To build strong brands, marketing theory advises managers to follow a consistency strategy and consolidate similar themes and associations for their brands over time (Beverland et al., 2015; Kalaignanam et al., 2021; Keller, 2008; Mafael et al., 2021; Parker et al., 2018). In advertising, a consistency strategy (Becker & Gisjenberg, 2023) implies using product-typical ads, whose visual content is consistent with ad elements typical of the product category, as when vehicle advertising features outdoor scenes or bank ads show consumers in financially safe situations (Elsen et al., 2016; Pieters & Wedel, 2012; Simola et al., 2020; Wedel & Pieters, 2015).

In contrast though, marketers often use an adaptation strategy and design event-typical ads that include visual elements evoking major sporting events—such as athletes or symbolic landmarks of the host city—to highlight a link to special events. Marketers also frequently mix consistency and adaptation by relying on ads typical of both a major sporting event and the product category, which we refer to as dual-typical ads. For example, brand advertising might feature symbols of the FIFA (i.e., Fédération Internationale Football Association) World Cup or the Summer Olympics (see Web Appendix A). By definition, events are out of the ordinary (Cambridge University Press, 2024), so when marketers adapt ads to be typical of events, they necessarily deviate from the precept of brand consistency and accordingly raise several challenges to conventional marketing thought.

First, managers’ use of event-typical or dual-typical ads contradicts empirical evidence that shows that product-typical ads are the most effective form if ad exposure lasts for less than 2 s (Pieters & Wedel, 2012; Simola et al., 2020; Wedel & Pieters, 2015)—the duration of most actual encounters with ads (Elsen et al., 2016). The research by Mazodier et al. (2018) on event-typical ads does not address this challenge to knowledge. Mazodier et al. (2018) strictly focus on event-typical ads and do not examine exposure duration, thus their research does not address how the relative effectiveness of different types of typical ads varies with exposure duration. This gap leaves unanswered the question of whether event-typical ads perform better than product-typical or atypical ads (i.e., not typical of any cognitive schema—structures representing people’s knowledge about a category) during very short exposures. Product-typical ads are considered a gold standard of advertising typicality effectiveness; atypical ads greatly benefit from incremental increases in exposure duration (Elsen et al., 2016). Because previous research focuses on stimuli typical of only one cognitive schema (Elsen et al., 2016; Mazodier et al., 2018; Noseworthy et al., 2014; Peterson & Malhotra, 2023; Pieters & Wedel, 2012), we also do not know how consumers receive dual-typical ads, which match both product and event schema. As our first research question, we ask: How do event-typical and dual-typical ads compare with product-typical and atypical ads in terms of consumer responses with brief exposure durations?

Second, the mechanism underlying the impact of advertising typicality following brief exposures may not rest entirely on the feeling of knowing, or entirely on image transfer, as posited in previous research (i.e., Elsen et al., 2016; Mazodier et al., 2018). A prominent theoretical explanation for the superiority of product-typical ads at brief exposures indicates that short processing time allows for only thin slices of information to be extracted (Peracchio & Luna, 2006). Consequently, product-typical ads are easier to match with an existing schema in memory than other types of ads when time is limited, which strengthens the feeling of knowing and improves stimuli evaluation (Elsen et al., 2016). According to categorization theory (Fiske & Neuberg, 1990), by evoking an event schema, event-typical ads facilitate the transfer of attributes from the event (Mazodier et al., 2018). Additionally, event-typical ads, and dual-typical ads, match the cognitive schema of the event, which should trigger a feeling of knowing similar to product-typical ads. Furthermore, because of their inconsistency with the product schema, event-typical ads could trigger viewers’ curiosity and enhance their responses to the ads (Mandler, 1982; Noseworthy et al., 2014). We thus ask, as our second research question: Do the feeling of knowing, attribute transfer, and curiosity drive the effectiveness of event-typical and dual-typical ads relative to product-typical and atypical ads with brief exposure durations?

To address these research questions, we conduct four experiments to compare the effectiveness of ads typical of a product category (i.e., product typical), a sporting event (i.e., event typical), both (i.e., dual typical), and neither (i.e., atypical) on brand attitudes and incentive-aligned brand choice, with varying ad exposure durations (i.e., from 100 ms [msec] to 5 s). With our empirical strategy, we consider consumer exposure duration rather than opportunity to see; only the former captures the time consumers’ eyes are on the ad (McGranaghan et al., 2022). Exposure durations within the realm of thin-slicing (Peracchio & Luna, 2006) are pervasive in the field: Average consumer exposures are less than 1 s for outdoor ads (Decker et al., 2015) and less than 2 s for print ads (Pieters et al., 2010). Similarly, most exposures to online display and banner ads are less than 1 s (MarketingWeek, 2016; Pieters & Wedel, 2012). Therefore, this research applies to outdoor, print, and digital media.

We also conduct our studies in relation to three major sporting events (Olympic Games, Wimbledon Championships, and FIFA World Cup), using three sets of ads that share similar picture–text relevance, visual appeal, brand familiarity, and visual complexity, according to pretests. Studies 1a and 1b, set in labs, test the effects of 24 ads from three product categories (cars, food and beverage, skin care) on brand attitudes. After a very brief exposure (≤ 500 ms), event-typical ads are more effective than atypical ads and equally effective as product-typical and dual-typical ads; the latter only outperforms atypical ads. After a longer exposure (≥ 2 s), event-typical ads become the most effective type, and dual-typical ads perform worse than atypical ads. With a sample of adult participants recruited online, Studies 2 and 3 provide process evidence and enhance generalizability by testing different brands, ads, and events. Both studies show that, in addition to the feeling of knowing (Elsen et al., 2016), curiosity and the transfer of attributes from the event to the advertised brand drive consumer responses to event typicality. We also provide external validity checks in both studies. For Study 2, a managerial implementation conducted on Facebook during the Wimbledon Championships, shows that the odds of the event-typical ad generating clicks are between 1.29 and 1.51 greater than the odds of the other three ad types, reducing campaign costs by 33% to 63%. Furthermore, in Study 3 we demonstrate that the brand attitude findings translate to actual consumer behavior: Incentivized brand choice share for the event-typical ad is nearly 71%, compared with less than 54% for the other ad types.

Our research thus advances substantive and theoretical knowledge in the advertising domain. Substantively, we show, for the first time, that ads that are not consistent with the product category but that are adapted to major sporting events can increase incentive-aligned brand choice even with just short exposures. This insight offers pronounced opportunities for advertisers facing thin-slicing conditions. Event-typical ads offer more fertile ground for creative executions than product-typical ads, due to their rich, multidimensional schemas (Mazodier et al., 2018), whereas the schemas of the latter often are characterized by only one or two key elements (Elsen et al., 2016; Peterson & Malhotra, 2023; Simola et al., 2020). In contrast with the beliefs of marketing experts from both industry and academia that dual-typical ads (mixing consistency and adaptation) are well-appreciated by consumers (see Web Appendix B), we find that these ads never perform best, in any experiment under any exposure duration. These findings indicate that advertisers such as official event sponsors should identify the most typical elements of an event to leverage them. This precept also applies to ambushers (i.e., advertisers willing to associate their brand with the event without any legitimate link; Kelly et al., 2012) with the caveat that, unlike event sponsors who have access to registered event-related symbols such as a logo, they can only use event-related symbols that are not registered.

Theoretically, we also expand existing ad typicality frameworks by showing that ad typicality effects cannot be fully explained by a single mechanism (Elsen et al., 2016; Mazodier et al., 2018; Noseworthy et al., 2014; Wedel & Pieters, 2015). Three mechanisms drive ad effectiveness: the feeling of knowing, attribute transfer, and curiosity. These mechanisms also are contingent on exposure durations, such that very small increments (i.e., the difference between 0.5 and 2 s) can significantly alter their strengths.

Whilst we focus on adaptations to major sporting events, our findings regarding the mechanisms of event typicality suggest that managers may benefit from adapting their ads to other types of events with positive associations or those that trigger viewers’ curiosity, such as religious holidays or popular festivals, even if exposure is very short. A post-hoc study indicates that consumers can also perceive ads as being typical of Christmas or Halloween. More generally, adaptation may strengthen the effects of other marketing communication strategies, such as direct marketing (e.g., event-typical emails), owned social media (e.g., event-typical posts), or product placement (e.g., brand–plot–event connections).

Conceptual background

Ad typicality

Typical ads are consistent with consumers’ expectations, which they form on the basis of existing schemas (Goodstein, 1993; Smith & Andrews, 1995)—i.e., the cognitive categories that establish which attributes, exemplars, and affect generalize to typical members (Desai & Ratneshwar, 2003; Fiske, 1982). Typicality differs from congruity or fit, which refers to “the ‘sense’ or ‘logic’ of a particular object” (Pappu & Cornwell, 2014). Ad typicality is determined by the extent to which the content of an ad is thematically normative of a cognitive schema (Goodstein, 1993). Cognitive schemas form through repeated exposures to advertisements that feature similar elements consistently; they may reflect any kind of recurring visual elements, such as scenes, people, or landscapes (Bobrow & Norman, 1975; Callister & Stern, 2008). The concept of ad typicality is therefore rather broad (Peterson & Malhotra, 2023), but most research on the memory structure of ad schemas focuses on their typicality relative to a product category (Elsen et al., 2016). For example, car ads are typical if they feature visual elements associated with the schemas for an automotive product category, such as a car, a road, or both (Pieters & Wedel, 2012).

Just as exposures to similar ads for the same product category can produce memory representations and typicality assessments, repeated exposures to ads that feature similar event-related elements may produce typical memory representations (Mazodier et al., 2018). Therefore, specific event-related ad schemas featuring typical elements likely exist among consumers; for sporting events for instance, these may include famous monuments, landmarks associated with the host city (e.g., Big Ben for London), or depictions of athletes (Kelly et al., 2012; Mazodier et al., 2018). Since cognitive schemas are formed through repeated exposure to ads, event typicality is individual and dynamic. Specifically, a given ad may be typical of an event for one person but not for another; likewise, its typicality may change in the future (as representative elements of the event fade away and/or new ones emerge). Prior typicality research highlights the individual determinants of typicality (Loken & Ward, 1990). Hence, building on the typicality literature, we define event typicality as the degree to which an ad is perceived to be representative of ads for an event. To operationalize event typicality, scholars and managers need to monitor the event’s representative elements according to their target market, similar to how they should monitor other context-specific drivers of advertising effectiveness—e.g., authenticity, fit, or self-congruity (Mazodier & Merunka, 2012; Nunes et al., 2021; Pappu & Cornwell, 2014).

Our research focuses on four types of typical ads—product, event, dual, and atypical—according to whether an advertisement is typical of a product category and/or of an event. Figure 1 displays some example ads, which we used in the experimental studies.Footnote 1

Fig. 1
figure 1

Product-typical, event-typical, dual-typical, and atypical ad examples

Full size image

Product-typical ads

Ads can be typical of the advertised product category but not of a particular event; their elements relate solely to the product category. They quickly activate this product category, which creates interest in the ad and improves brand attitude even after very brief exposures (Elsen et al., 2016). The visual elements of the product-typical ad in the first row of Fig. 1 signal the skincare category: a prominent face displayed next to a picture of the product. No elements evoke any particular event though.

Event-typical ads

These ads are typical of an event, but they do not immediately disclose the category of the product. Like product-typical ads, they can be categorized quickly, as ads related to an event. The Volvo ad in the first row of Fig. 1 is not typical of the car category, but the starting blocks and tagline “Full throttle to Rio” evoke the 2016 Rio Olympic Games. To identify the product category, viewers must recognize the skid marks on the track as a cue that a car, rather than a runner, just left the starting blocks.

Dual-typical ads

These ads are typical of both the product category and an event, so they activate two different schemas. The dual-typical ad in the first row of Fig. 1 represents the food and drink category, with its depiction of a large Gatorade bottle (Pieters & Wedel, 2012); it also features elements evoking the 2016 Olympics, including a running track and the silhouette of the Pano de Azucar.

Atypical ads

Finally, some ads are typical of neither a product category nor an event. They might activate the product category, but doing so requires longer exposures than product-typical ads (i.e., around 2 s; Elsen et al., 2016). Because they do not feature elements from the category, the product category is not disclosed in a straightforward manner. The atypical ad in the first row of Fig. 1 does not make it immediately obvious that it promotes a car; it features neither a car nor a road. Instead, with sufficient processing, viewers can draw a parallel between the dark side of personality and the limited Blackjack edition of the Fiat 500 with its matte black exterior paint.

Table 1 shows how the current research advances understanding of ad typicality effects on consumer responses in thin-slicing conditions. Pieters and Wedel (2012), Wedel and Pieters (2015), Elsen et al. (2016), and Simola et al. (2020) focus on product-typical versus atypical ads, and Mazodier et al. (2018) focus strictly on event-typical ads; neither approach compares the effectiveness of product typicality with another type of ad schema typicality (i.e., event typicality). No studies investigate ads that tap two schemas (i.e., dual typicality). Finally, the mechanisms have been examined in silos. Instead, we integrate the feeling of knowing (Elsen et al., 2016; Pieters & Wedel, 2012) and attribute transfer (Mazodier et al., 2018) from advertising typicality research with curiosity from schema congruity research (Hutter & Hoffmann, 2014; Noseworthy et al., 2014) in an ad categorization framework.

Table 1 Relevant research on ad typicality in thin-slicing conditions (i.e., short exposure durations)
Full size table

Ad categorization framework and empirical predictions

The ad categorization framework in Fig. 2 Panel A illustrates the relevant processing mechanisms that apply to different types of ads in thin-slicing conditions. The categorization process resulting from ad typicality activates the feeling of knowing, attribute transfer, and curiosity, the intensity of which vary across exposure durations. In turn, these mechanisms drive consumer responses. Brand attitude is a well-established predictor of brand choice (Watson et al., 2015). We expect the feeling of knowing, attribute transfer, and curiosity to be proximal mediators and brand attitude to be a distal mediator in the serial mediation of the effects of event typicality, product typicality, and exposure durations on incentive-aligned brand choice, as detailed next. Panels B and C of Fig. 2 graphically display the hypotheses resulting from the comparison of event-typical ads with product-typical ads (i.e., H1a-b and H2a-c) and of event-typical ads with atypical ads (H3a-b and H4a-c), respectively.

Fig. 2
figure 2

Ad categorization framework and hypotheses

Full size image

Effects of event-typical vs. product-typical ads

At the moment of exposure, viewers focus on identifying the ad (Pieters & Wedel, 2012). In thin-slicing conditions, uncertainty-reducing information accumulates rapidly until the viewer experiences a subjective sense of knowing the identity of the stimuli, which then drives evaluative judgments (Pouget et al., 2016). This feeling of knowing can arise very quickly; visual research shows that people feel confident about others’ traits (e.g., competence, aggressiveness, likeability) after only 100 ms of exposure to their physical appearance (Willis & Todorov, 2006). In addition, complex scenes can be categorized accurately within a single eye fixation—about 250 ms—such as people running on a track (sporting event) or playing musical instruments (stage performance) (Fei-Fei et al., 2007). The ad identification process on the left-hand side of Fig. 2 Panel A details how consumers match the ad with the schemas they have in memory. They seek a matching product category schema in their memories to identify the ad, but even if the ad does not match a product schema, it can induce a feeling of knowing if it matches another schema that enables identification, such as a sporting event. Consumers regularly encounter ads typical of product categories (Elsen et al., 2016; Simola et al., 2020), so the product category is the most natural schema for an ad (Stayman et al., 1992). The schemas of major events may be less salient; consumers only are exposed to event-typical ads during major events, such as once a year (e.g., Wimbledon Championships, Super Bowl) or less (e.g., four years for Summer Olympics and FIFA World Cup). The feeling of knowing induced by event-typical ads then may be more moderate than that evoked by product-typical ads.

Event-typical ads are more likely to benefit from curiosity and event attribute transfer than product-typical ads (Fig. 2 Panel A). According to the schema congruity effect (Mandler, 1982), curiosity is aroused when incongruity or disconfirmed expectations can be resolved (Daume & Hüttl-Maack, 2020; Noseworthy et al., 2014). Curiosity is the need to acquire new knowledge to bridge an information gap (Loewenstein, 1994), leading to positive feelings toward the stimuli that generate it (Daume & Hüttl-Maack, 2020). Event-typical ads may disconfirm what consumers expect, because they do not match a product schema, which is the natural expectation for ads (Stayman et al., 1992). However, this product schema incongruity seems solvable, because the ad can be identified through another theme, namely, its sporting event schema. Schema congruity theory thus predicts that event-typical ads enhance curiosity. In contrast, product-typical ads conform to product category codes and do not contradict expectations (Elsen et al., 2016), which limits curiosity.

Furthermore, as viewers match event-typical ads to sporting event schemas, they engage in category-based processing (Fiske, 1982), and the knowledge associated with the event can transfer to the ads. If viewers can successfully assign an object to a category using salient visual cues (Barsalou, 1992), the attributes of the category transfer to the object (Fiske & Neuberg 1990; Goodstein, 1993). Ads that depict typical elements of an event but not elements typical of the product category thus should be categorized with the event, leading to the transfer of its attributes. Sporting event attributes tend to be positive (Mazodier et al., 2018), so this transfer should improve responses to the ad and brand. In turn, event-typical ads may be more effective than product-typical ads. Although they trigger milder feelings of knowing, they benefit from more curiosity and more attribute transfer (Fig. 2 Panel B). We thus hypothesize:

H1

Event-typical ads affect (a) brand attitude and (b) incentive-aligned brand choice more positively than product-typical ads do.

H2

The positive impact of event-typical (cf. product-typical) ads results from (a) a negative indirect effect through the feeling of knowing, (b) a positive indirect effect through curiosity, and (c) a positive indirect effect through transfer of event attributes.

Effects of event-typical ads vs. atypical ads

Atypical ads feature different or unusual elements (Pieters & Wedel, 2012; Simola et al., 2020). Unlike event-typical ads, after very brief exposures, they are less likely to be categorized into existing schemas, hindering the feeling of knowing (Elsen et al., 2016). Yet Ruan et al. (2018) indicate that uncertainty about what the ad is about can engender curiosity if it stimulates viewers’ motivation to know more (Kashdan & Silva, 2009; Litman, 2005). Accordingly, atypical ads could trigger curiosity. Unlike event-typical ads though, the exposure duration needs to allow for a closer inspection, because atypical ads do not activate any cognitive schema. Their incongruity in turn is harder to resolve, and thin slices of information—resulting from less than 5 s of processing (Peracchio & Luna, 2006)—may not be enough to appreciate their logic. Therefore, in thin-slicing conditions, event-typical ads may trigger more curiosity than atypical ads, because short durations likely are sufficient to realize that the product schema incongruity is solvable, leading viewers to want more information (Kashdan & Silva 2009; Litman, 2005). This outcome is less likely for atypical ads. Also, unlike event-typical ads, atypical ads cannot be assigned to an event category, which may prevent attribute transfer. In summary, in thin-slicing conditions, event-typical ads should lead to more favorable consumer responses than atypical ads, due to a stronger feeling of knowing, curiosity, and attribute transfer (Fig. 2 Panel C). Thus:

H3

Event-typical ads affect (a) brand attitude and (b) incentive-aligned brand choice more positively than atypical ads do.

H4

The positive impact of event-typical (cf. atypical) ads results from (a) a positive indirect effect through the feeling of knowing, (b) a positive indirect effect through curiosity, and (c) a positive indirect effect through transfer of event attributes.

Effects of exposure duration

The effects of product-typical and atypical ads depend on ad exposure duration. Atypical ads perform worse than product-typical ads after very brief exposures (< 2 s) but equivalently after 2-s exposures (Elsen et al., 2016), and then they generate stronger brand preferences after 5-s exposures (Simola et al., 2020). The effects of event-typical ads also should be contingent on exposure durations, though the precise influences remain an empirical question (Fig. 2 Panel A). Increasing exposure duration to event-typical ads—while still within the thin-slicing range—may have opposing effects on the mechanisms that drive their impact. First, the feeling of curiosity is transient (Loewenstein, 1994); people remain curious only as long as they “almost know” the logic of the ad and feel compelled to close the knowledge gap (Noordewier & Van Dijk, 2017). Beyond very brief exposures, the curiosity triggered by event-typical ads may stagnate and decline, once consumers bridge the knowledge gap and fully understand the ads’ rationale (i.e., the link to the event). Second, longer exposure durations may intensify category-based processing (Fiske, 1982), which facilitates attribute transfer and likely boosts consumer responses. For this initial investigation of event-typical ads in varying thin-slicing conditions, it would be arbitrary of us to posit an exposure duration threshold for when the effects of event-typical ads change due to stronger/weaker curiosity or attribute transfer. Instead, we test H1–H4 across multiple exposure durations within a thin-slicing standard (100 ms, 500 ms, 2 s, and 5 s) and propose two empirical questions:

  • EQ1 What is the impact of exposure duration on the effects of event-typical ads versus product-typical and atypical ads?

  • EQ2 What is the impact of exposure duration on the indirect effects of event-typical ads, through the feeling of knowing, curiosity, and attribute transfer?

Effects of dual-typical ads

Contrary predictions are possible, regarding how well consumers respond to dual-typical ads. According to managers and academics we surveyed (see Web Appendix B), dual-typical ads accumulate benefits from matching both the product and the event schema, resulting in a strong feeling of knowing and attribute transfer. But the schema of dual-typical ads also may be ambiguous (Berlyne, 1960; Gregan-Paxton et al., 2005), which would undermine the feeling of knowing and attribute transfer, while bolstering curiosity. The ambiguity of dual schema activation may hinder ad identification (Fei-Fei et al., 2007; Goode et al., 2013) and categorization (Fiske & Neuberg, 1990), lowering the feeling of knowing and attribute transfer. Nevertheless, stimuli with multiple meanings also can trigger curiosity, such that ambiguous advertising slogans induce stronger curiosity than non-ambiguous ones (Daume & Hüttl-Maack, 2020). Considering the paucity of research on marketing stimuli with multiple schemas and the opposing forces on the mechanisms driving the effects of dual typicality, the comparison of ad types evokes an empirical question (Fig. 2 Panel A):

  • EQ3 What is the impact of dual-typical ads on brand attitude and incentive-aligned brand choice in comparison with product-typical, event-typical, and atypical ads?

  • EQ4 What role do the feeling of knowing, curiosity, and attribute transfer play in driving consumer responses to dual-typical ads in relation to other ad types?

Study 1

Stimuli development

In Study 1a, we operationalize ad event typicality with the 2016 Rio Summer Olympics, a very popular event (Gijsenberg, 2014). Product-typical ads include typical elements of the food and beverage, car, and skincare product categories, as identified by Pieters and Wedel (2012). Event-typical ads feature typical elements of the Summer Olympics, such as athletes and landmarks of the host city (Mazodier et al., 2018). Atypical ads do not feature either event-typical or product-typical elements; dual-typical ads include both types.

For generalizability, we created two ads for each of the three product categories and each of the four ad types, so we had 24 target ads for 24 different brands. We hired a professional graphic designer and shared a list of the typical elements of the three product categories and the event established in prior research (Mazodier et al., 2018; Pieters & Wedel, 2012). We also provided a list of 24 brands (8 per product category) whose familiarity level was estimated to be similar. The designer created 24 ads (including the examples in Fig. 1) according to our conceptualization of the four ad types and the event- and product-typical elements we provided; he also integrated the logos of the brands. Furthermore, we asked the graphic designer to create ads with similar picture–text relevance, visual appeal, and visual complexity (Heckler & Childers, 1992; Pieters et al., 2010). The ad copywriting ensured consistent picture–text relevance across the ads. A separate pretest with 24 students confirmed that the ads featured appropriate levels of product and event typicality (see Web Appendix C, Tables C1a and C1b). For example, the product category typicality of both product-typical and dual-typical ads was rated higher than that of the event-typical and atypical ads; the event typicality of both event-typical and dual-typical ads was also rated higher than that of product-typical and atypical ads (ps < 0.05). As intended, the four ad types did not differ in picture–text relevance, visual appeal, brand familiarity, or visual complexity.

Design, procedure, and measures

Two hundred sixty-eight undergraduate students from an Australian University (57.5% women) received course credit for their participation in a lab study, two years after the 2016 Rio Olympics. Each lab session included 10–15 participants, randomly assigned to one condition in a 3 × 2 × 2 mixed design with three exposure durations (100 ms, 2 s, and 5 s) as the between-subjects factor and product category typicality (yes, no) and event typicality (yes, no) as within-subject factors. An exposure of 100 ms is shorter than the duration of an eye fixation but sufficient to enable identification of the product in product-typical ads (Elsen et al., 2016; Pieters & Wedel, 2012), as well as objects or scenes from pictures (Grill-Spector & Kanwisher, 2005). A 2-s exposure is longer than two eye fixations and equivalent to the average time that consumers attend to print ads (Pieters et al., 2010). The 5-s exposure allows for a more detailed inspection of the ad. The design also included, as another between-subjects factor, prior exposure of the participants to either pictures of the 2016 Rio Olympic Games (e.g., athletes competing, different from the Rio and Brazil landmarks featured in the ads) or pictures from the National Geographic Channel (e.g., mountain, sea). We tested whether priming participants’ consideration of the event could affect the results. (We also verified that the Olympics and National Geographic Channel pictures did not affect moods differently.) Priming did not affect the results, so we exclude this factor in subsequent studies. (For full results, see Web Appendix D, which covers how we tested for alternative explanations, such as priming.)

In accordance with prior advertising and vision research (Elsen et al., 2016; Grill-Spector & Kanwisher, 2005), participants viewed the 24 ads on personal computers (24-inch screens, viewing distance approximately 25 inches); they had three practice trials with non-target ads to familiarize themselves with the task before proceeding to the 24 target ads. The instructions appeared first, before the practice and target trials. Then, for each trial, a fixation cross appeared on a blank screen for 900 ms to ensure that their viewing positions would be similar when processing the ads. One of the 24 ads randomly appeared (almost full screen), according to the duration assigned by the respective between-subjects condition. Then, to interrupt visual processing before participants proceeded to the ad rating task, a backward visual mask appeared (80 ms). To mark transitions between ad trials, a black screen appeared for 300 ms. The measure of the contribution of the ad to participants’ brand attitudes featured a seven-point scale, with the prompt “Due to this ad, my evaluation of the brand has…” and response options ranging from 1 = “become more negative” to 4 = “not changed” to 7 = “become more positive” (Elsen et al., 2016). The participants also rated their attitude toward the ad on a seven-point scale: “Your thoughts and feelings toward this ad are…” 1 = “negative” to 7 = “positive,” and indicated their immediate interest in the ad as a binary no–yes response to the question, “Is this ad interesting to you?” (Pieters & Wedel, 2012). The results for brand attitude, ad attitude, and ad interest are similar in all studies; we report the results for brand attitude in this manuscript (ad attitude and ad interest results are available at https://researchbox.org/1426&PEER_REVIEW_passcode=MAWOVL).

Results

All the data analyses rely on a Bayesian extension of an analysis of variance (BANOVA) (Wedel & Dong, 2019). The BANOVA Markov chains converged in all the models (Geweke, 1992).Footnote 2 We provide the model parameter estimates in Web Appendix E; the predicted means for brand attitude at each exposure time are in Table 2. We plot the effects on brand attitude in Fig. 3, Panel A.Footnote 3

Table 2 Predicted means as a function of event typicality, product typicality, and exposure duration (100 ms, 500 ms, 2 s, 5 s)
Full size table
Fig. 3
figure 3

Ad typicality and exposure duration effects

Full size image

In line with H1a and H3a, across exposures, event-typical ads (M = 4.91) affect brand attitude more positively than product-typical ads (M = 4.57, simple effect = 0.34 [0.22, 0.47]) and atypical ads (M = 4.66, simple effect = 0.25 [0.12, 0.38]). Furthermore, event-typical ads outperform dual-typical ads (M = 4.54, simple effect = 0.37 [0.24, 0.49]). Dual-typical, atypical, and product-typical ads do not differ significantly (ps > 0.05).

As Fig. 3 reveals, brand attitude after exposure to the product-typical ads is flat across exposure durations, but brand attitude after exposures to the event-typical, atypical, and dual-typical ads exhibits upward trajectories. Most improvements in brand attitude take place between 100 ms and 2 s. After 100 ms, brand attitude after exposure to event-typical ads (M = 4.53) is not significantly different from product-typical ads (M = 4.46, simple effect = 0.07 [-0.16, 0.29]). After 2 s however, event-typical ads (M = 5.04) outperform product-typical ads (M = 4.57, simple effect = 0.48 [0.27, 0.69]), and this pattern continues after 5 s (Mevent typical = 5.15; Mproduct typical = 4.67, simple effect = 0.48 [0.26, 0.70]).

Compared with atypical ads (M = 4.03), event-typical ads perform better after 100 ms (M = 4.53, simple effect = 0.49 [0.27, 0.73]), but atypical ads no longer differ after 2 s (Mevent typical = 5.04; Matypical = 4.93, simple effect = 0.12, [-0.09, 0.33]) and beyond (5 s: Mevent typical = 5.15; Matypical = 5.01, simple effect = 0.14, [-0.08, 0.36]). In response to our first empirical question, longer exposure durations allow the effectiveness of event-typical ads to strengthen relative to that of product-typical ads, but it weakens relative to atypical ads.

Dual-typical ads spark positive brand attitude after 100 ms (M = 4.35), outperforming atypical ads (simple effect = 0.32 [0.09, 0.54]). They are on par with event-typical ads (simple effect = -0.18 [-0.40, 0.05]) and product-typical ads (simple effect = -0.11 [-0.34, 0.12]). After 2 s, dual-typical ads (M = 4.62) fall behind event-typical ads (simple effect = -0.42 [-0.63, -0.23]) and atypical ads (simple effect = -0.31 [-0.52, -0.09]), but they are comparable to product-typical ads (simple effect = 0.05 [-0.16, 0.26]). These positions persist after 5 s: Dual-typical ads (M = 4.66) are inferior to event-typical ads (simple effect = -0.50 [-0.71, -0.28]) and atypical ads (simple effect = -0.35 [-0.58, -0.13]) but not different from product-typical ads (simple effect = -0.02 [-0.24, 0.20]). Addressing our third empirical question, dual-typical ads do not outperform the other ads and benefit less from longer exposures than atypical and event-typical ads do.

In summary, after 100 ms, event-typical, product-typical, and dual-typical ads perform best; after 2 or 5 s, event-typical ads yield the highest brand evaluations. Regardless of exposure durations, event-typical ads always evoke the best evaluations. In Study 1b, to replicate our findings from Study 1a, we also include a 500-ms exposure condition to pinpoint more precisely at what duration, beyond 100 ms, the effects of a longer exposure materialize. We omitted the 5-s condition, for which the results did not change relative to 2 s. Furthermore, as in our subsequent studies, Study 1b did not include the priming factor. The results (detailed in Web Appendix F) are analogous: Exposure duration produces its effect between 500 ms, which implies two eye fixations and is the minimum to synthesize information from spatially distinct areas (Rayner & Castelhano, 2007), and 2 s.

Study 2

At very brief exposures, evaluations of event-typical ads do not differ significantly from those of product-typical ads or dual-typical ads but are more positive than those of atypical ads. At longer exposure durations, event-typical ads surpass all the other ad types. With Study 2, we investigate the processes responsible for these findings, with the prediction that the feeling of knowing, attribute transfer, and curiosity are underlying mechanisms. Whereas pretests confirmed that the ads and brands in Study 1 have similar characteristics, all the ad stimuli in Study 2 feature the same brand, “Lunch Garden Restaurants,” which does not operate in the region where the study participants live. This precaution eliminates brand difference by design and ensures that participants’ responses are not biased by prior brand knowledge.

Method

Four hundred four participants were compensated for their time, as members of the Prolific platform residing in the United Kingdom, and were randomly assigned to one condition of a 2 × 2 × 2 experimental design, with exposure durations (100 ms, 2 s), event typicality (yes, no), and product typicality (yes, no) as between-subject factors. The procedure is similar to that in Study 1, except that the study was conducted online rather than in the lab, and each participant saw one of the four target ads and eight filler ads. The experiment ran on July 11, 2021, during the final day of the Wimbledon Championships in England—one of the most prestigious tennis tournaments in the world. To determine which attributes are susceptible to being transferred, we performed a pretest of the strength of the association between the Wimbledon Championships and several attributes that characterize sporting events (Carrillat et al., 2010; Mazodier et al., 2018) on a 5-point Likert-type scale (“Please rate your degree of agreement or disagreement regarding whether each of the adjectives below reflects the Wimbledon Championships as a brand”). The results from 51 Prolific participants showed that “iconic” (M = 4.66, SD = 0.63) and “popular” (M = 4.53, SD = 0.64) were most strongly associated with the event (their average ratings are also significantly greater than the fourth point of the scale corresponding to the “Agree” response: ts > 5.88, ps < 0.001).

To specify the underlying processes driving brand evaluations, we also measured three mediators: (1) attribute transfer, operationalized as the extent to which the ad is perceived as popular and iconic, using the pretest scale; (2) curiosity, adapting a scale from Ruan et al. (2018) that is anchored at 1 = “not at all” and 7 = “very much” (“Did you feel curious when viewing the ad?”); and (3) the feeling of knowing, adapted from Elsen et al. (2016). Respondents first identified the theme of the ad from six categories presented randomly on screen (“Please indicate which theme below is the most representative of the last ad you saw: Food and Beverage, Cars, Skin care, Wimbledon Championships, FIFA World Cup, Others”). For the feeling of knowing (Pleskac & Busemeyer, 2010), participants indicated their certainty of knowing the theme of the ad (“Please indicate your degree of certainty regarding your response to the previous question”), from 1 = “absolutely not certain” to 7 = “absolutely certain.” As before, ad and brand attitudes were the dependent variables.

We hired the same graphic designer as for Study 1 and provided similar guidelines (e.g., list of typical visual elements of the Wimbledon Championships and restaurants) to develop four new ads, one for each of the ad types (see Fig. 1, row 2). A separate pretest with 31 paid Prolific participants confirmed that the ads appropriately manipulated product and event typicality (ps < 0.001) but did not differ on picture–text relevance, visual appeal, or visual complexity (see Web Appendix C, Tables C2a and C2b).

Results

Effects on brand attitude, feeling of knowing, attribute transfer, and curiosity

Table 2 displays the predicted means by ad type and exposure duration. The plots of the feeling of knowing, attribute transfer, and curiosity in Fig. 3, Panels B–D, confirm the mechanisms illustrated in our ad categorization framework (Fig. 2, Panel A). We provide the model parameter estimates in Web Appendix G, Panel A.

The results confirm the superiority of the event-typical ad, consistent with H1a and H3a, as well as the lackluster performance of the dual-typical ad. Across exposure durations, event-typicality (M = 4.48) affects brand attitudes more positively than product typicality (M = 3.95, simple effect = 0.53 [0.26, 0.80]), atypicality (M = 3.89, simple effect = 59. [0.32, 0.86]), or dual typicality (M = 4.20, simple effect = 0.29 [0.02, 0.55]). As in Study 1a, dual typicality is equivalent to product typicality (p > 0.05), but dual typicality influences brand attitude more positively than atypicality does (simple effect = 31. [0.04, 0.57]).

Comparing the effects of 100-ms and 2-s exposures, we find that brand attitude for the event-typical ad is more positive than for the product-typical ad after both 100 ms (Mevent typical = 4.38 vs. Mproduct typical = 3.90, simple effect = 0.47 [0.09, 0.86]) and 2 s (Mevent typical = 4.59 vs. Mproduct typical = 4.00, simple effect = 0.59 [0.22, 0.96]). Unlike Study 1a, the positive impact on brand attitude of the event-typical ad compared with the product-typical ad is significant after a 100-ms exposure. We address this discrepancy subsequently, with mediation analyses. The same pattern emerges from the comparison with the atypical ad (100 ms: Mevent typical = 4.38, Matypical = 3.72, simple effect = 0.65 [0.27, 1.04]; 2 s: Mevent typical = 4.59, Matypical = 4.06, simple effect = 0.53 [0.17, 0.89]). After 100 ms, the dual-typical ad is superior to the atypical ad (Mdual typical = 4.14, Matypical = 3.72, simple effect = 0.42 [0.04, 0.80]) but on par with the other ad types (event-typical: simple effect = -0.24 [-0.63, 0.15]; product-typical: simple effect = 0.24 [-0.14, 0.62]). After 2 s, the dual-typical ad is marginally less effective than the event-typical ad (Mdual typical = 4.25, Mevent typical = 4.59, simple effect = -0.33 [-0.69, 0.02], p = 06) and on par with the product-typical (simple effect = 0.25 [-0.10, 0.61]) and atypical (simple effect = 0.20 [-0.16, 0.55]) ads. That is, the pattern of results is consistent with the findings of Studies 1a and 1b.

Mediation analyses

The effects of the event and product typicality factors, as well as of exposure durations, on the dependent variable when we include the feeling of knowing, attribute transfer, and curiosity in the model provide evidence that all three mediators enhance brand attitudes (feeling of knowing: \(b\) = 0.30 [0.17, 0.42]; attribute transfer: \(b\) = 0.77 [0.57, 0.98]; curiosity: \(b\) = 0.30 [0.16, 0.44]), but attribute transfer exhibits the strongest effect (non-overlapping credible interval [CI]). The indirect effects in Table 3, computed with the BANOVA R package (Wedel & Dong, 2019), confirm that the feeling of knowing, attribute transfer, and curiosity all drive the impacts of event typicality and product typicality on brand attitudes across exposure durations.

Table 3 Effects on mediators (\(a\)) and indirect effects (\(a\times b\)) on brand attitude: Study 2
Full size table

Event-typical vs. product-typical ads

In line with H2a and H2c, the feeling of knowing and attribute transfer act as opposite forces driving the effect of the event-typical ad relative to the product-typical ad. Across exposure durations, the indirect effect through the feeling of knowing is negative (\(a\times b\) = -0.46, p < 0.001)—it lowers the effect of the event-typical ad on brand attitude relative to the product-typical ad—whereas that of attribute transfer is positive (\(a\times b\) = 0.75, p < 0.001)—it improves the effect of the event-typical ad on brand attitude relative to the product-typical ad. The indirect effect through curiosity is not significant (\(a \times b\) = 0.11, p = 0.12), so we cannot confirm H2b. As mentioned, attribute transfer explains more variance in brand attitude (b = 0.77) than the feeling of knowing (b = 0.30). Therefore, the stronger positive indirect effect through attribute transfer overcomes the negative indirect effect through the feeling of knowing, as reflected by more positive brand attitudes after exposure to the event-typical compared with the product-typical ad (p < 0.001). As reported in Table 3 and illustrated in Fig. 4, Panel A, the indirect effect through the feeling of knowing, attribute transfer, and curiosity remains stable between the two exposure durations. This result also suggests that the more positive impact of the event-typical ad on brand attitude, compared with the product-typical ad, after 100 ms is due to a stronger attribute transfer in Study 2 than in Study 1. The Wimbledon tournament takes place every year in the same location, whereas the Summer Olympics occur every four years, in different locations, which may decrease the salience and homogeneity of event attributes.

Fig. 4
figure 4

Indirect effects through feeling of knowing, attribute transfer, and curiosity on brand attitude: Study 2

Full size image

Event-typical vs. atypical ads

As expected, across exposure durations, the superiority of the event-typical ad compared with the atypical ad (p < 0.001) reflects the stronger attribute transfer and feeling of knowing (\(a\times b\) = 0.82, p < 0.001 \(; a\times b\)= 0.13, p = 0.05, respectively), but the indirect effect through curiosity is not significant (\(a\times b\)= -0.02, p = 0.82). These results are in line with H4a and H4c but fail to support H4b. Table 3 and Fig. 4, Panel C, show that a longer exposure duration (i.e., 2 s rather than 100 ms) increases the positive indirect effect of the event-typical ad compared with the atypical ad through attribute transfer; the indirect effect through the feeling of knowing becomes non-significant. Furthermore, the indirect effect through curiosity becomes negative and significant (i.e., the atypical ad triggers greater curiosity than the event-typical ad after a 2-s exposure; Table 2). The stronger indirect effect through attribute transfer compared with curiosity at 2 s (Δ \(a\times b\)= 1.01) indicates that the event-typical ad outperforms the atypical ad (p = 0.004). For our second empirical question, we note that exposure duration exerts a stronger contingency effect in comparing event-typical with atypical ads, rather than with product-typical ads.

Dual-typical ad vs. others

Addressing our fourth empirical question, we find that the effects of attribute transfer, curiosity, and the feeling of knowing are contingent on exposure durations. The varying influences between 100 ms and 2 s reveals why the dual-typical ad outperforms the atypical ad at the former duration (p = 0.03), whereas they do not differ for the latter duration (p = 0.28). The positive indirect effect through the feeling of knowing weakens between 100-ms (\(a\times b\) = 0.75 [0.40, 1.15], p < 0.001) and 2-s (\(a\times b\) = 0.18 [-0.01, 0.39], p = 0.06) exposures (non-overlapping CIs); curiosity is not significant at the shorter duration (\(a\times b\) = 0.14, p = 0.15) but is negative and significant at the longer duration (\(a\times b\) = -0.18, p = 0.05); and attribute transfer tends to grow with longer exposures (100 ms: \(a\times b\) = 0.34, p = 0.08; 2 s: \(a\times b\) = 0.49, p < 0.01). In addition, the event-typical ad marginally outperforms the dual-typical ad after 2 s (-0.33 [-0.69, 0.02]); p = 0.06), but no difference occurs at 100 ms (-0.24 [-0.63, 0.15]), mainly due to the positive indirect effect through attribute transfer, which is significant after 2 s (\(a\times b\) = 0.90, p < 0.01) but not after 100 ms (\(a\times b\) = 0.41, p = 0.18). It compensates for the significant negative effect of the feeling of knowing at the longer exposure (100 ms: \(a\times b\) = -0.04, p = 0.80; 2 s: \(a\times b\) = -0.34, p = 0.02) (Fig. 4, Panel B).

External validity check

To demonstrate the external validity and the managerial implications of our findings, we tested the “Lunch Garden Restaurants” ads during the 2021 and 2022 sessions of the Wimbledon tournament. We launched an advertising campaign through Facebook’s targeting settings and randomly delivered the four types of ads created for Study 2 to adult U.K. residents interested in tennis. Although exposure time could not be controlled in this setting, the durations likely represented thin-slicing, considering evidence that 91% of digital display ads are gazed at for less than 1 s (MarketingWeek, 2016). We first distributed the ads during July 10–14, 2021. The number of clicks was deemed too low (i.e., 21 for the event-typical ad, 12 for the atypical and product-typical ads, and 11 for the dual-typical ad), so we extended the data collection by distributing the ads again the following year for the entire duration of the tournament (i.e., June 27–July 10, 2022), displayed only in Facebook News feeds. People who clicked on the ads were directed to a Facebook page, explaining the purpose of the study and disclosing that “Lunch Gardens Restaurant” was unavailable in their region. There was no evidence that the click rates representing the 2021 and 2022 samples come from different populations for any of the four ads (ps > 0.16), so we merged the data.

The total reach varied between 97,228 and 101,607 across ad types. Seventy-seven (77) people clicked on the event-typical ad, 58 on the product-typical ad, 51 people clicked on the dual-typical ad, and 49 on the atypical ad. The performance of the advertisements confirmed our prior results. The odds of clicking on the ad typical of the Wimbledon tournament were 1.46 times greater than those of clicking on the dual-typical ad (odds ratio CI95% [1.02, 2.07]) and 1.51 times greater than those of clicking on the atypical ad ([1.05, 2.15]). The cost-per-click for the event-typical ad ($1.44) was 51% and 63% lower than the cost of the dual-typical ($2.18) and atypical ($2.35) ads, respectively. The event-typical ad yielded a 33% lower cost-per-click than the product-typical ad ($1.92), though the odds ratio in favor of the former ad (1.29) did not reach statistical significance (i.e., the CI spanned 1 [0.92, 1.81]). This result aligns with controlled exposure duration findings of a greater impact of event-typical versus product-typical ads after exposures longer than 500 ms. In summary, the benefits of event-typical ads extend to the field.

Study 3

We extend our findings in three ways with Study 3. First, to confirm external validity, we measure participants’ actual decisions between two real brands with an incentive-aligned brand choice design. Second, to improve internal validity, we manipulate product and event typicality with a single visual element each. Third, to increase generalizability, we feature a new brand, new ads, and a new event.

Design

Three hundred eighty-seven participants (50.1% women, average age 40 years) were recruited in France through an online panel company (Made in Surveys), one month after the 2022 FIFA World Cup. They were randomly assigned to one condition of a 2 × 2 × 2 experimental design with exposure durations (500 ms, 2 s), event typicality (yes, no), and product typicality (yes, no) as between-subject factors. The shorter exposure duration of 500 ms rather than 100 ms tests the generalizability of the effects when consumers have the opportunity for two eye fixations.

The procedure was similar to that of Study 2, except that we added an incentive-aligned brand choice, such that participants had to choose between two real German apparel brands—Merz b. Schwanen and Schiesser Revival—that are largely unknown in France. We manipulated the ad for Merz b. Schwanen, the target brand, and kept the ad for Schiesser Revival constant (i.e., actual ad featuring the brand as the benchmark). Participants had incentives to reveal their true preferences, because we indicated they would be entered in a lottery, and the winner would receive a T-shirt from their selected brand. We collected the same dependent measures as in our previous studies. A pretest confirmed that “popular” (M = 4.53, SD = 0.64) and “iconic” (M = 3.40, SD = 1.34) are the attributes most strongly associated with the FIFA World Cup (ps < 0.07). We also test the plausibility of two alternative theoretical accounts for the underlying mechanisms: priming and confirmation/disconfirmation. We measured participants’ skepticism to capture the former and their response latency for the latter; response latency is strongly linked to fluency, a manifestation of expectation confirmation (Schwarz, 2004). Web Appendix D provides the theoretical arguments and full test results.

The same graphic designer created four different ads for Merz b. Schwanen, similar to two real ads used by the brand. One ad showed a male model wearing a white T-shirt from the front (typical for the product category); another showed the same male model wearing the same T-shirt from the back (atypical for the product category). The graphic designer also created an event-typical background (i.e., painting-like silhouettes of two soccer players on each side of the model) and a neutral background (i.e., painting-like silhouettes of two tree branches of the same size and color as the soccer players’ silhouettes and located in the same area of the ad), which were combined with the front- or back-facing model to create product-typical (front + tree), event-typical (back + soccer players), dual-typical (front + soccer players), and atypical (back + tree) versions (see last row of Fig. 1). All ads featured the same slogan: “Nothing beats attitude—Turn your back on fast fashion.” A within-subject pretest with 40 respondents confirmed that the ads had the desired levels of product and event typicality but did not differ on picture–text relevance, visual appeal, or visual complexity (Web Appendix C, Tables C3a and C3b).

Results

Effects on incentive-aligned brand choice and brand attitude

The incentive-aligned brand choice mirrored our brand attitude findings, and the model results are consistent with our previous studies too (model parameter estimates are in Web Appendix G, Panel B). Across exposure durations, 70.9% of respondents chose Merz b. Schwanen over Schiesser in the incentivized task when it was featured in the event-typical ad, compared with 53.8% for the product-typical, 43.1% for the atypical, and 41.7% for the dual-typical ads. In support of H1b and H3b, across exposure durations, the event-typical ad affected incentive-aligned brand choice more positively than the product-typical ad (simple effect = 0.17, [0.03, 0.30]) and atypical ad (simple effect = 0.27, [0.14, 0.41]). Specifically, no significant difference arose after 500 ms (simple effect = 0.08, [-0.12, 0.27]), but the event-typical ad outperformed the product-typical ad after 2 s (simple effect = 0.25 [0.06, 0.44]). Incentive-aligned brand choice after exposure to the event-typical ad was significantly higher than after the atypical ad, at both 500 ms (simple effect = 0.34 [0.14, 0.54]) and 2 s (simple effect = 0.20 [0.02, 0.39]). The brand attitude results (Table 2) were consistent with the incentive-aligned brand choice.

These results confirm the underwhelming performance of the dual-typical ad. Compared with the event-typical ad, its effect on incentive-aligned brand choice is weaker at 500 ms (simple effect = -0.19 [-0.39, 0.01], p = 0.06) and 2 s (simple effect = -0.38 [-0.57, -0.20]), even if marginally so with the former exposure duration. The dual-typical ad is on par with the product-typical ad across exposure durations (500 ms: simple effect = -0.11 [-0.31, 0.09]; 2 s: simple effect = -0.13 [-0.32, 0.06]), and with the atypical ad after 500 ms (simple effect = 0.15 [-0.05, 0.36]), but it is marginally weaker than the atypical ad after 2 s (simple effect = -0.18 [-0.37, 0.01], p = 0.06).

Mediation analyses

As in Study 2, the three mediators enhanced brand attitudes (feeling of knowing \(b\) = 0.14 [0.09, 0.18]; attribute transfer \(b\) = 0.29 [0.21, 0.37]; curiosity \(b\) = 0.07 [0.01, 0.13]). As expected, brand attitude enhanced incentive-aligned brand choice (\(b\) = 0.13 [0.07, 0.18]).

In a moderated serial multiple mediation analysis, event typicality (0 = no; 1 = yes) is the independent variable; product typicality (0 = no; 1 = yes) and exposure duration (0 = 500 ms; 1 = 2 s) are the moderators; incentive-aligned brand choice is the dependent variable; the feeling of knowing, curiosity, and attribute transfer are stage-one (proximal) mediators; and brand attitude is the stage-two (distal) mediator.Footnote 4 The serial indirect effects through brand attitudes, in Table 4 and Fig. 5, align with the indirect effects documented in Study 2.

Table 4 Effects on mediators (\(a\)) and serial indirect effects (\(a\times b\times c\)) on incentive-aligned brand choice: Study 3
Full size table
Fig. 5
figure 5

Serial indirect effects through feeling of knowing, curiosity, and attribute transfer and then through brand attitude on incentive-aligned brand choice: Study 3

Full size image

Across exposure times, for the event-typical versus product-typical ad, the serial negative indirect effect through the feeling of knowing is significant (\(a \times b \times c\)= -0.019, p < 0.001); it is positive and significant for attribute transfer (\(a \times b \times c\)= 0.049, p < 0.001) and curiosity (\(a \times b \times c\)= 0.007, p = 0.02). That is, the three serial mechanisms are significant and in the expected direction, in support of H2a–H2c. As in Studies 1a and 1b, but not Study 2, the difference between the event-typical ad and the product-typical ad is not significant after a very brief exposure (≤ 500 ms), driven by the negative serial indirect effect through the feeling of knowing, which is as strong as the positive serial indirect effect through attribute transfer (Fig. 5, Panel A). Study 3 took place shortly after the FIFA World Cup, which is even more popular than the Wimbledon Championships. Therefore, a lower salience of event attributes does not offer a plausible explanation for the different results. Instead, it appears that consumers need longer exposures to identify the event schema of some event-typical ads and facilitate their attribute transfer—depending on the specific event-related elements featured. This explanation also is supported by the increased strength of attribute transfer between the 500-ms and 2-s exposure durations (Table 4).

Across exposure durations, the superiority of the event-typical ad over the atypical ad can be explained by the positive serial indirect effects through attribute transfer (\(a \times b \times c\) = 0.061, p < 0.001) and curiosity (\(a \times b \times c\) = 0.005, p = 0.04), whereas the serial indirect effect through the feeling of knowing is not significant (\(a \times b \times c\) = -0.001, p = 0.76), in support of H4b and H4c but not H4a. As mentioned, the results suggest that the event-typical ad used in Study 3 requires longer exposure to be categorized, which explains why respondents did not feel more certain about its identity, relative to the atypical ad, after 500 ms, according to the non-significant difference for the feeling of knowing between the two ads (p = 0.29). Furthermore, the serial indirect effect through curiosity is significant after 500 ms (\(a \times b \times c\) = 0.011, p = 0.02) but not after 2 s (\(a \times b \times c\) = -0.001, p = 0.79). As we address in the General Discussion, some atypical ads may require exposures longer than 2 s to reach peak curiosity, whereas curiosity triggered by event-typical ads stagnates or declines beyond very brief exposures if people easily resolve the incongruity. This explanation is in line with the observed pattern of curiosity levels across durations, which increases sharply for the atypical ad between 500 ms (M = 3.00) and 2 s (M = 4.48) but remains stable for the event-typical ad (500 ms: M = 4.29; 2 s: M = 4.39; see Table 2).

The results for the dual-typical ad confirm the contingency effect of exposure duration. When comparing the event-typical and dual-typical ads, the positive serial indirect effect through curiosity is significant after 500 ms (\(a \times b \times c\) = 0.007, p = 0.03) but not after 2 s (\(a \times b \times c\) = 0.005, p = 0.13). Similarly, for the dual-typical versus product-typical ads, the serial indirect effect through the feeling of knowing is positive and significant after 500 ms (\(a \times b \times c\) = 0.031, p < 0.001) but is not significant after 2 s (\(a \times b \times c\) = 0.005, p = 0.45); for the dual-typical versus atypical ads, the serial indirect effect through attribute transfer is not significant after 500 ms (\(a \times b \times c\) = 0.001, p = 0.91) but is positive and significant after 2 s (\(a \times b \times c\) = 0.023, p = 0.02).

The direction and significance of the indirect effects in Study 3 thus are very similar to those of Study 2, except for curiosity. Unlike Study 2, the indirect effect through curiosity drives the positive effects of the event-typical ad here, in support of H2b. To be curious, people need to be on the verge of resolving the incongruity (Loewenstein, 1994). The mediating effect of curiosity may have been significant in Study 3 but not in Study 2 because the stimuli we used facilitated incongruity resolution to varying extents. First, the Wimbledon Championship (Study 2) takes place every year in the same location, whereas the FIFA World Cup (Study 3) happens every four years in different locations, so the event schema of the former likely is more homogeneous than that of the latter. Ads matching the Wimbledon Championships schema thus should be easier to identify than those matching the FIFA World Cup schema, making it easier for respondents to resolve product schema incongruity. Second, one of the objectives of Study 3 was to enhance internal validity, so we manipulated event and product typicality by making only one small change to the ad visuals, which precluded the inclusions of iconic visual elements of the event. More respondents may have remained in an “almost knowing” state, because of the greater difficulty of resolving product schema incongruity when the ad evokes the event schema using non-iconic visual elements.

General discussion

Managers often choose to adapt their marketing activities to major sporting events, which actually runs counter to conventional marketing wisdom: The positive effects of consistency are well-established in marketing and branding literature, because consistency facilitates recognition, cultivates strong brand associations (Beverland et al., 2015; Keller, 2008) and triggers positive attitudes (Becker & Gisjenberg, 2023). The present research offers an initial evaluation of the effectiveness of ads that adapt to events versus those that stay consistent with product category schemas. It shows that marketers should adapt their marketing mix to events. Specifically, they can improve incentive-aligned brand choice (Study 3) and brand attitudes (all studies) by running event-typical ads, even with short exposures, as are prevalent in digital, print, and outdoor advertising (Elsen et al., 2016).

The use of event-typical ads not only defies the gold standard set by product-typical ads in short exposures conditions (Elsen et al., 2016; Pieters & Wedel, 2012) but also can outperform all the other ad types tested in our research. After 100 ms, event-typical ads perform as well as product-typical and dual-typical ads, and they surpass atypical ads. After longer exposures, event-typical ads outperform all three other ad types. This change occurs after exposures lasting between 500 ms and 2 s; the relative evaluations of the ads remain unchanged between 2 and 5 s. Our findings also challenge the psychological mechanisms that underlie advertising typicality effects, in that both curiosity (due to incongruency with the product category schema) and attribute transfer (due to being categorized with the event schema) are crucial, along with the feeling of knowing (Elsen et al., 2016). The evidence is very strong regarding attribute transfer. Comparing event-typical ads with product-typical ads provides support for the feeling of knowing and attribute transfer effects in Studies 2 and 3; curiosity receives support only in Study 3. In a comparison with atypical ads, attribute transfer is supported in both Studies 2 and 3, the feeling of knowing is supported in Study 2, and curiosity is supported in Study 3. In addition, exposure durations act as contingent factors, by altering the strength with which the feeling of knowing, curiosity, and attribute transfer drive the effects of event-typical and dual-typical ads. Notably, we consider alternative priming- and fluency-based accounts, but none of the empirical tests supported either explanation (see Web Appendix D).

Managers’ use of dual-typical ads also challenges brand consistency precepts. In a separate study, involving 24 experienced marketing managers and 24 advertising scholars, we learned that these experts believed dual-typical ads would perform best—providing reasons such that the link to the event could enhance ad impact—while still cautioning that the “product should not be overruled” (see Web Appendix B), reflecting their continued prioritization of consistency. Our results challenge these experts' intuition though. We show that dual-typical ads are never the optimal advertising strategy, because the ambiguity of their schema (Berlyne, 1960) can impede categorization mechanisms (Gregan-Paxton et al., 2005).

Theoretical contributions

First, we contribute to typicality literature by integrating and comparing three underlying mechanisms to explain the effects of event-typical ads on brand attitudes and incentive-aligned brand choice. Specifically, the transfer of event attributes predicts brand attitudes and incentive-aligned brand choice more strongly than the feeling of knowing what the ad is about (Elsen et al., 2016) or curiosity (Noseworthy et al., 2014). It is the main reason event-typical ads outperform product-typical and atypical ads.

Second, we extend the growing body of knowledge on ad exposure duration by showing that it is a contingency factor of the ad processing mechanisms in thin-slicing conditions. Curiosity for the atypical ad and attribute transfer for the event-typical ad strengthen when the exposure duration extends by as little as 1.5 s (i.e., from 100 or 500 ms to 2 s). Prior ad curiosity (Hutter & Hoffmann, 2014; Noseworthy et al., 2014) and attribute transfer (Mazodier et al., 2018) studies do not include exposure duration. Our findings clarify though that small increments in duration, even within thin-slicing parameters, can significantly alter consumer response mechanisms to advertising typicality.

Third, we expand the conceptual scope of ad identification frameworks after short exposures. Identifying an ad quickly with the product category (Elsen et al., 2016) is a particular case of a more general process of identifying the dominant ad theme, whether a product category or a sporting event. As we show, the indirect effect of product-typical versus atypical ads through the feeling of knowing (Elsen et al., 2016) generalizes to the comparison of event-typical versus atypical ads.

Fourth, we contribute to schema congruity theory, according to which an incongruent rather than congruent schema yields better evaluations by increasing curiosity (Mandler, 1982), by addressing this effect for the first time for stimuli that activate two schemas. Our findings about dual-typical ads are consistent with the schema congruity effect; this type of ad, which is congruent with both the event and the product category schema, is never optimal and is outperformed by the event-typical ad that is incongruent with the product schema. This insight is particularly compelling because it is counterintuitive. The experts we surveyed anticipated that dual-typical ads would be rated best; moreover, ads featuring two schemas are common in practice.

Fifth, our research suggests that schema incongruity is conducive to curiosity when consumers are in an “almost knowing” state (Kashdan & Silva 2009; Litman, 2005). In Study 3, the effects of the event-typical ad through curiosity are stronger than in Study 2. The homogeneity of Wimbledon’s cognitive schema and the greater evocative power of the event’s visual elements seemingly made it easier for consumers to understand the event-related theme in Study 2, which in turn prevented them from remaining in an “almost knowing” state.

Implications for managers

Embrace event typicality and creative flexibility

Prior research questions the efficacy of sporting event advertising and indicates considerable advertising elasticity heterogeneity across advertisements (Gijsenberg, 2014). Our results partially explain this heterogeneity and establish event typicality as a key driver of advertising efficacy. Specifically, event typicality is an effective adaptation strategy for designing ads that deliver results if brief or very brief viewing conditions are expected. Evoking major events through ad visual content can be a winning move, even when consumers have little time to view the ad. If this creative strategy is embraced though, product-typical elements should be avoided. Although it might be tempting to capitalize on both product category and event typicality, this strategy is not the most effective. Rather, advertisers should design ads that are clearly typical of the event to ensure better results across the thin-slicing spectrum.

Event-typical ads rely on the events’ cognitive schemas, which provide opportunities for originality. Events like the Olympic Games, FIFA World Cup, or Wimbledon Championships can be evoked with various elements, such as landmarks, symbols, or people (Mazodier et al., 2018). Event-typical ads thus allow for creativity and remain effective regardless of exposure duration. In contrast, atypical ads may pay off if audiences are likely to access thicker slices of information, which is unusual in many real-life situations in which ad exposure time is less than a couple of seconds, such as while driving by billboards, paging through magazines, or scrolling websites or mobile applications.

Enhance sponsorship leveraging efficacy

Our findings provide new insights for sponsorship leveraging, a sponsorship-linked marketing strategy that promotes the association between the sponsor and the event (Cornwell & Kwon, 2020). Sponsorship literature has paid limited attention to how the effectiveness of leveraging ads depends on their visual content; instead, it tends to consider whether leveraging should be performed at all, how much should be invested in it, and how prominent the event and the sponsor names should be (Carrillat et al., 2015; Olson & Thjømøe, 2009; Quester & Thompson, 2001). For example, Henderson et al. (2019) show that managers should adapt the color of their branding elements to match the color of the sponsored entity. Together with our research, these findings indicate that adapting marketing activities to a sponsored entity is key to increasing sponsorship returns.

In detail, event-typical ads are effective in leveraging major sporting events, even if they do not employ registered symbols exclusive to sponsors (e.g., Olympic logo). They can rely instead on generic event-themed representations available to all advertisers. Study 3 offers a strong illustration of this possibility, because the elements we used to evoke the event were less iconic. Therefore, our results are also relevant to managers interested in ambush marketing (i.e., using unrestricted event-related elements to create an implicit link to an event to improve brand equity; Kelly et al., 2012). On the one hand, ambushers can find unregistered typical symbols of an event to increase the effects of their campaign. On the other hand, event organizers and sponsors must identify typical symbols and register as many as they can, then help sponsors maximize their returns by making them the only brands able to use these highly representative attributes. For example, sponsors should work with organizing committees of the Olympic Games to register typical symbols of the host city and country, in addition to the event logo and tagline, to create leveraging ads that are more typical of the event than those that can be created by any other (non-sponsoring) advertisers.

Limitations and further research directions

Some limitations of our study could be addressed by further research. First, our results for dual-typical ads are consistent with categorization ambiguity (Berlyne, 1960; Gregan-Paxton et al., 2005), according to which being typical of more than one cognitive category impedes gist perception. An in-depth investigation of the reasons dual-typical ads might underperform relative to event-typical ads thus is warranted.

Second, the weaker-than-expected curiosity effects for the event-typical ad in Study 2 could imply the incongruity was fully resolved within 2 s. The homogeneity (i.e., fixed location) and yearly frequency of the Wimbledon Championships may have decreased the influence of curiosity, whereas it emerges for events that occur every four years in different locations, like the FIFA World Cup. Further research could test whether event-typical ads for events that change locations and happen infrequently trigger more curiosity than ads for more frequent events in the same location. Similarly, event frequency and location may influence the attribute transfer triggered by event-typical ads. We found that attribute transfer increases with exposure duration in Study 3, but not in Study 2, such that the more positive impact of event-typical ads compared with product-typical ads is statistically significant after very brief exposures (≤ 500 ms) in Study 2 but not in Study 3. The event schema related to the FIFA World Cup arguably is more heterogeneous, so the event-typical ad becomes harder to categorize, which inhibits attribute transfer after very brief exposures to some degree. Other event-typical ads also may require longer exposures to be categorized, so further research should examine which event-related elements facilitate the identification of the event schema to optimize attribute transfers.

Third, we use major sporting events as an alternative ad typicality schema, due to their massive following and large advertising investments, but even in thin-slicing conditions, other events could be relevant too, such as holiday ads. To test this assumption, we ran a post-hoc study in Australia. We asked 95 respondents (one was excluded due to failing twice the same attention check, leaving 94 respondents) whether eight real ads featuring elements typical of Christmas or Halloween were event-typical (i.e., mainly displaying elements that are representative of advertisements for an event). Confirming our assumption, more than 75% of respondents perceived these ads as event-typical (detailed results are available at https://researchbox.org/1426&PEER_REVIEW_passcode=MAWOVL). Furthermore, sporting events generate strong (usually positive) emotions (Decrop & Derbaix, 2010), which may facilitate attribute transfer (Fiske & Neuberg, 1990). Further research should test the effects of event typicality for less emotional events. Beyond product categories and events, other typicality schemas may exist too, whose relevance in thin-slicing conditions remains to be tested, such as ads typical of celebrities or consumer lifestyles (e.g., do viewers identify an ad as targeting a particular segment if it depicts people whose appearance is archetypical of this segment?). Event typicality also might boost the effectiveness of other communication strategies, like direct marketing (e.g., emails that are typical of events), owned social media (e.g., event-typical posts), and product placement (e.g., brand-plot-event connection).

Fourth, ad typicality research has not investigated the role of ad exposure frequency. The effects of ad typicality may weaken after multiple exposures, because it becomes easier to identify even an atypical ad. Nor do we examine how product typicality, event typicality, and exposure duration together affect recall. McAlister et al. (2012) propose (but do not test) that sponsors can improve identification when ads are thematically linked to the event. This proposition offers an avenue for further research.

Conclusion

Saturated communication environments and consumers’ reduced attention (Elsen et al., 2016) to ads create challenges for advertisers seeking to target consumers effectively. We show that event-typical ads can circumvent shorter exposure durations better than product-typical, dual-typical, or atypical ads, because they create a feeling of knowing, trigger curiosity, and benefit from the transfer of attributes from events. Event-typical ads not only allow marketers to leverage the global following of major international sporting events but also offer fertile ground for creativity. We hope this research spurs more investigations into the promising area of marketing mix adaptations to an event, especially in brief viewing conditions.