I. Introduction
The practice of celebrity endorsement has been a worldwide phenomenon. The use of celebrity endorsers in the US accounts for about 25 percent of TV commercials[1], while, in Japan and Korea[2][3], more than 70 percent of TV commercials feature one or two celebrity endorsers. Regardless of product or service categories, celebrity endorsers are likely to appear in TV commercials. Since marketers want to select a celebrity endorser or endorsers with credibility, attractiveness, favorability, expertise, or popularity, the celebrity endorser has frequently been shared by many companies in a form of multiple celebrity, or product, endorsement. For instance, Gillette has employed famous athletes, such as Rodger Federer (a tennis player), Thierry Henry (a soccer player), and Tiger Woods (a golfer), to promote its Fusion men’s razor.
Thus, the objective of the current study is to investigate which factors influence the effects of multiple celebrity endorsement. Three factors, such as portfolio-brand fit, portfolio-celebrity fit, and portfolio celebrity-consumer fit, will be examined as predictors which affect a consumer’s attitude toward the ad, the endorsed brand, and the purchase intention in the context of multiple celebrity endorsement. According to Yu and Pappu[4], portfolio-brand fit refers to “the fit of the multiple elements in a campaign portfolio with a single brand. Portfolio-celebrity fit refers to the fit among celebrity endorsers collectively involved in an instance of multiple celebrity endorsement[4].
Lastly, in a similar vein, portfolio-consumer fit refers to the extent with which consumers identify with celebrity endorsers in the multiple celebrity endorsement context. The current study is deemed significant in the sense that this research will further broaden our understanding of the effectiveness of multiple celebrity endorsement.
To date, research on effects of multiple celebrity endorsement has been scarce. Since use of multiple celebrity endorsement is prevalent in Korea, study findings may provide advertising practitioners with practical implications. A conceptual model, which explains factors in the effectiveness of multiple celebrity endorsement, will be tested in this study.
II. Literature Review
1. Effects of Celebrity Endorsement
Agrawal and Kamakura[5] measured the economic worth of celebrity endorsers and found that companies garnered a significant percentage of positive abnormal returns. According to Atkin and Block[6], people perceive that celebrity endorsers are more attractive than non-celebrity endorsers. Furthermore, their study suggests advertisements featuring celebrity endorsers can lead to greater purchase intention and brand evaluation than featuring non-celebrity endorsers.
In a similar vein, Petty, Cacioppo, and Schumann[7] also found that people tended to prefer products endorsed by celebrity endorsers than by non-celebrity endorsers. Lee and Kim[8] found that the effects of brand trait transference were more notably observed in the cases of a non-celebrity endorser than in celebrity endorser. Kim and Cho found [9] that pet naming types using celebrity endorser are effective.
Hsu and McDonald[10] found that fit between the endorsed product and various celebrities is a key factor for using multiple celebrity endorsers in advertising. In their study they found that multiple celebrity endorsement an effective advertising strategy. In terms of effects of celebrity endorsement on political decision making process, Jackson and Darrow[11] found that celebrity endorsement could be effective in changing voters’ political opinions.
2. Portfolio-Brand Fit: Fit between Multiple Celebrity Endorsers and Brand
Yu and Pappu[4] define portfolio-brand fit as “the fit of multiple elements in a campaign portfolio with a single brand.” The portfolio-brand fit could be interpreted as how well the consumer perceives the congruence between the celebrity endorsers as a whole and their endorsed brand. In short, portfolio-brand fit is an extension of the ‘match-up hypothesis,’ which suggests that the effectiveness of celebrity endorsement depends on the fit/congruence between a celebrity endorser and the endorsed brand[12]. The match-up hypothesis, according to Rossiter & Percy[13], suggests that visual imagery conveys information over and above the information contained in explicit verbal arguments.
From a multiple celebrity endorsement perspective, it is plausible to assume that the fit between multiple celebrity endorsers and their endorsed brand may have a positive impact on attitude toward multiple celebrity endorsement, resulting in a positive evaluation of the ads and the endorsed brand. A recent study [4] found that portfolio-brand fit positively affects attitude toward multiple celebrity endorsement; consequently, leading to a favorable attitude toward the brand. Previous research also found that multiple celebrity endorsement had a positive impact on consumers’ brand evaluation [14]. The above discussion leads to the following hypothesis:
H1: Portfolio-brand fit has a positive impact on attitude towards the ad.
3. Portfolio-Celebrity Fit: Fit among Multiple Celebrity Endorsers
Independent of the endorsed brand, portfolio-celebrity fit refers to the fit among celebrity endorsers collectively involved in an instance of multiple celebrity endorsement[4]. Yu and Pappu [4] suggest that “the advantage of the celebrity portfolio perspective is its consideration of the interdependencies among the brand and multiple celebrity endorsers with respect to consumer perceptions (p. 158).” Congruence among multiple celebrity endorsers may influence how consumers attribute their reasons for endorsing a brand.
According to Heider[15], people tend to attribute the causal structure of events to the environment or to something within the person involved in the event. When individuals evaluate celebrity endorsement, they are likely to make either an internal attribution in which they believe that the endorser is recommending the product because of its good characteristics, or an external attribution in which they believe that the endorser is recommending the product because they are paid for it[10].
For instance, when three athletes (Rodger Federer, Tiger Woods, and Thierry Henry) endorse the Fusion men’s razor, consumers may perceive them as ‘congruent’ in terms of their images. Under this condition, consumers may make an internal attribution, thinking that the reason for their endorsing a razor is the good quality of the brand. However, if an athlete (Rodger Federer), an actor (Leonardo DiCaprio), and a singer (Adam Levine; Maroon 5 main vocalist) endorse the same razor, consumers may feel an ‘incongruence’ in terms of their images, irrespective of their world-wide popularity. Thus, when consumers attribute multiple celebrity endorsement internally rather than externally, they are likely to have a more favorable attitude toward celebrity endorsement, positively affecting their attitude toward the ads and the brand. Thus, the above discussion leads to the following hypothesis.
H2: Portfolio-celebrity fit has a positive impact on attitude toward the ad.
4. Portfolio-Consumer Fit: Fit between Multiple Celebrity Endorsers and Target Audiences
Portfolio-consumer fit refers to the extent to which consumers feel ‘matched’ with celebrity endorsers in a multiple celebrity endorsement context. With regard to celebrity endorser selection criteria, a study conducted with advertising practitioners proposed that practitioners view ‘celebrity-target audience match’ as the most important criterion, followed by ‘celebrity-brand/product match’ [16].
Celebrity-target audience match may be understood as how similar the celebrity is to the target audience. In this sense, identification theories can be applied to explain how the match between celebrity endorsers and target audiences influence consumers’ evaluations of celebrity endorsement ads and the endorsed brand.
Burkes’ dramatism theory[17], Kelman’s theory of opinion change[18], and Bandura’s social cognitive theory[19], all emphasize the role of identification in changing individuals’ attitudes and behavior. The para-social relationship, according to Horton and Wohl[20], occurs when individuals, via the media, develop a sense of intimacy and identification with a celebrity. This perspective indicates that identification is a natural outcome of the communication situation.
People who are highly identified are more likely to adopt the thoughts, feelings, and behaviors advocated by that celebrity than those who are lowly identified. Prior research also found that the more people identified with a celebrity, the more likely they were to credit that celebrity with positive attributes. Identification with celebrities brings about a desire to change the aspects of one’s own attitudes, beliefs, and behaviors, to more closely emulate those of celebrities[21][22].
H3: Portfolio-consumer fit has a positive impact on attitude toward the ad.
H4: Attitude toward the ad will have a positive impact on attitude toward the brand.
H5: Attitude toward the brand will have a positive impact on purchase intention.
III. Method
1. Study Design
To test the proposed hypotheses, three single celebrity endorsement ads were created, with the help from a professional advertising practitioner, as stimuli for this study. Of two types of multiple celebrity endorsements (the simultaneous or sequential type of endorsement), the current study took the format of the sequential type of multiple celebrity endorsement since this type of multiple celebrity endorsement is more frequently observed in the real world. Each celebrity endorsement ad features one of three celebrities chosen for this study. The layout, headlines, body copy and visual elements were identical in the three celebrity endorsement ads. Three celebrities, voted as the most favorable celebrities among coffee brand celebrity endorsers by Tillion Panel[23], were chosen for the current study. In terms of the choice of product, coffee was selected as a product because coffee consumption is reported high among 20-something aged individuals in Korea. A fictitious coffee brand name was created for the study.
2. Data Collection Procedure
To create an online survey, this study utilized Qualtrics, a web-based survey tool. The web-based survey opened with an informed-consent notice; participants were then asked to click on the “proceed” button if they agreed to participate in the study. Then, participants were exposed to multiple celebrity endorsement print ads (three single celebrity endorsement ads). Following that, they were asked to answer several questions regarding portfolio brand fit (i.e., congruence between three celebrity endorsers and their endorsed brand), portfolio celebrity fit (i.e., congruence among three celebrity endorsers), and portfolio consumer fit (i.e., congruence between three celebrity endorsers and consumers). Finally, they were asked to answer demographic questions such as age, gender, years in college, and so forth.
3. Measure
To measure portfolio-brand fit, portfolio celebrity fit, and portfolio consumer celebrity fit, Rifon, Choi, Trimble, and Li’s[24] scale from the sponsorship research was adopted and modified accordingly. Each variable was measured, using three, 7-point semantic differential scales: not compatible/compatible, not a good fit/good fit, and congruent /incongruent. Portfolio brand fit was measured once for each portfolio and portfolio celebrity fit was measured for each celebrity included in the portfolio. Lastly, portfolio consumer fit was measured once for each portfolio. Attitude toward the ad was measured using three, 7-point semantic differential scales: very bad-very good, very unfavorable-very favorable, like very much-dislike very much[25]. Attitude toward the brand (Ab) was measured using five, 7-point semantic differential scales: unappealing-appealing, bad-good, unpleasant-pleasant, unfavorable-favorable, and unlikable-likable. To measure purchasing intentions, the study used five items–never/definitely, definitely do not intend to buy/definitely intend to buy, very low/high purchase interest, definitely not buy it/definitely will buy it, and probably not buy/probably will buy. To assess purchase intention, the following statement was given to subjects: “If you were in the market today for this product/brand, how likely do you feel it is that you would purchase/use this product /brand?”[26].
IV. Results
1. Sample Profile
For this study, a total of 375 subjects participated. However, after removing 21 incomplete responses, 354 subjects (university students) remained for further analysis. Of 354 subjects, 39.5% (n = 140) were male and 60.5% (n = 214) were female. Their mean age was 22.2 years old. Sophomores made up the majority (77.4%, n = 274); the rest were juniors (17.5%, n = 62), and seniors (5.1%, n = 18).
2. Hypotheses Tests
The relationships between the celebrity-brand fit portfolio, celebrity-consumer fit portfolio, celebrity-celebrity fit portfolio, attitudes toward multiple celebrity endorsement, attitudes toward the ad, attitudes toward the brand and their correlates are shown in [Table 1]. The correlation results indicate significant relationships among measured variables. To test the structural model concerning the relationships among the variables, the path analysis was performed via SPSS AMOS 21.0. [Figure 1] shows the original model. As [Table 2] indicates, the overall fit indices for the model were not acceptable, revealing a weak fit of the model to the data (x2 = 35.70, df = 10, p < .001; NFI = .60; GFI = .77; CFI = .60; SRMR = 1.24; RMSEA = .314).
Table 1 Correlation Matrix, CR, AVE, and Descriptive Statistics for the Measurement Model
Notes: Bold, values on the diagonal of the correlation matrix are the square root of AVE. CR, construct reliability; AVE, average variance of extracted.
Table 2 Parameter estimates for casual paths: Original Model
Goodness-of-fit statistics x2 = 35.70, df = 10, p<.001; NFI=.60; GFI=.77; CFI=.60; SRMR=1.24; RMSEA=.314
*p<.05, **p<.01, ***p<.001
Figure 1 Original Model
Goodness-of-fit statistics x2 = 35.70, df = 10, p<.001; NFI=.60; GFI=.77; CFI=.60; SRMR=1.24; RMSEA=.314
*p<.05, **p<.01, ***p<.001
In this study, the model fit was judged on the basis of five primary criteria – Normed Fit Index (NFI), Comparative Fit Index (CFI), Goodness of Fit Index (GFI), the Standardized Root Mean Squared Residual (SRMR), and the Root Mean Squared Error of Approximation (RMSEA). A model is regarded acceptable if NFI and GFI exceed .90 and CFI exceed .93, and when SRMR is less than .06 [27] and RMSEA is less than .08 [28][29].
Thus, the original model was rejected and the modification indices were examined as a way of improving the model fit[30]. The modification indices showed that the model fit could be improved by adding covariance paths between the following: portfolio brand fit and portfolio consumer fit, portfolio brand fit and portfolio celebrity fit, and portfolio consumer fit and portfolio celebrity fit. After the model modification, the goodness of fit statistics demonstrated that the modified model provided a better fit (x2 = 5.73, df = 7, p < .001; NFI = .96; GFI = .97; CFI = .96; SRMR = .062; RMSEA = .078). The NFI, the GFI, and the CFI suggest a good fit, and the RMSEA suggests it is marginally acceptable. In addition, the SRMR also suggests a good fit. [Figure 2] shows the modified model and [Table 3] indicates the parameter estimates for paths.
Figure 2 Modified Model
Goodness-of-fit statistics x2 = 5.73, df = 7, p<.001; NFI=.96; GFI=.97; CFI=.96; SRMR=.062; RMSEA=.078
*p<.05, **p<.01, ***p<.001
Table 3 Parameter estimates for casual paths: Modified Model
Goodness-of-fit statistics x2 = 5.73, df = 7, p<.001; NFI=.96; GFI=.97; CFI=.96; SRMR=.062; RMSEA=.078
*p<.05, **p<.01, ***p<.001
H1 posits that portfolio-brand fit has a positive impact on attitude toward the ad. According to Anderson and Gerbing (1998), a t-value of greater than 2 for each coefficient indicates a statistical significance. Study results show that a portfolio-brand fit had a significant relationship to attitude toward multiple celebrity endorsement (t = 5.0, p < .001). H2 proposes that portfolio-celebrity fit has a positive impact on attitude toward multiple celebrity endorsement. Study results show that portfolio-celebrity fit had a significant relationship to attitudes toward multiple celebrity endorsement (t = 2.28, p < .05). H3 states that portfolio-consumer fit has a positive impact on attitudes toward multiple celebrity endorsement. As expected, the study results show that portfolio-consumer fit had a significant impact on attitudes toward multiple celebrity endorsement (t = 2.16, p < .05). H4 posits that attitudes toward an ad will have a positive impact on attitudes toward a brand. As shown in [Table 3], attitudes toward an ad are found to have a significant impact on attitudes toward a brand (t = 18.38, p < .001). Lastly, H5 suggests that attitudes toward a brand will have a positive impact on purchase intentions. As [Table 3] indicates, attitudes toward an ad are found to have a significant impact on purchase intentions (t = 15.91, p < .001). In sum, H1, H2, H3, H4, and H5 were supported in the study.
Three additional paths were added in the modified model. As shown in [Table 3], study results suggest that a portfolio brand fit-portfolio consumer fit path coefficient is .64 with a t-value of 6.28 (p < .001). This result indicates that portfolio brand fit positively influences portfolio consumer fit, and vice versa. The portfolio brand fit-portfolio celebrity fit path coefficient is 1.49 with a t-value of 11.05 (p < .001). This result indicates portfolio brand fit positively influences portfolio celebrity fit, and vice versa. Lastly, the portfolio consumer fit-portfolio consumer fit path coefficient is .72 with a t-value of 5.82 (p < .001). This result indicates portfolio consumer fit positively influences portfolio celebrity fit, and vice versa.
IV. Discussion
Marketers employ multiple celebrity endorsement in hopes that multiple celebrity endorsers could appeal to a broad spectrum of audiences. In short, it is vital to use multiple celebrities to communicate effectively with such a broad spectrum of audience. In the current study, three factors were assumed to influence the effects of multiple celebrity endorsement. First, fit between multiple celebrity endorsers and brand, termed as ‘portfolio-brand fit’ was examined as one important factor. As study results found, fit between multiple celebrity endorsers and brand is known to positively impact attitudes toward the ad. A practical implication may be drawn from this result. When employing multiple celebrity endorsement, it is imperative that advertising practitioners consider the fit between celebrity endorsers and the endorsed brand in a collective manner.
In addition, the fit among multiple celebrity endorsers, termed as ‘portfolio-celebrity fit’ was investigated as a factor influencing effects of multiple celebrity endorsement. Drawing upon attribution theory, it was suggested that consumers’ perceived congruence among multiple celebrity endorsers, in terms of their images, could lead to dispositional attribution rather than situational attribution. Consequently, the fit among multiple celebrity endorsement could result in positive impact on attitudes toward the ad. Thus, another implication could be drawn from this finding. When it comes to selecting multiple celebrity endorsers, the fit among multiple celebrity endorsers could be considered important criterion.
Lastly, the fit between multiple celebrity endorsers and target audiences, termed as ‘portfolio-consumer fit,’ was proposed as a factor which could influence the effect of multiple celebrity endorsement. As the study found, the fit between multiple celebrity endorsers and target audiences had a positive impact on attitudes toward the ad. Theoretically, this finding supports the notion that a celebrity’s similarity with a targeted audience could positively affect the target audiences’ thoughts, feelings, and behaviors. Practically, this suggests that advertising practitioners should consider ‘portfolio-consumer fit’ as a criterion for selecting celebrity endorsers for an advertising campaign.
To wrap up, a year-long advertising campaign for a brand or service, particularly one which employs multiple celebrity endorsers, needs more strategic considerations in terms of selecting celebrity endorsers. First, celebrities themselves need to be considered as a whole, not as a part. Second, their congregated images should be congruent with targeted audiences as well as with their endorsed brand or service if the advertising campaign is to communicate effectively. Third, fit between multiple celebrity endorsers and their endorsed brand needs to be considered when it comes to creating an ad campaign since the fit plays an important role in evaluation of ads and the endorsed brand.
References
- A. Stephens and A. Rice, "Spicing up the message," Finance Week, Vol.76, No.26, pp.46-47, 1998.
- D. Kilburn, "Star power," Adweek, Vol.39, No.2, p.20, 1999.
- H. Kim, "Big model," Ad Information, Vol.307, pp.40-42, 2006.
- M. Yu and R. Pappu, "Celebrity Portfolio Effects on Consumer Brand Evaluations," In Looking Forward, Looking Back: Drawing on the Past to Shape the Future of Marketing(pp.736-736), Springer, Cham, 2016.
- J. Agrawal and W. Kamankura, "The Economic Worth of Celebrity Endorsers: An Event Study Analysis," Journal of Marketing, Vol.59, No.3, pp.56-63, 1995. https://doi.org/10.1177/002224299505900305
- C. Atkin and M. Block, "Effectiveness of Celebrity Endorsers," Journal of Advertising Research, Vol.23, No.1, pp.57-61, 1983.
- R. E. Petty, T. J. Cacioppo, and D. Schumann, "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, Vol.10, No.2, pp.135-146, 1983. https://doi.org/10.1086/208954
- 이계명, 김희진, "광고모델로 이전되는 브랜드특성 전이효과에 관한 연구," 한국콘텐츠학회논문지, 제15권, 제12호, pp.493-503, 2015. https://doi.org/10.5392/JKCA.2015.15.12.493
- 김혜진, 조창환, "유명인을 이용한 제품 애칭화 유형에 따른 소비자 반응 연구," 한국콘텐츠학회논문지, 제17 권, 제1호, pp.532-549, 2017. https://doi.org/10.5392/JKCA.2017.17.01.532
- C. K. Hsu and D. McDonald, "An examination on multiple celebrity endorsers in advertising," Journal of Product & Brand Management, Vol.11, No.1, pp.19-29, 2002. https://doi.org/10.1108/10610420210419522
- D. J. Jackson and T. I. Darrow, "The influence of celebrity endorsements on young adults' political opinions," Harvard International Journal of Press/Politics, Vol.10, p.80, 2015.
- L. Kahle and P. Homer, "Physical attractiveness of the celebrity endorser: A social adaptation perspective," Journal of Consumer Research, Vol.11, No.4, pp.954-961, 1985. https://doi.org/10.1086/209029
- J. R. Rossiter and L. Percy, "Attitude change through visual imagery in advertising," Journal of Advertising, Vol.9, No.2, pp.10-16, 1980. https://doi.org/10.1080/00913367.1980.10673313
- D. H. Rice, K. Kelting, and R. J. Lutz, "Multiple endorsers and multiple endorsements: The influence of message repetition, source congruence and involvement on brand attitudes," Journal of Consumer Psychology, Vol.22, No.2, pp.249-259, 2012. https://doi.org/10.1016/j.jcps.2011.06.002
- F. Heider, "The psychology of interpersonal relations," Psychology Press, Hillsdale, New Jersey, 2013.
- B. Z. Erdogan, M. J. Baker, and S. Tagg, "Selecting celebrity endorsers: The practitioner's perspective," Journal of Advertising Research, Vol.41, No.3, pp.39-48, 2001. https://doi.org/10.2501/JAR-41-3-39-48
- K. Burke, "A rhetoric of motive," Englewood Cliffs, NJ: Prentice-Hall, 1950.
- H. Kelman, "Process of opinion change," Public Opinion Quarterly, Vol.25, No.1, pp.57-78, 1961. https://doi.org/10.1086/266996
- A. Bandura, "Social foundations of thought and action: A social-cognitive theory," Englewood Cliffs, NJ: Prentice-Hall, 1986.
- D. Horton and W. R. Richard, "Mass communication and para-social interaction: Observations on intimacy at a distance," Psychiatry, Vol.19, No.3, pp.215-229, 1956. https://doi.org/10.1080/00332747.1956.11023049
- M. D. Basil, "Identification as a mediator of celebrity effects," Journal of Broadcasting & Electronic Media, Vol.40, No.4, pp.478-495, 1996. https://doi.org/10.1080/08838159609364370
- W. J. Brown and M. D. Basil, "Parasocial interaction and identification: Social change processes for effective health interventions," Health Communication, Vol.25, No.6-7, pp.601-602, 2010. https://doi.org/10.1080/10410236.2010.496830
- D. W. Kang, "Gong Yoo, Song Jung-Gi, and D. W. Kang, High Model Preference for Coffee Brand," Retrieved in September 8, 2017 from http://mnb.moneys.mt.co.kr/mnbview.php?no=2016112013508063987.
- N. J. Rifon, S. M. Choi, C. S. Trimble, and H. Li, "Congruence effects in sponsorship: The mediating role of sponsor credibility and consumer attributions of sponsor motive," Journal of Advertising, Vol.33, No.1, pp.30-42, 2014.
- S. B. MacKenzie and R. J. Lutz, "An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context," Journal of Marketing, Vol.53, No.2, pp.48-65, 1989. https://doi.org/10.1177/002224298905300204
- N. Spears and S. N. Singh, "Measuring attitude toward the brand and purchase intentions," Journal of Current Issues & Research in Advertising, Vol.26, No.2, pp.53-66, 2004. https://doi.org/10.1080/10641734.2004.10505164
- K. G. Jöreskog amd D. Sorbom, "LISREL 8: User's reference guide," Scientific Software International, 1996.
- B. M. Byrne, "Structural equation modeling with EQS and EQS/Windows: Basic concepts, applications, and programming," London: Sage, 1994.
- M. W. Browne and R. Cudeck, "Alternative ways of assessing model fit," Sage focus editions, Vol.154, pp.136-136, 1993.
- J. C. Anderson and D. W. Gerbing, "Structural equation modeling in practice: A review and recommended two-step approach," Psychological Bulletin, Vol.103, No.3, pp.411-423, 1989. https://doi.org/10.1037//0033-2909.103.3.411