/20/how-brands-should-use-celebrities-for-endorsements/#52a220ae5593

The effect of celebrity endorsement on customer-based brand equity concerning fast moving consumer goods : an experimental investigation