This item is the archived peer-reviewed author-version of:
Do I know you? How brand familiarity and perceived fit affect consumers' attitudes towards brands placed in movies
Reference: Verhellen Yann, Dens Nathalie, De Pelsmacker Patrick.- Do I know you? How brand familiarity and perceived fit affect consumers' attitudes towards brands placed in movies Marketing letters - ISSN 0923-0645 - (2015), p. 1-11 Full text (Publishers DOI): http://dx.doi.org/doi:10.1007/s11002-015-9347-0 To cite this reference: http://hdl.handle.net/10067/1226720151162165141
Institutional repository IRUA
1. Introduction In today’s oversaturated media environment, advertisers are increasingly resorting to brand placement, the (paid) inclusion of brands in media content as an alternative way to promote their brands. Although prior studies have unearthed some core dimensions underlying brand placement effectiveness, the practice is constantly evolving, and there is much more to learn about how brand placement operates. Brand placement effects are generally attributed to the connection that is forged between the placed brand and the editorial content (e.g., a movie) in which it is embedded (Russell and Stern 2006). One of the key characteristics of brand placements that influences their effectiveness is plot connection (Balasubramanian et al. 2006; Dens et al. 2012; Russell 2002). Highly plot connected brand placements are intimately tied to the plot and relevant for the plot in some way (cfr. AOL in “You’ve Got Mail”). The effect of the plot connection of a placed brand is explained by the transferal of meaning from the editorial content (the movie or program) to the placed brand (Van Reijmersdal et al. 2007). In the present paper, we propose that the perceived fit between the placed brand and the movie in which it is placed acts as a mediator to explain how plot connection impacts brand attitude. A second gap in existing research is the lack of integration of various determining factors of brand placement effects in a single model. As noted by Balasubramanian et al. (2006), there is a need for more integrative models that include the effects of both placement characteristics and audience perceptions and characteristics. Besides testing the mediating role of perceived fit between the brand and the movie (an audience perception variable), we also investigate whether the effects of connecting a placement to the plot are moderated by the prominence of the placement (a placement characteristic) (cfr. Dens et al. 2012), and consumers’ levels of familiarity with the placed brands (an audience characteristic) in a single integrated model.
Lastly, we conduct a field experiment in a natural exposure setting (i.e., a real-life movie theatre). Most brand placement studies are performed in a laboratory setting or use only short fragments of movies or television programs (e.g, Homer 2009). Under these circumstances, the movie-going experience cannot be validly replicated and its nuances are not captured (Balasubramanian et al. 2006). Our study measured responses of actual moviegoers for a fulllength movie under natural viewing conditions. This greatly enhances the ecological validity of the findings (see discussion in Deighton et al. 1989).
2. Literature review and hypotheses The findings of Russell (2002) show that highly plot connected brands produce more positive brand attitudes than weakly plot connected brands. We propose that this effect of plot connection is mediated by the perceived fit between a movie and brands that are placed in the movie. According to the Associative Network Theory (ANT), human memory is a network of individual, interconnected nodes that activate each other in relevant contexts. Brand knowledge can be conceptualized as a node in memory, to which various associations are linked, forming the brand’s associative network (Dens and De Pelsmacker 2010). A spreading activation process from node to node determines the extent of retrieval in memory . The information that a cue activates in memory , i.e., its associative network, forms the basis for evaluative judgments or attitudes about that cue. It is argued that the attitudinal outcome of this process will be determined by the relative fit or congruence between the associative network of the brand and that of the movie (Gawronski and Bodenhausen 2006). This is where plot connection comes into play, as a way to meaningfully integrate the brand into the story of a movie (Russell 2002). By weaving a brand more strongly into the plot, a higher degree of congruence is created between the associative networks of the brand and the movie by activating common nodes. In other words, plot connecting a placement is
expected to improve the perceived fit between the brand and the movie. Note that we consider plot connection and perceived fit as two related, but conceptually distinct dimensions. As mentioned, plot connection refers to the degree to which a brand is relevant to the story or plot of the movie. This is a factor that characterizes the execution of brand placements themselves (Dens et al. 2012; Russell 2002). Weakly plot connected placements do not contribute much to the story (e.g., driving by a supermarket, seeing a random advertisement on television, etc.), while a strong plot connection means the placement is closely tied in to the plot and constitutes a major thematic element (Dens et al. 2012). Perceived fit, on the other hand, is an audience-side measure that expresses the degree to which viewers perceive a match-up between (the image of) the brand and the movie. In other words, how well do they go together? For example, one of the episodes of Friends is almost completely devoted to the tension between Rachel and Phoebe concerning a Pottery Barn apothecary table, which makes the brand highly plot connected. A priori, viewers may not consider Pottery Barn as a brand that would fit the program well, as the program revolves around the friendship of 6 young adults struggling to survive in the “real world”. However, because of the high degree of plot connection, the perception of fit may increase after this particular episode. Thus, we argue that plot connection positively influences viewers’ perceptions of fit. In turn, we expect perceived fit between the brand and the movie to exert a positive effect on brand attitude. Consumers exposed to brand placements which are congruent with the movie and the character, will more easily assimilate the brand placements to existing activated knowledge structures and develop more favorable brand attitudes. Considering these arguments, we hypothesize that: H1: The perceived fit between the brand and the movie mediates the effect of plot connection on brand attitude.
Apart from plot connection, the literature on brand placement identifies prominence, the conspicuousness of the placement, as an influential placement characteristic (e.g., Cowley and Barron 2008). Low prominence (Cowley and Barron 2008; Homer 2009) brand placements are those in which visual or verbal identifiers of the brand name or logo are not shown or mentioned prominently (e.g. small in size, a background prop outside of the main field of visual focus, , etc.). High prominence (prominent) placements are those that are frequently repeated and/or emphasized either verbally, visually, or both (Dens et al. 2012). For the exact conceptualization of prominence and its different dimensions in the study, we refer readers to the research method section. The findings of Dens et al. (2012) indicate that prominence moderates the effect of plot connection on brand attitude. They demonstrate that, while plot connection has no impact on brand attitude for prominent placements, it exerts a positive effect on brand attitude for subtle placements. They argue that the more prominent a brand placement, the more likely that viewers will consider the appropriateness of the placement in light of its manipulative intent (Cowley and Barron 2008). The persuasive intent may be interpreted by viewers as intrusive, causing irritation and/or distraction and an increase in counter-arguing (Friestad and Wright 1994). This should result in a negative shift in brand attitude , regardless of whether the brand is connected to the plot or not (Dens et al. 2012). For subtle placements, counter-arguing is less likely to occur. As argued by Gawronski and Bodenhausen (2006), viewers’ brand responses will also depend on their existing knowledge structures, i.e., how familiar they are with the placed brands. We propose that brand familiarity will further moderate the interaction between plot connection and prominence on brand attitude. As consumers’ brand familiarity increases, so does their knowledge of how the brand tries to persuade consumers. Thus, it is especially for familiar brands that a prominent placement may instigate conscious deliberation of the
appropriateness of its presence. In contrast, subtle placements may prevent the activation cognitions or thoughts about the brand placement and its appropriateness . In this scenario, it is less likely that persuasion knowledge structures will be activated, thus decreasing the risk of skeptical responses. In that case, plot connection can benefit brand attitude more than with prominent placements. H2: For consumers who are familiar with the placed brand, the positive effect of plot connection on brand attitude will be greater with a subtle placement than with a prominent placement. For unfamiliar brands, consumers lack prior brand associations in memory, and are thus more likely to use the placement to learn about the brand. In that case, the more the brand is connected to the plot, the easier associations from the movie will transfer to the brand, resulting in a positive effect on brand attitude. Prominent placement increases the visibility of the brand and will likely make it easier to link the brand to the context, reinforcing the effect of plot connection on brand attitude. Moreover, because people are less familiar with the brand and its persuasive practices, prominence is less likely to activate persuasion knowledge and negatively impact brand attitude than with familiar brands. We expect that the beneficial effect of plot connection will be smaller for subtle placements. As subtle placements are less noticeable, they might fall under the perceptual threshold that activates the brand node in memory. If this node is not, or less strongly activated, the subtly placed brand does not benefit from being plot connected. Following this reasoning we hypothesize: H3: For consumers who are unfamiliar with the placed brand, the positive effect of plot connection on brand attitude will be greater with a prominent placement than with a subtle placement.
3. Research design and method
3.1. Procedure Our study was carried out at the occasion of an edition of ‘Ladies at the movies’, a monthly women-only event organized in several movie theatres in Belgium. During this event, a preview of a Hollywood blockbuster is shown, due to officially open in theatres a few weeks later. This offers the advantage that none of the participants could have seen the movie before or were likely exposed to promotion for or information about the movie, which may otherwise bias the results of the study. On the night of the study, the movie shown was Bride Wars. Prior to the event, all brand names or logos that were referred to verbally and/or shown visually for at least one second in the movie were content analyzed by two independent coders (cfr. Ferraro and Avery 2000). A coding scheme was developed based on the existing instruments of La Ferle and Edwards (2006) and Ferraro and Avery (2000) to assess prominence and plot connection of each placement. In line with Dens et al. (2012), prominence and plot connection are considered as separate, conceptually independent dimensions. The prominence of each brand placement was rated using a multidimensional coding scheme, encompassing the most important dimensions of prominence: (1) modality of presentation (visual, auditory, audiovisual), (2) time on screen/number of mentions in the same scene, (3) visibility of the brand name, (4) appearance of the brand in close up, (5) appearance of the brand in the fore-/background, (6) amount of other branded products shown or mentioned in the same scene (clutter), and finally (7) character interaction (is the character shown using or handling the brand or actively mentioning the brand name?) (La Ferle and Edwards 2006). For each brand, an overall evaluation of the prominence level in the movie was made based on the characteristics of the individual occurrences. A three-item plot connection measure used in an experiment by Russell (2002) (“…plays an important role in the story”; “Without the references to…, the story would be different”; and “…is connected to
the plot”), served as a guideline to classify the level of a brand placement’s plot connection into four categories (none to high). The level of plot connection was assessed per brand across the entire movie, for it is only after seeing the entire movie that the complete plot and the scope of the role of the brand therein can be assessed. Both coders participated in a training session in which they were thoroughly instructed on the use of the coding scheme. Inter-coder agreement on the prominence level of brands was 91.8%, agreement on plot connection level was 93.9%. Cohen’s Kappa for the overall categorization was K = .755 (+ 1 indicates perfect agreement, 0 indicates no agreement other than expected by chance) (Perreault and Leigh 1989). In order to achieve a substantial amount of variation in prominence and plot connection, and to be able to investigate the interaction effect, we set up a 2 (prominence: subtle, prominent) x 2 (plot connection: weak, strong) fullfactorial design. Based on the coding for all 15 brands in the movie, we selected four brands that, according to both coders, best represented the four conditions of our design (Table 1). The subtle, low plot connected placement was represented by a DHL truck that stood in the background in one of the scenes. The subtle placement with high plot connection was a small Tiffany’s box containing an engagement ring one of the two lead characters finds hidden between her boyfriend’s things. Her finding the box plays a significant role in the further development of the plot. The brand is verbally mentioned in a whisper. It is also mentioned in one other scene. The signature Tiffany’s box is small and only shown briefly. An iPod was prominently placed in a weakly plot connected way. The brand name was mentioned three times in the dialogue, and the product was also visible in one scene. The iPod further had no significance to the plot of the movie, though (it was used while the lead characters were running). Lastly, The Plaza Hotel was placed in a highly prominent and plot connected fashion. The story is centered on two weddings taking place at the hotel. The hotel’s name was verbally mentioned 11 times, the hotel façade is filmed 3 times and numerous scenes take
place within the hotel. The selected brands each represent a different product category. Moreover, to exclude potential confounding effects due to the variation in brands the research design includes both a control of pre-existing attitudes for these brands, and brand familiarity as a moderator. We collected e-mail addresses from the women at the event just before the movie, and emailed them a web link to the questionnaire a few days later. This yielded 167 completed questionnaires (response rate: 39.6%). Because the four brands under study were existing brands, an online survey was also conducted with a control group of women (N = 85, movie: Last Chance Harvey’) as way to control post-exposure attitudes towards the existing brands studied for pre-existing brand attitudes (see discussion in Webb 1979). The e-mail addresses for this control group were collected in an identical fashion at the previous month’s “Ladies at the Movies” event and cross-checked with those of the main study’s respondents to rule out any potential overlap.
Table 1: Overview of brands per condition Prominence
Tiffany & Co.
The Plaza Hotel
3.2. Measures First, we measured respondents’ movie liking (which will be used as a control variable) on a 7-point Likert scale consisting of 7 items (e.g., “I enjoyed watching __”, I don’t regret watching __”, etc., α = .936). Afterwards, respondents completed a 3-item 7-point semantic differential scale measuring their level of familiarity with the test brands and several filler brands (e.g., “I don’t know __ at all - I know __ very well”, etc., α = .796). Next, the 8
respondents completed a 4 item 7-point scale that measured their perceived fit between the brands and the movie (e.g., “(Brand) and (Movie) do not fit together”, etc., α = .915). Finally, brand attitude (Ab) was measured for each of the four separate brands under study in the movie by means of a four-item seven-point semantic differential scale (e.g., negative – positive, = .878). Brand attitude was also measured in the control group, for the four brands under study ( = .883). Per brand, summated scales were calculated for both the main group and the control group. For each brand, the control group’s mean Ab scores were subtracted from each individual’s post-exposure scores in the main group (cfr. Dens et al. 2012; Russell 2002). This resulted in an Ab difference measures that expresses the shift in Ab due to the experimental exposure. These scores are used in subsequent analyses.
4. Results The hypotheses are tested using the PROCESS-macro (Model 12) for SPSS (Hayes 2008). This procedure allows statistical testing of indirect effects by generating asymmetric bootstrap confidence intervals for inference. In order to effectively assess mediation, this procedure estimates two Ordinary Least Squares regression models. The first model regresses the mediator (perceived fit) on plot connection, prominence and brand familiarity, including all two-way interactions, and the three-way interaction. The second model regresses the dependent variable (brand attitude) on plot connection, prominence and brand familiarity, their two-way interactions, and the three-way interaction, together with perceived fit (the mediator) (Table 2). Both models contained movie liking (M = 5.60) as a covariate, as previous research has demonstrated that program liking may influence viewers’ attitudes toward embedded brands (e.g., Lehu and Bressoud 2008). The Condition Indices indicated no potential problems with multicollinearity (CI = 12.061 for model 1, and CI = 15.851 for model 2, well below the suggested cut-off of 30), which evidences good discriminant validity
between all independent variables (Janssens et al. 2008). It could be argued that brand familiarity and placement prominence are related constructs (brand familiarity could increase perceived prominence, or prominence could positively affect subsequent brand familiarity), which could confound our results. We therefore examined the correlation between these two constructs before proceeding to the hypotheses tests. The results indicate that the constructs are not related. The Point-Biserial Correlation (which is the logical equivalent of the t-test when correlating a binary variable (prominence) with a continuous measure (familiarity)) evidences independence, r = .064, t = -1.652, p = .099. Table 2: Unstandardized model coefficients Perceived
Plot connection x prominence
Plot connection x brand familiarity
Prominence x brand familiarity
Plot connection x prominence x brand familiarity
Note: *** p ≤ .001, ** p ≤ .01, * p ≤ .05
The results of the regressions show that plot connection exerts a significantly positive effect on the perceived fit between the brand and the movie (b = 2.932, t = 5.057, p < .001). Perceived fit in turns has a significantly positive effect on brand attitude (b = .108, t = 3.391, p < .001). A Sobel test indicates that the indirect effect of plot connection on brand attitude, through perceived fit, is positive and significant (t = 3.107, p = .002). This result confirms H1. As the results in Table 2 show that the direct effect of plot connection on brand attitude is not significant (b = .293 , t = .751, p = .453), this suggests a case of indirect-only mediation (Zhao et al. 2010). In regression 2, the three-way interaction between plot connection, prominence and brand familiarity on brand attitude is significant (b = -.298, t = 2.442, p = .015). In order to test H2 and H3, we calculated the conditional effects of plot connection on brand attitude at different levels of both moderators. For prominence, these levels were ‘subtle’ (0) or ‘prominent’ (1), while for brand familiarity we evaluated the conditional effects at – 1 and +1 standard deviation from the mean . The results of these analyses are depicted in Figure 1 and Table 3. Our results show that, for more familiar brands, the effect of plot connection on Ab is marginally significant with a subtle placement (b = .345, t = 1.822, p = .069), but not with a prominent placement (b = .116, t = .650, p = .516). As recommended by , a t-test was performed to test the difference between both regression coefficients. The result of this test shows that the difference between both regression coefficients is actually not significant (t = .882, p = .378). Consequently, Hypothesis 2 is rejected.
Table 3: Conditional effects of plot connection on brand attitude Brand familiarity
Effect size Std. Err.
High (+1 SD)
High (+1 SD)
Low (-1 SD)
Low (-1 SD)
As predicted in H3, for rather unfamiliar brands, the effect of plot connection on Ab is higher with a prominent placement (b = .890, t = 6.188, p < .001) than with a subtle placement (b = .320, t = 1.977, p = .048). A t-test of the difference of between both regression coefficients shows that the effect of plot connection on Ab is significantly larger for the prominent placement than for the subtle placement (t = 2.637, p = .008). H3 is thus confirmed.
Figure 1: The effect of plot connection on brand attitude across different levels of brand familiarity and prominence.
Effect size of Plot connection on Ab
1 0,8897 0,9 0,8 0,7 0,6 0,5
0,1 0 2
3,5 4 4,5 Brand Familiarity
5. Discussion, limitations and managerial implications The present study offers an ecologically valid investigation of how movie brand placements differing in plot connection and prominence influence the formation of consumers’ brand attitudes through perceived brand-movie fit, and how brand familiarity moderates this process. By incorporating perceived fit between the brands and the movie as a mediator, we confirmed our assertion that plot connection impacts brand attitude by increasing the convergence between the associative networks of the brands and that of the movie. These findings support the notion that the effectiveness of brand placement is driven by its connection to media content (Russell and Stern 2006), and elucidate how this connection influences brand attitude. Connecting a brand to the plot of a movie gives meaning to a placement by causing a higher perceived fit between the brand and its context. In other words, plot connection creates congruence between both attitude objects. It is through this increased congruence, or fit, that plot connection influences brand attitude. The results of this study confirm the idea embedded in associative network theory that it is crucial to take existing knowledge structures into account when studying the attitudinal impact of new information (Gawronski and Bodenhausen 2006). This is something previous brand placement research has ignored. For familiar brands, the effect of plot connection was only (marginally) significant with a subtle placement, but not with a prominent placement. Although prominently placing a familiar brand leads to less positive brand attitudes than subtly placing a familiar brand, suggesting that negative counter-arguing does occur, it does not alter brand attitudes significantly. The idea behind brand placement is that the brand attitude benefits from new (positive) movie associations. However, for an already familiar brand, the existing associations diminish the impact of new associations on brand attitude (Finn and Roediger 2013). Therefore, the upward potential for familiar brands is limited.
For unfamiliar brands, the effect of plot connection is significantly more positive with a prominent placement than with a subtle placement. By stimulating the activation of brand relevant associations in memory, prominence creates a higher opportunity for the associative networks of the movie and the unfamiliar brand to converge. Furthermore, skeptical responses to unfamiliar brands, as a consequence of prominent placement, are less likely to occur with unfamiliar brands, because persuasion knowledge structures related to the brand are less available than for a familiar brand. As such, it is less likely that prominent placement will be negatively evaluated and decrease brand attitude. In addition to its theoretical contributions, the present study offers several managerially relevant implications. Currently, advertisers investing in brand placement campaigns are mainly paying for the amount and explicitness of exposure their brand gets in media content. Prior studies have illustrated that prominent placements can actually harm brand attitude (e.g., Cowley and Barron 2008), especially when they are not connected to the plot of the media content (Dens et al. 2012). The results of the present study point out that this pitfall does not seem to apply to novel or less familiar brands. For these brands, prominently connecting a brand to the plot of a movie is a good way to attach meaningful associations to the brand, and it does not lead to negative evaluative brand responses on behalf of viewers. If the brand’s target audience is highly familiar with the brand, subtle placement is the safest strategy. To inform future research into the workings of brand placement, a number of limitations of this study are discussed. Although capturing the experience of a real visit to the movie theatre enhances the external or ecological validity, a field study offers the potential drawback of a decrease in internal validity caused by the lack of control over external circumstances (Deighton et al. 1989). For instance, we were unable to control for exposure to additional marketing communications for the brands studied between respondents’ watching of the movie, and respondents’ completion of the survey a couple of days later. Secondly, the
present study is also limited by the fact that only one movie is studied, and the audience consisted of women only. Thirdly, the brands under study belong to different product categories and therefore may differ on other dimensions other than the ones that were measured in this study. While this is partly mitigated by the inclusion of a control measure of brand attitudes, and by explicitly including brand familiarity in the model, we would urge researchers to include more detailed category information (e.g., a measure of product category involvement) in their future studies. A fourth limitation relates to the application of the ANT as our the theoretical basis of our predictions. Although the logic of the framework was carefully applied to explain these effects, we do not directly assess the underlying theoretical mechanism itself. This is beyond the scope of the present study. Future research should perform a more detailed and systematic experimental scrutiny of the exact structure of the associative networks of brands and the movie they are placed in to uncover the specific mechanics of the attitude formation process.
6. Acknowledgement The authors gratefully acknowledge the financial support of the Flemish Agency for Innovation Through Science and Technology (IWT Vlaanderen).
7. References Balasubramanian SK, Karrh JA, Patwardhan H (2006) Audience Response to Product Placements: An Integrative Framework and Future Research Agenda. Journal of Advertising 35 (3):115-141 Cowley E, Barron C (2008) When Product Placement Goes Wrong: The Effects of Program Liking and Placement Prominence. Journal of Advertising 37 (1):89-98 Deighton J, Romer D, McQueen J (1989) Using Drama to Persuade. Journal of Consumer Research 16 (3):335-343 Dens N, De Pelsmacker P (2010) Advertising for extensions: Moderating effects of extension type, advertising strategy, and product category involvement on extension evaluation. Marketing Letters 21 (2):175-189
Dens N, De Pelsmacker P, Wouters M, Purnawirawan N (2012) Do you like what you recognize? The effects of brand placement prominence and movie plot connection on brand attitude as mediated by recognition. Journal of Advertising 41 (3):35-53 Ferraro R, Avery RJ (2000) Brand Appearances on Prime-Time Television. Journal of Current Issues and Research in Advertising 22 (2):1-15 Finn B, Roediger HLI (2013) Interfering Effects of Retrieval in Learning New Information. Journal of Experimental Psychology: Learning, Memory, and Cognition 39 (6):1665-1681 Friestad M, Wright P (1994) The Persuasion Knowledge Model: How People Cope With Persuasion Attempts. Journal of Consumer Research 21 (1):1-31 Gawronski B, Bodenhausen GV (2006) Associative and Propositional Processes in Evaluation: An Integrative Review of Implicit and Explicit Attitude Change. Psychological Bulletin 132 (5):692-731 Hayes AF (2008) Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press, Homer PM (2009) Product Placements: The impact of Placement Type and Repetition on Attitude. Journal of Advertising 38 (3):21-31 Janssens W, De Pelsmacker P, Wijnen K, Van Kenhove P (2008) Marketing research with SPSS. Prentice Hall, La Ferle C, Edwards SM (2006) Product Placement: How Brands Appear on Television. Journal of Advertising 35 (4):65-86 Lehu J-M, Bressoud E (2008) Effectiveness of Brand Placement: New Insights About Viewers. Journal of Business Research 61 (10):1083-1090 Perreault WD, Jr., Leigh LE (1989) Reliability of Nominal Data Based on Qualitative Judgments. Journal of Marketing Research 26 (2):135-148 Russell CA (2002) Investigating the Effectiveness of Product Placements in Television Shows: The Role of Modality and Plot Connection Congruence on Brand Memory and Attitude. Journal of Consumer Research 29 (3):306-318 Russell CA, Stern BB (2006) Consumers, characters, and products: A balance model of sitcom product placement effects. Journal of Advertising 35 (1):7-21 Van Reijmersdal EA, Neijens P, Smit E (2007) Effects of Television Brand Placement on Brand Image. Psychology & Marketing 24 (5):403-420 Webb PH (1979) Consumer Initial Processing in a Difficult Media Environment. Journal of Consumer Research 6 (3):225-236 Zhao X, Lynch JG, Jr., Chen Q (2010) Reconsidering Baron and Kenny: Myths and Truths about Mediation Analysis. Journal of Consumer Research 37 (2):197-206