A Multiple Product Endorser can be a Credible SourceAndrew Redenbach (Griffith University)
This article examines celebrity endorsement with a new approach. Taking source credibility and multiple product endorsement into account, when assessing the effects on purchase intention. Previous studies have examined credibility and multiple product endorsement separately. Previous studies found that it was more appropriate to use single product endorsers. This study found that a multiple product endorser could be a more effective source than a single product endorser. Implications for future research are discussed.
The use of a celebrity endorser as the source for a communications message has increased dramatically in the past few years. Celebrities have been used to promote everything from sports drinks to digital mobile phones and the use of communication mediums seems endless. Endorsers are used in broadcast media, both Television and Radio, in print media and even in outdoor billboard advertising, and their effectiveness in each has been tested.
The value of celebrity endorsements has been recognized world wide, but it is the United States, which has most readily accepted the practice. Firms such as Burns Sports Celebrity Service Inc. operate with their primary function being negotiating celebrity endorser contracts. "IEG Endorsement Insider estimates that $800 million will be spent this year in the U.S. to acquire talent - entertainers, athletes and other high profile personalities - to spotlight in advertising, promotion and PR campaigns." (Thompson, 1998) These facts suggest the ever-growing popularity of using celebrities as a spokesperson or endorser in marketing communication. In fact "according to industry sources, approximately 20% of all television commercials feature a famous person." (Agrawal and Kamakura, 1995, p.56) This equates to almost 10% of all advertising expenditure being spent to pay the endorser. (Agrawal and Kamakura, 1995)
The use of the celebrity endorser is not limited to any one industry. "Companies using endorsement cross all industry categories - including packaged goods, telecommunications, financial services and auto." (Thompson, 1998). Collectively the figures are astounding, but individually the money a top line celebrity endorser can make is phenomenal. Michael Jordan is recognized as the most expensive of all athlete endorsers with his total income from endorsements alone reaching a staggering US$ 40,000,000. (Lane and Spiegel, 1996)
It was reported in 1997 that "Jordan was endorsing the products of General Mills (Wheaties), McDonald's (Quarter Pounders, Value Meals), Nike (Air Jordan), Quaker Oats (Gatorade), and Sara Lee (Hanes Underwear)". (Mathur, Mathur and Rangan, 1997, p.67) As a result of endorsing such an array of products, some related to his sport, such as Nike and others totally unrelated (E.g. McDonalds) some researchers would question his effectiveness. The findings of authors such as Tripp, Jensen and Carlson (1994) suggest that a celebrity such as Jordan, who endorses more than 1 product would be less effective. "The number of products a celebrity endorses negatively influences consumers perceptions of endorser credibility and likability, as well as attitudes towards the ad" (Tripp et al, 1994, p.543). This is a major factor for consideration by organisations when selecting celebrity endorsers.
The person who is involved in communicating the marketing message in either a direct or an indirect manner is know as the source. (Belch & Belch 1995). This process of social influence "results in an individual adopting the attitude advocated by the communicator: compliance, identification and internalization" (Kamins, 1989, p.35) and this is what gives a source their influence. Compliance relates to social acceptance and approval and is therefore not relevant in non-personal communication. This study concentrates on the use of a celebrity as the source, therefore the influence either through identification and internalization are two areas of importance. "Celebrity endorsements are expensive for firms" (Agrawal and Kamakura, 1995, p.56) and therefore there must be careful consideration taken in the selection process. There are many differing views on what characteristics make a celebrity endorser a good source.
The most recognized way to evaluate a celebrity's ability to be an endorser or spokesperson is source credibility. "Most of the research on source credibility has focused on the expertise and trustworthiness of the communicator" (Homer and Kahle, 1990, p.31), but authors such as Ohanian (1991), also included physical attractiveness in the analysis. Ohanian's (1991) study on how the perceived image of a celebrity endorser affected consumer's intention to purchase found that "only the perceived expertise of the celebrities was a significant factor explaining the respondents intention to purchase". (Ohanian, 1991, p.51).
Ohanian (1991) analyzed celebrity endorsers on three measures, to examine which characteristics of the source would be most effective in influencing purchase intention. The three areas that were used in the survey were physical attractiveness, trustworthiness, and expertise. These characteristics of the source were related to each celebrity to investigate their effectiveness in creating purchase intention. The results of the study found that trustworthiness and physical attraction had little effect on the purchase intention of the sample and it was the perceived expertise, which has the most significant impact on consumers. (Ohanian 1991, p.50) The use of a physically attractive endorser does have its place. Kamins suggest "the use of an attractive celebrity spokesperson appears to be effective for a particular category of product, those which are attractiveness related" (1990, p.11). This can also be linked to how related the endorser is to the product endorsed.
O'Mahony and Meenaghan (1997/98) supported Ohanian's (1991) findings. They concluded that "the source characteristic with the greatest impact on the consumers' intention to purchase were those of credibility and expertise. This study also concurs with many other academic works in the area.
There are many factors that effect the success or failure of a celebrity endorser. Credibility of the source is recognized as a key element and the characteristic can be effected by multi product endorsement. Multi-product endorsement is when a celebrity endorses more than one product at the same time. The most famous example of a celebrity who endorses a number of products or brands (as sited earlier) must be Michael Jordan. Research has shown that as the number of products endorsed decreases consumer perceptions of a celebrity's credibility and likability, in addition attitude towards the advertisement becomes less favorable. (Tripp et al 1994) With credibility as the most important feature of the source model, using a celebrity in advertising that endorses multiple products can be very harmful. "Simply knowing that a celebrity endorses multi products is sufficient to erode consumer perceptions of endorser trustworthiness." (Tripp et al 1994, p. 535)
Further evidence that indicates the importance of endorser credibility came from the same study by Tripp, Jensen and Carlson. Their second and third hypotheses looked at the effect of multi product endorsement and it's effectiveness over a number of exposures. Using in-depth interviews or qualitative research techniques, Tripp, Jensen and Carlson (1994) found that "consumers who are actually exposed to the multi product endorsers ads may perceive the celebrity as less trustworthy, expert and likeable" (Tripp et al 1994, p. 544). Importantly this study found that consumers' interpretation of credibility was effected by exposure. The study recognizes it's own failing in relation to multiple exposure. "Taken as a whole, the interviews suggest that consumers do not actively process information regarding how many products a celebrity endorses" (Tripp et al, 1994, p.543). So, although they draw conclusions relating to multiple product endorsement, exposure and the individual nature of the subjects, Tripp et al recognize the fact they made these evaluation points salient.
Consumers are individuals and therefore the majority of them are responsible for forming personal opinions. This is the case, but other studies, including Ohanian found that "gender and age of respondents had no significant effect on their intention to purchase or on how they evaluated attractiveness, trustworthiness and expertise " (1991, p.51).
Both celebrities and non-celebrities can generate the source of a marketing communications message. The use of "famous people holds the viewer's attention" (Dyson and Turco, 1998, p.1), but is it a more effective marketing communications method, and does a celebrity source add more to the message? Kamins (1989) looked at the use of celebrity versus non-celebrity as a spokesperson.
Kamins (1989) undertook a study, which evaluated using a two-sided communications approach as a method of adding credibility to endorsers. Kamins established that "celebrities are well liked" (1989, p.36) but what other techniques could be extended from traditional advertising to make them more effective. The study was working on the premise that is well supported in endorsement theory, that a more credible source would increase purchase intention. From this it was hypothesized that producing and advertisement, giving both positives and negatives about a product or brand would increase credibility of the celebrity endorser. (Kamins, 1989) "Research has shown that two sided appeals were more effective than one sided communication" (Kamins 1989, p. 36) in non-celebrity source situations. This was because it increased overall perceptions of credibility.
What Kamins found was that there was an "enhancement in brand attitude through the use of a two sided relative to one sided appeal" (1989, p.41). The success of the two-sided approach to celebrity endorsement came from the linking of this advertising technique which increased two recognized source characteristic of likability and trustworthiness. Apart from source credibility there is another element which makes an endorser effective.
As mentioned earlier Tripp et al (1994) suggest that it is a great disadvantage to have an endorser whom acts as a spokesperson for multiple products or brands. Authors such as Dyson and Turco (1998) echo this sentiment. "Getting a celebrity to represent their product exclusively" (Dyson and Turco, 1998, p.2) should be of the utmost priority of any organisation. Many examples can be sited for multiple brand endorsers and it is even common place for endorsers to switch to rival brands.
What effect does this have on the consumer? Remember, that in the end it is the consumer whom is most important and it is the consumer who must be convinced by the marketing effort, either with or without a celebrity to purchase a product. Gotlieb, Schlacter and St Louis (1992) look at source credibility and its effects on consumers. They overlooked the specifics of who takes on the role of providing the information as the source. But comparison can be drawn when looking at non-celebrity source. Research has "suggested that credibility of information source mediates how consumers perceive and interpret stimuli" (Gotlieb et al 1992, p.193). Gotlieb et al advocates that source credibility was a significant variable in consumers' response to an advertising message. (Gotlieb et al 1992). Gotlieb et al made the distinction between high and low involvement products and how source credibility affects consumers differently depending on the product being endorsed. When the product is high involvement the endorser must have higher credibility. When there is an effective source the high price of the item is evaluated in a more positive light.
Further more "given the credibility of the information source, the consumer converts the actual price into a psychological price" (Gotlieb et al 1992, p.195) which can encourage them to pay a higher price. What can be inferred is that consumers are more inclined to purchase a higher price product if the source is considered to be more credible. When low involvement products are considered, consumers can be persuaded to change their purchase on the advice from a more likeable source. This is because in this case, credibility is more socially based than economically based. In other words, the consumer wishes to identify with the source and therefore the small differences in price have little or no effect on the consumer (Gotlieb et al 1992). In addition, the study found that the greater the credibility of the source the easier it is to alter the purchase decision of the consumer.
O'Mahony and Meenaghan (1997/98) found consumers response to endorsement messages was also linked to relatedness. This supports previous findings relating to athletes and sporting products. Moreover, it can be generalized to other celebrities including musicians and actors. O'Mahony and Meenaghan confirmed "that consumers expected congruity between the celebrity endorsers' perceived images and the type of product they endorse" (1997/98, p.23).
The two basic source models of credibility and attractiveness when analyzed in conjunction give an accurate picture of their effect on purchase intention, but what effect do environment influences and factors outside the control of marketers have on source credibility. The major area of concern is negative media coverage of celebrities' actions (e.g., OJ Simpson). Till and Shimp have most recently examined "the impact that negative information about celebrities might have on consumers evaluations of endorsers brands" (Till and Shimp, 1998, p.67). What was discovered was that the negative information, which related to an endorser could adversely effect the consumers' evaluation of their particular brand or brands.
Till and Shimp (1998) took a very psychological approach to celebrity endorsement. They suggested that "when knowledge structures for a brand and/or celebrity are more fully developed, a brand may be somewhat insulated from negative press about endorsing celebrities" (Till and Shimp, 1998, pp.72 - 73). What Till and Shimp really tried to achieve was to broaden the knowledge base related to celebrity endorsement issues and put them "within the context of an associated memory framework (1998, p.79).
To tie the theories together it is important to analysis celebrity endorsement from another perspective. McCracken (1989) looked at the celebrity endorsement process from its cultural foundations. McCracken (1989) did endorse that credibility was an important source factor in creating the appropriate advertising message. McCracken " contends that a message depends for its effectiveness on the expertise and trustworthiness of the source" (1989, p. 310). The study analyzed cultural meanings that are attached to celebrity endorsers. McCracken (1989) believes that the source model makes no provisions for a celebrity not being compatible to a product and it allows any celebrity to be a persuasive source for any advertising message. (McCracken 1989) Where downfalls appear in the argument is that the credibility model has sub sections that can be linked to relatedness. The expertise criteria suggests that if a person has no experience with a product or it is not related to them in any way then they will not make an effective source.
The effectiveness of a celebrity endorser depends in part upon the meaning he or she brings to the endorsement process (McCracken 1989). Demographic characteristics are relatively easy to establish but it is more subjective categories such as expertise, likability and trustworthiness which are used more regular to determine the effectiveness of a source. Therefore Ohanian (1991) and others have better explained the endorsement process. McCracken (1989) put the cultural foundation of endorsement theory into perspective when analyzing the social implication of internalizing the qualities of unattractiveness, untrustworthy and unbelievable endorser.
Purpose of the Study
With the current pool of literature relating to celebrity endorsement it is important to retest the models in a different geographic context. Using a random selection of two endorsers their personal attributes will be tested in relation to consumer perceptions.
H1: A more credible endorser will create a higher intention to purchase.
H2: The credibility of an endorser is of greater importance in terms of purchase intention than the number of product they endorse.
The quantitative research was conducted using a stimulus and the questionnaire. This experiment will therefore control the possibility of any interference.
The research was implemented using a convenience sampling technique. The whole sample (n = 108) was divided into two homogeneous sub groups (n = 54) with equal numbers in each. The sample was drawn from an undergraduate class at Griffith University. The sample consisted with a relatively even distribution of males (n = 50) and females (n = 58). The majority of the sample were under the age of 24 years (n = 44). These give an indication of the breakdown of the sample demographics. It is important to remember that because of the sampling technique that was used "we cannot measure sampling error, and clearly cannot make any definitive statements about the results from such a sample" (Kinnear, Taylor, Johnson and Armstrong, 1996, p.289).
The data was collected using two surveys. The questions were identical for both sample groups except for the celebrity endorser they were evaluating. The first sample evaluated Michael Richards (Kramer from Seinfeld) with his endorsement of Vodafone. The second sample looked at Michael Jordan (Chicago Bull Basketballer) and his endorsement of Gatorade, a sports drink. The celebrities were each selected randomly from an accessible list of single product endorsers (Kramer) and multi product endorsers (Jordan).
The questionnaire consisted of 15 questions. Their were12 likert scales 1 metric scale to find out the number of products endorsed and 2 nominal scales for the demographic characteristics. The questionnaire covered the areas of attractiveness, trustworthiness, expertise, purchase intention, number of products endorsed and demographics. The categories of attractiveness, trustworthiness, expertise and purchase intention were collected in accordance with Ohanian (1991). The quantity of products endorsed linked to Tripp, Jensen and Carlson (1994).
The measurement scales used on the questionnaire were all validated by earlier research. They were selected for each of the required categories from Bruneril and Hensel (1996), a recognized source of marketing research scales. The questions were worded exactly the same for both samples, except for the endorsers name. This was to ensure that there was no measurement error in the questionnaire.
The data was analyzed in SPSS. The data was coded into one file with the ability to select either group for specific analysis. The whole data set was used to get data such as the descriptive about the sample (e.g. age and gender). For the comparisons, the specific cases were selected and statistical procedures were carried out, ensuring the correct statistical technique was used for the specific data (e.g. means were only calculated on data interval or higher).
To determine if the endorsers met the source characteristics of expertise, attractiveness and trustworthiness, summated means were calculated. To allow for the summation a factor analysis was carried out on each variable. Each factor loaded on a single component.
With a possible range 1 being strongly agree to 5 being strongly disagree Jordan rated on the agree side of neutral on two of the three salient attributes. Attractiveness (8 = 3.0123), trustworthiness (8 = 2.9506) and expertise (8 = 2.6543) for Jordan were all categories rating substantially better than for Kramer. Kramer was consider by the sample to be less attractive (8 = 3.9753) less experienced (8 = 3.3333) and less trustworthy (8 = 3.4074). An ANOVA was calculated to determine what were significant factors influencing purchase intention. The results in table 1 show that all source credibility factors as well as the celebrity are influencing factors.
Table 1: Analysis of Variance (ANOVA)
*Significant at a level of .05
To examine further difference between the celebrities a chi-square was calculated for each credibility factor. In favour of Jordan, their was a significant difference at the point .05 level on expertise (sig. = .000) trustworthiness (sig. = .000) and attractiveness (sig. = .000)
Table 2: Chi Square results
*Significant at a level of .05
The purchase intention scale was set up with a 1 to 5 range where 1 represented no definitely not purchase and 5 representing yes definitely purchase. Although neither of the endorsers got positive reactions to the purchase of the products they endorsed, Jordan (8 = 2.4813) rated significantly higher than Kramer (8 = 2.0987). Again to allow for the summation a factor analysis assigned the three questions to one component.
These figures agree with authors such as Ohanian (1991) and O'Mahony and Meenaghan (1997/98) and support H1. Even though the purchase intention figures are low, the sample appears to have taken into account that Jordan is a more attractive, trustable and expert source than Kramer is. This is an expected result.
The second area for evaluation was the effect consumers' knowledge regarding the number of products that the endorser endorses. Earlier research by Tripp, Jensen and Carlson (1994) showed that the optimum number of products one should endorse is one. Dyson and Turco (1998) supported this in their urging for exclusivity. What would be expected is that the expertise and trustworthiness of a multi product endorser would be lower than that of a single product endorser.
The findings of this study suggest otherwise. Jordan endorses many more products than Kramer does in the samples geographic area. Yet Jordan rated higher on all the credibility categories. The sample recognized that Jordan endorsed multiple products (8 = 3.4185) which was significantly higher than Kramer, (8 = 1.3704) who actually only endorses one product and yet Jordan had higher purchase intention.
A second ANOVA was calculated to assess if the number of products that a celebrity endorses effected purchase intention. The results suggested that at a significance level of .05, multiple product endorsement has no effect on purchase intention (sig. = .244).
These findings support H2 that consumers put more emphasis on the credibility of an endorser than the number of products that he/she endorses. This is important when looking at whom to select as a celebrity endorser. The perceptions of a sample in relation to numbers of products endorsed are also very important. Jordan has endorsed at least 5 different products in the Australian market yet he is only recognized to endorse 3.4815, this suggests that people may link an endorser to a certain number of products if they endorse more. Conversely, if an endorser such as Kramer only endorses one product he may be mistakenly linked with more brands. The sample believed that Kramer endorsed 1.3704 with 14.8% of the sample under the impression that he endorsed 3 products. Kramer only endorses one product.
The present study has been designed to look at source factors, and how they relate to purchase intention. The study considered the source personal factors of attractiveness, trustworthiness and expertise, as well as the number of products that they endorsed. The random selection of the subject stimuli gave two totally opposite celebrities and the sample rated them accordingly on all measures.
It was discovered that the source personal characteristic appeared to be given higher weighting by the sample and linked closer to purchase intention rather than the number of products endorsed. This contrasts the previous studies by Tripp, Jensen and Carlson, (1994) and Dyson and Turco (1998) which suggested that it was necessary to have an endorser who exclusively endorses one organisations products.
The first hypothesis supported many previous academic works, but the second study provided some more interesting insights. The fact that a known multi product endorser rated highly on a factor such as trustworthiness contrasts other analysis. Michael Jordan was rated high as a communications source and his purchase intention was higher than a single product endorser.
The empirical finds on Jordan's worth as a celebrity endorser was supported by economic reports relating to his value. It was reported that when Jordan announced his comeback to professional basketball that the companies to whom he was contracted as endorser had their "stock value sky-rocket: General Mills +US$500 million, McDonalds +US$1 billion, Nike +US$200 million, Quaker Oats +US$100 million, Sara Lee +US$500 million " (Arace, 1995, p.29).
Knowles, Mathur and Rangan, (1997) who completed an event study analysis into Michael Jordan and supported these figures. They found "compelling evidence that a major celebrity endorser has the potential to profoundly influence the profitability of an endorsed product" (Knowles et al, 1997, p.72). These finds add to the endorsement literature by suggesting that a celebrity may still be an effective endorser even if they endorse many products. It must also be remembered that they must be a credible source to begin with.
The importance of this study is that it suggests that a highly credible source can also be a multi product endorser without negatively effecting the purchase intentions of a consumer. "Previous research studies investigating the use of celebrities in advertising have shown ads to vary in effectiveness for such dependent measures as believability and a purchase intention as a function of the product type investigated" (Kamins, 1989, p.34). This study has shown that consumers place higher weighting on different factors depending on the endorser.
Limitations of the Study
The study is a step in the right direction. It provides some valuable insights, but is only the beginning. It is very important to recognize the limitations of the study when interpreting the findings. The sample was not generalizable to the population. The results would need to be replicated on a more representative sample.
In addition the stimuli used is not representative of enough endorsers or product categories. The only definitive statements that could be made from the research relate to Jordan and sports drinks and Kramer and mobile phones. The future direction in research in this area would need to use a wider variety of endorser in a greater number of product categories.
Despite it limitation, the study adds valuable information to the field of endorsement literature. Given the amount of celebrities whom endorse products it is a valuable area of study and this research has added new areas for consideration. It is very important that future work in the area consider source credibility and multi product endorsement simultaneously.
Agrawal, J., and Kamakura, W., (1995) 'The Economic Worth of Celebrity Endorsers: An Event Study Analysis' Journal of Marketing, Volume 59 July, pp. 56 - 62
Arace, M., (1995) 'Michael Jordan: The Second Coming' One On One: Australia's official NBA Magazine, Volume 4 Number 5 May 1995, pp. 16 - 37
Belch, G. & Belch, M. (1995) 'Introduction to Advertising and Promotion: An Integrated Marketing Communications Perspective, Third Ed' United States of America: Richard D. Irwin INC.
Bruneril, C., and Hensel, P., (1996) ' Marketing Scales Handbook: A Compilation of Multi-item Measure Volume 2, American Marketing Association, Chicago, Illinois, USA
Dyson, A., and Turco, D., (1998) 'The State of Celebrity Endorsement in Sport', The Cyber Journal of Sport Marketing, Volume 2, Number 1, Retrieved September 1st 1998 from the World Wide Web
Gotlieb, J., Schlacter, J., and St Louis, R., (1992) 'Consumer Decision Making: A model of the Effect of Involvement, Source Credibility and Location on the Size of the Price Difference Required to Induce Consumers to Change Suppliers' Psychology and Marketing, Volume 9 Number 3, pp.191- 208
Kamins, M., (1989) 'Celebrity and Non Celebrity Advertising in a Two Sided Context" Journal of Advertising Research, June/July pp. 34 - 42
_________, (1990) 'An Investigation into the Match-Up Hypothesis in Celebrity Advertising: When Beauty May be only Skin Deep' Journal of Advertising, Volume 19, Number 1, pp. 4 - 13.
Kinnear, T., Taylor, J., Johnson, L. and Armstrong, R., (1996) 'Australian Marketing Research' McGraw Hill, New South Wales, Australia
Lane, R., and Spiegel P., (1996) 'The Year of Michael's' Forbes Magazine, December 16th 1996, Retrieved 22nd September 1998 from the World Wide Web http://www.forbes.com/forbes/121696/5814244a.htm
Mathur, L. K., Mathur, I. and Rangan, N., (1997) 'The Wealth Effects Associated with a Celebrity Endorser: The Michael Jordan Phenomenon' Journal of Advertising Research, May - June, pp. 67 - 73.
McCracken, G., (1989) 'Who is the Celebrity Endorser? Cultural Foundations of the Endorsement Process' Journal of Consumer Research, Volume 16 December, pp. 310 - 321
Ohanian, R., (1991) 'The Impact of Celebrity Spokespersons' Perceived Image on Consumers Intention to Purchase" Journal of Advertising Research. February/March, pp. 46 - 54
O'Mahony, S., and Meenaghan, T., (1997/98) 'The Impact of Celebrity Endorsement on Consumers' Irish Marketing Review, Volume 10 Number 2, pp.15 - 24
Thompson, J, W., (1998) 'What Price Fame? $800 Million', IEG Endorsement Insider, Retrieved September 22nd 1998, from the World Wide Web http://www.endorsements.com/0101/frame.html
Till B., and Shimp, T., (1998) 'Endorsers in Advertising: The case of Negative Celebrity Information' Journal of Advertising, Volume XXVII, Number 1, Spring pp. 67 - 82
Tripp, C., Jensen, T. and Carlson, L., (1994) 'The Effect of Multiple Product Endorsements by Celebrities on Consumers' Attitudes and Intentions' Journal of Consumer Research, Volume 20 March pp. 535 - 547
All material presented in The Cyber-Journal of Sport Marketing is copyright unless otherwise stated. For academic and personal use, The Cyber-Journal of Sport Marketing's papers may be downloaded, or read, free of charge. Material published in the The Cyber-Journal of Sport Marketing may not be further reproduced, except as described in the previous sentence, sold or published by use of any existing or future media without the express permission of the executive editor. The Journal is registered as a journal (issn:1327-6816) with the Australian National Library and the ISSN International Centre in Paris.
This is an archive copy of a web document originally published in the Cyber-Journal of Sport Marketing. All copyright remains with the creator.
National Sport Information Centre Web Archive