ABSTRACT
Every year, executives debate whether to release their Super Bowl advertisements early on social media, or whether this early release will hurt the surprise and impact of being seen for the first time during the most viewed televised sporting event in the United States. This research examines the impact of prior social-media engagement with a Super Bowl advertising campaign on consumers' moment-to-moment affect traces and retrospective evaluations of television advertisements. Benefits include affect traces characterized by a higher peak, final moment, and linear trend as well as higher retrospective advertisement and brand evaluations. Reflecting the positive influence of social media on attitudes and processing, this research also examines the drivers of social-media interaction. Findings include that watching an advertisement on social media was driven mainly by how engaging, serious, or humorous consumers believed an advertisement would be. Liking or sharing an advertisement on social media was driven primarily by consumers' cognitive thoughts about the advertisement or brand.
MANAGEMENT SLANT
There are many benefits to releasing an advertisement early on social media, including affect traces characterized by a higher peak, final moment, and linear trend as well as better advertisement and brand attitudes.
If a manager's goal is to get people to watch an advertisement on social media, this behavior is motivated by how engaging, serious, or humorous an advertisement is perceived to be.
If a manager's goal is to get people to like or share an advertisement on social media, this behavior is motived by whether consumers perceive the advertisement as influential to their purchase decision.
INTRODUCTION
Companies spent approximately $5 million dollars for 30 seconds of advertising time on the Super Bowl in 2018; many industry executives wondered whether this was money well spent. Although not all executives agree, companies that pursue Super Bowl advertising often cite the added benefits of the pregame and post-game buzz that their advertisements generate. Research shows that companies that release their advertisements prior to the Super Bowl generate more than four times as much buzz, in terms of traditional and social-media mentions, as companies that do not (Nail, 2007).
The publicity benefits associated with purchasing a Super Bowl advertisement cannot be argued, but there is much debate surrounding the early release of Super Bowl advertisements on social media. Many argue that the early release will ruin the surprise and impact for the more than 100 million viewers who tune in to the big game specifically to watch the advertisements. Although some consumer-behavior research from the variety-seeking tradition indicates that consumers prefer novel stimuli to provide them with new and exciting experiences and emotions (Mowen, 1988), research on the mere-exposure theory shows that consumers often prefer stimuli that they have encountered before (Janiszewski, 1993). Recognizing such inconsistencies in the literature, the current authors posed the following research question:
RQ1: Is there a benefit to releasing advertisements early on social media, especially before high-viewership and high-impact programming, such as the Super Bowl?
Social media are an increasingly important part of the integrated marketing-communications mix, used to build relationships with consumers by engaging them with informative or entertaining content that they can watch, like, or share with their contacts. In previous years, advertisers focused on the number of likes a fan page or post received instead of the number of consumers sharing their content. The number of likes resulting in a relationship with a consumer was considered the primary indicator for a company to measure return on investment (Berry, 2013).
Recently, marketers have realized that social-media likes do not translate always to desirable business outcomes (Lake, 2011). Instead, marketers are turning their focus to the extent to which consumers share marketing content on social media as the key quantifiable metric for engagement. The shift from liking to sharing content makes sense given that the sharing of content is tied more strongly to campaign business objectives (Heine, 2013).
Despite this shift, academic literature has not explored yet what aspects of an advertisement campaign differentially affect the three distinct social-media behaviors of watching, liking, and sharing marketing content. Thus, the current authors posed a second research question:
RQ2: What types of advertisement characteristics are associated most with the distinct behaviors of watching, liking, and sharing an advertisement on social media?
In a departure from previous research that focused on the importance of the entertainment value of a campaign, this study found a strong role for the advertisement- and brand-related cognitions generated by a campaign.
The current research addresses several gaps in the marketing literature. To the authors' knowledge, this is the first study to examine the impact of prior social-media engagement on consumers' moment-to-moment emotional responses to advertisements to provide a better understanding of how prior social-media engagement affects subsequent emotional processing of advertising content. Whereas prior literature focused on the peak, final moment, and linear trend of consumers' affect traces as being important determinants of consumers' advertisement and brand attitudes, this research found two new characteristics of affect traces that were linked strongly to improved advertisement and brand attitudes—the time before the peak, and the duration of the peak level of affect.
This research also aids practitioners in the navigation of conflicting evidence surrounding the early release of a Super Bowl advertisement on social media. Does prior research from the variety-seeking or mere-exposure traditions best predict how consumers will respond to advertisements on subsequent viewings? The current findings provide companies with added assurance on how their large advertising investment will perform on game day.
This study also answers the numerous calls for additional research on what advertisement characteristics differentially affect watching, liking, and sharing content on social-networking sites (de Vries, Gensler, and Leeflang, 2012; Shankar and Batra, 2009). Prior research stressed the importance of a wide range of “attitude toward the advertisement” variables covering different emotional and cognitive reactions consumers may have toward an advertisement. The current research suggests that whether an advertiser uses an emotional or cognitive approach should depend on the exact social-media behavior the practitioner wants to encourage with the campaign.
THEORETICAL DEVELOPMENT
The Impact of Prior Social-Media Engagement On Advertisement Attitudes, Brand Attitudes, And Purchase Intentions
This article first addresses the question as to whether prior social-media engagement with an advertisement campaign is beneficial or detrimental to consumers' subsequent advertisement attitudes, brand attitudes, and purchase intentions. Although top executives argue about the merits of an early release of a Super Bowl advertisement on social media, existing research has provided no definitive evidence as to whether prior social-media engagement with an advertisement campaign enhances subsequent exposure to that campaign.
Consumer-behavior research on variety seeking argues that consumers have a natural tendency to seek new sensations and pursue variety in their lives to create new experiences and reach higher levels of positive emotion (Mowen, 1988). A dynamic-attribute satiation model of variety seeking states that consumers seek out certain attributes in a product, and when these desires are satisfied, consumers then seek products delivering an entirely different set of attributes (McAlister, 1982). Attribute satiation is a similar concept to wear-out in the advertising world, whereby an advertisement no longer has any positive effect on consumers' attitudes or behaviors because the consumer has been overexposed to the advertisement (Pechmann and Stewart, 1988).
The mere-exposure theory, conversely, argues that consumers prefer stimuli that they have encountered before, and research has shown that consumers have more favorable attitudes toward advertisements, brands, and packages that they have seen before (Janiszewski, 1993). More recent research examining musical preferences shows that even though consumers said they preferred listening to new music, they actually selected music that they had heard before in choice tasks (Ward, Goodman, and Irwin, 2013). This stream of research shows that prior exposure is a stronger predictor of consumers' choices than liking of the stimuli.
Despite the conflicting implications of prior research, the authors hypothesized that companies would benefit from releasing their Super Bowl advertisement early on social media because of the advertisement attitudes, brand attitudes, and purchase intentions the advertisement would generate. Although prior research does suggest that wear-out is a potential problem for advertising campaigns, current studies show that wear-out typically occurs not within a period of weeks but rather within a period of months after initial exposure to an advertisement campaign (Blair, 2000; Pechmann and Stewart, 1988). This stream of research also acknowledges a period of wear-in, or the idea that it takes multiple exposures to an advertisement for the message to reach its full impact on a consumer (Blair, 2000; Pechmann and Stewart, 1988).
Support for the benefits of prior exposure to a campaign on social media also comes from research on the synergistic effects of multiple forms of media. This literature illustrates that the persuasive impact of two advertisements seen on different media by a single consumer is greater than the additive impact of two consumers experiencing those same advertisements separately (Varan, Murphy, Hofacker, Robinson, et al., 2013). Research suggests that synergistic effects happen because exposure to an advertisement in the first medium primes interest in the advertisement in a second medium (Voorveld, Neijens, and Smit, 2011). For example, hearing an advertisement in multiple formats forms a perception that a brand is high quality and popular (Voorveld et al., 2011). This is probably why the National Football League has a website that allows viewers to preview the upcoming Super Bowl advertisements as soon as they are released prior to the game.
The authors thus formulated the following hypothesis:
H1: Consumers who engage with Super Bowl advertisements on social media will experience higher advertisement appeal, stronger brand attitudes, and higher purchase intentions than consumers who do not.
The Impact of Prior Social-Media Engagement on Consumers' Emotional Processing of Advertisements
Moment-to-moment evaluations of an advertisement are a useful tool for understanding how prior social-media engagement with an advertising campaign affects consumers' subsequent emotional processing of an advertisement. This technique consists of collecting continuous self-report measures of the level of affect consumers experience in each second of the advertisement. The data are collected through a computer; consumers watch an advertisement while using their mouse to guide a slider scale to reflect how appealing or unappealing they find the advertisement at every single moment. The advertisement runs on approximately 90 percent of the computer screen, and the slider scale occupies the lower 10 percent of the computer screen.
Researchers and practitioners can use the resulting graphs, or affect traces, to interpret what executional elements enhanced or detracted from consumers' overall advertisement and brand evaluations. They also can examine key aspects of the resulting affect trace (e.g., peak, final moment, and linear trend), which prior research has shown to be correlated highly with advertisement and brand evaluations (Baumgartner, Sujan, and Padgett, 1997; Woltman Elpers, Mukherjee, and Hoyer, 2004). The tool used in this study was developed by BDJ Solutions in Medford, Massachusetts, and is very similar to the “worm” and “reactor” tools developed by Roy Morgan Research. These tools typically are proprietary and developed by communications-research agencies that specialize in this type of research. The authors are unaware of any off-the-shelf software packages that would allow anyone to conduct this type of research easily.
Prior academic research using moment-to-moment advertisement evaluations has focused on the importance of the peak, final moment, and linear trend of consumers' affect traces. The peak is the highest level of positive affect a consumer experiences while watching an advertisement. The final moment is the level of affect consumers experience during the last second of an advertisement. The linear trend is the slope of the linear-regression function that best fits the data points of the affect trace.
One study found that key aspects of a consumer's affect trace (i.e., the peak, final moment, and linear trend) were better predictors of retrospective advertisement and brand evaluations than the sum or average of all moments (Baumgartner et al., 1997). Another replicated these findings in the context of humor and found that affect traces characterized by a high peak, final moment, and linear trend also were rated the most humorous (Woltman Elpers et al., 2004). Likewise, the authors of the current study predict that individuals with prior social-media engagement with a campaign will experience affect traces characterized by a high peak, final moment, and linear trend:
H2: Consumers who engaged with Super Bowl advertisements on social media prior to the advertisement test will have affect traces characterized by a higher peak, final moment, and linear trend.
An interesting aspect of prior research on affect traces is the finding that the peak level of affect occurs at the end of the advertisement for advertisements with the highest retrospective evaluations (Baumgartner et al., 1997; Woltman Elpers et al., 2004). These prior studies quantified this finding by showing that retrospective advertisement and brand evaluations were related positively to the time before the peak level of affect occurred and negatively to the time after the peak level of affect occurred.
An aspect overlooked by these studies is the duration of the peak. This is due to the significant information loss that results from the aggregation of moment-to-moment data to use the advertisement, instead of the individual, as the unit of analysis (Burton, McAlister, and Hoyer, 2015). Aggregating the data typically results in a clearly defined moment that is classified as the peak because the average of all values typically goes out many decimal places. Individual affect traces, however, illustrate that consumers may spend several moments experiencing a peak level of affect. This is analogous to a consumer providing a high rating to an entire scene of an advertisement that contains a funny punch line. Consumers who like an advertisement more likely will have longer peaks that begin earlier in the advertisement.
In a departure from prior research, therefore, the authors believe that consumers with prior social-media engagement with a campaign will experience peak levels of affect occurring earlier in the advertisement. Because these consumers also will experience longer peaks, however, the authors expect that the peak level of affect also will occur closer to the end of the advertisement, consistent with prior research. The authors thus proposed the following hypothesis:
H3: Consumers who engaged with Super Bowl advertisements on social media prior to the advertisement test will experience a longer duration of peak affect with fewer moments before the peak level of affect occurs and fewer moments after the peak level of affect occurs.
The Drivers of Watching, Liking, and Sharing An Advertisement on Social Media
Next, this article examines advertising characteristics and perceptions that drive the various levels of social-media interaction—watching, liking, and sharing. These three social-media behaviors differ in the level of interactivity they offer and amount of effort required. Watching an advertisement on social media has the least amount of interactivity and effort. In comparison, sharing an advertisement on social media requires the most interactivity and effort. Liking an advertisement on social media falls between watching and sharing behaviors and requires a medium level of interactivity and effort. Additionally, the activities of liking and sharing content on social media presume that an individual has watched the content, so they inherently involve a more complex interaction with the advertisement.
According to the elaboration likelihood model, watching can be considered a low-involvement activity whereby peripheral route processing takes place and persuasion occurs because of affect transfer from pleasant and entertaining stimuli to the brand itself (Petty and Cacioppo, 1986). Peripheral processing also likely will occur with exposure to positive emotional or humorous content. This suggests that advertisements featuring positive emotion and humor likely will encourage lower involvement behaviors on social-media sites, such as simply watching the advertisement (Eisend, 2011). The drivers of getting people to watch an advertisement on social media more likely are related to the entertainment motives that generally determine consumers' reactions to social media (Taylor, Lewin, and Strutton, 2011). In other words, people are drawn into social-media content looking to satisfy their entertainment needs.
In the context of the elaboration likelihood model, liking and sharing can be considered higher involvement activities in which central-route processing takes place and advertisement- and brand-related cognitions play a more important role in persuasion (Petty and Cacioppo, 1986). These higher involvement behaviors are driven by higher level motives, such as helping others and appearing more knowledgeable (de Valck, van Bruggen, and Wierenga, 2009). For example, a study of Chinese consumers' motivations for visiting and interacting on corporate social-media pages revealed that the participants liked or visited a company's site primarily for informational purposes, whereas entertainment and social needs were considered secondary and less-important benefits (Men and Tsai, 2013).
The authors therefore predict that entertainment motives will be the strongest drivers of getting consumers to watch a campaign on social media. They also predict, conversely, that advertisement and brand cognitions will be the primary driver of whether a consumer likes or shares that content. The authors thus formulated the following hypotheses:
H4: Perceptions of an advertisement's entertainment value—perceived engagement, humor, and seriousness—will be the biggest drivers of getting consumers to watch a Super Bowl advertisement on social media.
H5: The ability of an advertisement to generate advertisement- and brand-related cognitions will be the biggest driver of getting people to like or share a Super Bowl advertisement on social media.
METHOD
The authors worked in conjunction with a marketing-research and survey-panel company to collect the data for this study. The survey-panel company recruited a total of 650 consumers from 46 states in the United States to complete an online survey in which consumers watched and answered questions about the advertisements from Super Bowl XLVII. Data were collected over a one-month period starting approximately three and a half weeks after the airing of Super Bowl XLVII, from February 28, 2013, to March 26, 2013.
Consumers were asked to evaluate either four or five advertisements in a session (depending on the length of the advertisements being tested) to prevent survey fatigue. The authors tested a total of 25 Super Bowl advertisements in a series of five sessions over that month. Advertisements were from different product categories, including automobiles, milk, soft drinks, beer, salty snacks, fast food, sweets, insurance, web services, real estate, cleaning supplies, and financial services.
After agreeing to participate in the study, consumers received detailed instructions about how to view the test advertisements and use a slider scale to provide moment-to-moment evaluations of the Super Bowl advertisements. These instructions included a video module that demonstrated the technique visually and allowed participants to practice using the measurement tool until they were comfortable with it. The evaluation process involved consumers watching advertisements on their computer screen, much as they would on YouTube, and using their mouse to guide a slider scale that appeared below the advertisement they were watching. The slider scale measured appeal and contained a scale from 0 to 10, where 0 represented that the advertisement was “extremely unappealing” and 10 represented that the advertisement was “extremely appealing.” Level 5 on the slider scale represented a neutral level of appeal (i.e., the advertisement was not appealing or unappealing).
The advertisement would not play until consumers used their mouse to place the cursor over the midpoint of the scale (or 5) for at least five seconds. This ensured that all participants started with a neutral level of appeal as their baseline, and they adjusted their assessments from there to reflect their continuous assessments of appeal over the course of the advertisement. The computer recorded the position of the slider scale at every second of the advertisement, which produced a graph of the level of affect consumers experienced during each moment of the advertisement.
The authors asked consumers a number of questions after each advertisement to gauge their emotional and cognitive reactions to it. Consumers also answered questions that examined how the advertisement had influenced their brand attitudes and purchase intentions. Of particular interest to this study, the authors first asked whether the consumer had interacted with the advertisement on social media.
The authors made distinctions among “watching,” “liking,” and “sharing” behaviors. “Watching” was considered watching an advertisement on any social-media site. “Liking” was considered the action of liking an advertisement on Facebook or re-Tweeting a posted advertisement on Twitter. “Sharing” was considered the action of actually posting a Super Bowl advertisement to one's Facebook or Twitter account.
Finally, to better understand which advertisement perceptions drive social-media behavior, the authors collected consumers' ratings of the advertisements on
advertisement cognitions (how informative the advertisement was and whether the advertisement provided helpful information; Cronbach's α = .65),
brand cognitions (“The advertisement gave me a positive brand attitude,” “The advertisement improved my product impression,” “The advertisement increased my purchase intentions”; Cronbach's α = .89),
perceptions of advertisement seriousness,
perceptions of advertisement humor, and
perceptions of advertisement engagement.
All postviewing advertisement and brand perceptions were measured on a 5-point scale. If multiple scale items were used to measure a construct, the average for all items was used for the overall measure.
RESULTS
Multivariate regression was used to test the first three hypotheses that examined the relationship between prior social-media engagement with a campaign and consumers' reactions both during and after watching an advertisement. To demonstrate the impact of an early release of a Super Bowl advertisement on social media, the authors dropped from the analysis six advertisements that were not released early or were pulled from ultimately appearing during the Super Bowl.
H1 predicted that prior engagement would lead to higher levels of advertisement appeal, better brand attitudes, and stronger purchase intentions. H1 was confirmed, because consumers who engaged with a campaign through social media rated the advertisement as more appealing (MPrior = 4.26 versus MNoPrior = 3.57), F(1, 12443) = 132.81, p < .01, adjusted R2 = .05, and had higher brand attitudes (MPrior = 3.84 versus MNoPrior = 3.08), F(1, 12443) = 200.27, p < .01, adjusted R2 = .08, and purchase intentions, (MPrior = 3.97 versus MNoPrior = 3.19), F(1, 12443) = 278.22, p < .01, adjusted R2 = .10, than consumers who did not engage with the campaign on social media (See Figue 1).
H2 stated that prior social-media engagement would have a positive impact on the level of affect consumers experienced at the peak, final moment, and linear trend of their affect traces. H2 was confirmed, because consumers who had prior social-media engagement experienced a higher
peak (MPrior = 8.22 versus MNoPrior = 7.70) F(1, 12443) = 22.53, p < .01;
final moment (MPrior = 7.52 versus MNoPrior = 6.49), F(1, 12443) = 52.54, p < .01; and
linear trend (MPrior = 0.53 versus MNoPrior = 0.39), F(1, 12443) = 8.04, p < .01 (See Figure 1).
These findings are significant because they illustrate that prior social-media exposure leads to higher levels of appeal while consumers are viewing the advertisement subsequently.
Similarly, consumers who had prior social-media engagement with an advertising campaign experienced a peak level of affect that occurred earlier in the advertisement (MPrior = 15.32 versus MNoPrior = 18.21), F(1, 12443) = 13.64, p < .01, and included fewer seconds between the peak level of affect and the end of the advertisement (MPrior = 7.94 versus MNoPrior = 10.43), F(1, 12443) = 13.59, p < .01. This finding is because consumers with prior social-media engagement experienced a peak level of affect for a longer duration than consumers without prior social-media engagement with an advertising campaign (MPrior = 15.27 versus MNoPrior = 9.87), F(1, 12443) = 61.13, p < .01, which confirms H3 (See Figure 1).
H4 and H5 were designed to address the advertisement perceptions that differentially influence the social-media behaviors of watching, liking, and sharing. The remaining analyses therefore only looked at the 229 consumers who interacted with any of the 26 Super Bowl advertisements previously on social media. The authors used discriminant analysis to understand the impact of various advertisement perceptions on the dichotomous variables of watching, liking, and sharing an advertisement on social media.
The discriminant function that predicts whether someone will watch an advertisement on social media was significant (See Table 1). Confirming H4, perceived engagement was the top driver of whether a consumer watched an advertisement on social media, followed by perceived seriousness and perceived humor. Perceptions that an advertisement would enhance a consumer's brand or advertisement cognitions did not predict whether a consumer watched an advertisement on social media.
The discriminant functions predicting liking and sharing behaviors on social media were also significant (See Table 1). The biggest predictors of liking advertisements on social media were brand cognitions, advertisement cognitions, and perceived engagement. The biggest predictors of sharing advertisements on social media were advertisement cognitions, brand cognitions, perceived engagement, and perceived seriousness. Both discriminant functions showed that advertisement and brand cognitions were the biggest predictors of liking and sharing advertisements on social media, confirming H5.
DISCUSSION
The current research offers a few important theoretical contributions. First, the study provides substantial evidence that prior social-media engagement with a Super Bowl advertisement results in stronger emotional processing of advertising content, leading to higher advertisement attitudes, brand attitudes, and purchase intentions. This finding is theoretically important because literature on variety seeking suggests that consumers prefer to encounter stimuli that promise to deliver novel experiences and sensations (McAlister, 1982; Mowen, 1988), whereas mere-exposure theory suggests that consumers prefer to encounter stimuli that they have seen before (Janiszewski, 1993; Ward et al., 2013). The evidence provided in this study suggests that prior social-media exposure to an advertising campaign facilitates the wear-in process (Pechmann and Stewart, 1988), or the idea that it takes multiple exposures to an advertising campaign before it has the maximum impact on consumers' attitudes.
The results also provide support for the idea that seeing an advertisement on multiple forms of media produces synergistic effects on consumers' attitudes and behaviors because exposure to the advertisement in one medium primes interest in the second medium (Varan et al., 2013; Voorveld et al., 2011). Comparing the impact of previously seeing an advertisement on television versus seeing that same advertisement previously on social media reveals a much stronger impact of social media on subsequent advertisement viewings.
Second, the present study found two new characteristics of affect traces not addressed in prior research that illustrate higher emotional processing of advertisements that lead to higher advertisement attitudes, brand attitudes, and purchase intentions. Prior social-media engagement with an advertising campaign led to affect traces characterized by a higher peak, final moment, and linear trend, consistent with prior research on affect (Baumgartner et al., 1997) and humor (Woltman Elpers et al., 2004). Deviating from prior research, the authors have shown that prior social-media engagement with a campaign led to a longer duration of the peak and a peak level of affect that occurred earlier in the advertisement.
Prior affect trace research did not address the duration of the peak level of affect, and the authors found that this variable also was related strongly to advertisement attitudes, brand attitudes, and purchase intentions. When consumers experienced a longer peak, the peak occurred earlier in the advertisement (a departure from prior research) and ended later in the advertisement (consistent with prior research). The authors expect that these differences occurred because prior research (e.g., Baumgartner et al., 1997; Woltman Elpers et al., 2004) aggregated the moment-to-moment data and used the advertisement as the unit of analysis, which resulted in a clear moment that was labeled the peak. When using the individual as the unit of analysis, however, researchers will notice that many consumers experience a peak level of affect occurring over many moments. The average duration of the peak for this study's sample was 13 seconds, with more than 75 percent of the sample having a peak level of affect lasting more than one second.
Third, the current study departs from prior research that emphasized the role of emotions in predicting the success of Super Bowl advertisements (e.g., Kim, Freling, and Grisaffe, 2013; Siefert, Kothuri, Jacobs, Levine, et al., 2009). Instead, the authors found that an advertisement's ability to generate advertisement- and brand-related cognitions was a bigger predictor of social-media behaviors than emotions such as engagement and humor. Additionally, prior research examining social-media behavior generally focused on one behavior (see Naylor, Lamberton, and West, 2012, for a notable exception), instead of examining how personal and advertisement characteristics differentially affect the behaviors of watching, liking, and sharing advertisements on social media. The authors found that people were drawn to watch content on social media because of emotional perceptions that the advertisement is engaging, serious, or humorous. More important than emotional perceptions, however, was the extent to which an advertisement generated advertisement- and brand-related cognitions and drove consumers to like or share content on social-media networks.
Distinguishing among different social-media behaviors helps reconcile conflicting findings in the research emphasizing the importance of emotions (e.g., Stieglitz and Dang-Xuan, 2013; Taylor et al., 2011) versus cognitions (e.g., Berger and Milkman, 2012; Saxena and Khanna, 2013) in predicting social-media behaviors. The results are consistent with the implications of central-route processing from the elaboration likelihood model (Petty and Cacioppo, 1986). Liking and sharing an advertisement on social media require a greater amount of involvement, which suggests that product- and advertisement-related cognitions are more important in attitude formation.
Managerial Implications
From a practical standpoint, the current research provides several contributions to managerial practices. The findings demonstrate that prior social-media engagement with an advertising campaign does not spoil the surprise or novelty of a Super Bowl advertisement. Instead, releasing an advertisement on social media prior to the Super Bowl aids in the processing of the advertisement when game day arrives. Practitioners should use the strategy of an early advertisement release because of the high level of liking and impact the advertisement will have on the audience. This high impact helps practitioners create a more memorable advertisement through strong advertisement attitudes, brand attitudes, and purchase intentions.
It is also important to point out that the mere-exposure effect found by the authors (i.e., the tendency to rate advertisements higher on subsequent viewings) was more pronounced when that previous exposure occurred on social media. Media planners taking an integrated marketing-communications perspective should emphasize social media more in their media mix when trying to achieve their desired levels of frequency.
Conflicting research concerning the role of information and entertainment value in the processing of social-media campaigns creates additional questions for practitioners. The current research emphasizes the importance of entertainment value in social media but found that the higher levels of social-media engagement were driven by advertisements that generated advertisement- and brand-related cognitions. Although an advertisement's perceived entertainment value encourages consumers to watch an advertisement on social media, managers should be sure to incorporate informative content and useful brand information embedded in engaging content to encourage liking and sharing of a social-media campaign.
Moment-to-moment assessments of advertising campaigns allow practitioners to examine how an audience emotionally is processing content of an advertisement. It is important that managers be able to identify which parts of the advertisement are appealing to the audience and which are not appealing. Accordingly, practitioners not only should watch the traditional measures but also should pay attention to the duration of the peak.
Although prior research suggests that there should be a longer elapse of time prior to the peak, the long duration of a peak (something practitioners will encounter when using individuals as the unit of analysis) makes it possible for a peak to occur earlier in the most impactful advertisements. This is also the case if consumers have previous exposure to the advertisement campaign through social media. Managers should create advertisements in which the peak level of affect occurs earlier in the advertisement, lasts longer, and ends near the conclusion of the advertisement to ensure maximum effectiveness.
Limitations and Future Research
The limitations of any research project open the doors for future research and investigation into important marketing matters. A potential limitation of this research is that the advertisements tested were all Super Bowl advertisements, and there are unique characteristics associated with them. Advertisements aired during the Super Bowl are perceived as some of the most entertaining advertisements developed; consumers actually anticipate watching these advertisements and process them with high involvement (Kim et al., 2013). Although this limitation might seem to bias the results toward emphasizing emotions over cognitions, the authors still found a surprisingly strong role for advertisement- and brand-related cognitions in predicting liking and sharing behaviors on social media. Additionally, the advertisements selected for this research represented a wide range of both low- and high-involvement product categories. One nevertheless should exercise caution in generalizing these results to all advertisements, and future research should try to extend these findings to all types of advertisements.
Another potential limitation of this research is the empirical nature of the data and that rating advertisements using a computerized affect monitor does not mimic entirely the natural way consumers watch advertisements on television or social media. The affect monitor was necessary to capture the way consumers were processing the advertisements emotionally and to capture the differences in processing between consumers who had seen the advertisement before on social media and those who had not. Future research should continue to try to understand how affective and cognitive processing interrelate and work together to influence attitudes and behaviors. It is hard to see how this can be accomplished in a more natural setting. The authors urge future researchers to try to isolate and control the interrelation between these processes in a behavioral lab and examine their influences on attitudes and social-media behaviors.
ABOUT THE AUTHORS
Jennifer Lee Burton is an assistant professor of marketing at The University of Tampa. Her research focuses on the topics of persuasion, marketing communications, and social media and has been published in top journals, such as the Journal of Advertising Research and International Journal of Research in Marketing.
Kristen M. Mueller is the owner of the Accent Your Style Boutique inside Jensen Home Furnishings, located in Central Illinois. She specializes in product marketing, merchandising, buying, and selling in the home furnishings, apparel, and gift markets. Mueller worked on the research for this paper as an undergraduate student at Bradley University.
Jan Gollins was the principal and founder of the Delta Modelling Group. Delta provides advanced analytics, predictive models, and consumer-research services to consumer packaged goods and pharmaceutical companies. At the time of his death in November 2017, Mr. Gollins held adjunct faculty positions at three leading universities in Chicago—DePaul University, University of Chicago, and Loyola University.
Danielle M. Walls is the cofounder and research director of BDJ Solutions, a custom market-research company founded in 2010. Ms. Walls specializes in research design, project management, data analysis, and data visualization.
- Received May 15, 2017.
- Received (in revised form) June 11, 2018.
- Accepted July 17, 2018.
- Copyright© 2019 ARF. All rights reserved.