Abstract
Marketers commonly use ads that associate brands with success in persuasive communications. Yet, these ads may not be the most effective way to promote brands, particularly masculine brands. The current research examines when and why failure messaging can be an effective promotional approach. Across eight studies using both observational and experimental data from field and lab settings, the authors demonstrate that experiences of failure in achievement contexts, and ads that employ failure (vs. success) messaging, are more effective in promoting masculine but not feminine or neutral brands. An increase in consumers’ hostility mediates these effects. Feminine brands that employ aggressive branding cues (e.g., the color red) can also benefit from failure messaging. Additionally, the benefits of failure messaging are enhanced when people do (vs. do not) take responsibility for their failures, and this moderating effect is intensified (weakened) when failures are attributed to unstable (stable) causes. Finally, the consumption of masculine options was found to assist consumers in recovering from achievement failures.
The use of success messaging has long held a place in marketers’ promotional arsenal to influence consumers’ choices (Cavanaugh, Bettman, and Luce 2015; Garretson and Burton 2005). Success messaging evokes positive emotions such as pride and contentment that “engineer positive environments for consumers” (Griskevicius, Shiota, and Nowlis 2010, p. 238) and increase consumer engagement (Teixeira, Wedel, and Pieters 2012). In contrast, failure messaging—an ad that narrates, reminds consumers of, or targets consumers with failure experiences—evokes negative emotions that can adversely affect consumers’ attitudes toward brands (Burke and Edell 1989; Jain, Mathur, and Maheswaran 2009). As a result, failure messaging is not widely used as a marketing tactic. A notable exception involves the use of failure messaging in promoting products or behaviors that can prevent or resolve specific failures (Krishen and Bui 2015; Zheng 2020). For example, highlighting one's failure to lose weight as part of a healthy food campaign can motivate consumers to eat healthier options as a way to resolve failure. Yet, little research has explored whether failure messaging can influence preferences beyond instances where there is a remedial connection between the failure and the product portrayed in ads. For instance, can marketers promote a coffee brand by using an ad that primes consumers with common failures, even though coffee doesn’t resolve those failures?
The current research examines when and why failure messaging can be employed as a more generalized and effective promotional approach. In this multimethod research, we focus on failures within achievement contexts (e.g., jobs, education, sports) to explore how promotional content that primes personal and vicarious experiences of such failures can affect consumers’ preferences. Drawing on self-determination theory (Deci and Ryan 2000; Leander and Chartrand 2017), attribution theory (Weiner 2018), and the branding literature (Grohmann 2009), we propose and demonstrate that achievement failures can increase consumers’ desire for masculine brands—those that are perceived as dominant, brave, daring, aggressive, sturdy, and adventurous (Grohmann 2009). Additionally, since achievement failures produce hostility (Camacho-Morles et al. 2021; Çelik et al. 2013), and hostility leads to the pursuit of aggressive and dominant behaviors (Adachi and Willoughby 2013; Leander et al. 2019; Przybylski et al. 2014) often associated with masculinity (Eliot 2021; Weisbuch, Beal, and O’Neal 1999), we propose that the effect of failure messaging on preferences for masculine brands is mediated by consumers’ heightened levels of hostility (e.g., anger, irritability).
Extending this theorizing to feminine brands, we examine whether feminine brands that employ aggressive branding cues (e.g., the color red) in their advertising can also benefit from failure messaging. Further, because consumers’ responses to failures are primarily influenced by responsibility and stability attributions (Tsiros, Mittal, and Ross 2004; Weiner 2000), we suggest and demonstrate that the benefits of failure messaging are enhanced when people do (vs. do not) take responsibility for their failures, and this moderating effect is intensified (weakened) when failures are attributed to unstable (stable) causes. Finally, we extend our examination to consumers’ use of products by demonstrating that postfailure consumption of masculine products can help consumers recover from failure experiences.
Theoretical Development
Failure and Success Message Frames
Ads can directly or vicariously evoke psychological states, thoughts, and feelings (Vakratsas and Ambler 1999). According to classic advertising theory (Lutz, MacKenzie, and Belch 1983; Shimp 1981), advertising “works” by generating favorable attitudes, which then generalize to the featured product or brand. Advertisers often portray people experiencing success through the use of a product in pleasant surroundings to persuade consumers of the benefits derived from using the product or brand (Hung and Wyer 2009). Research indicates that consumers mentally simulate the events depicted in these ads, imagining themselves in the situation (Escalas 2007; Hung and Wyer 2009). Hence, success ads evoke vicarious experiences of positivity associated with achieving an outcome, and these associations are thus transferred to the featured product.
Negatively valenced ads are common in noncommercial contexts such as charitable giving (e.g., guilt-inducing appeals for donations) and public service advertising (e.g., fear-based appeals for reducing high-speed driving) (Zheng 2020). The use of such ads in commercial contexts has been limited and generally avoided in marketing campaigns, as negative emotions can adversely affect consumers’ attitudes toward brands (Burke and Edell 1989; Jain, Mathur, and Maheswaran 2009). Failure ads are a specific type of negatively valenced ad, where the negatively valenced failure can be related or unrelated to the featured product. When related, the advertised product simply resolves the salient failure depicted in the ad. For example, Weight Watchers ads illustrate how a failure to lose weight can be remedied by use of the focal product (e.g., the Weight Watchers program; Boone 2015). However, much less is understood about how failure ads that are unrelated to the featured product might influence consumer preferences. We extend prior literature by focusing on failure messaging's role in promoting brands and products that do not provide a direct resolution to failures depicted in ads; rather, they provide a specific symbolic value that becomes appealing as a result of consumers’ postfailure affective state. Since individuals vicariously experience emotional states depicted in advertisements (Escalas 2007; Hung and Wyer 2009), failure messaging can produce psychological effects that are nearly indistinguishable from actual experiences of success and failure. Thus, our theoretical framework considers failure messaging and direct experiences of failure as interchangeable, and they are expected to produce similar effects.
Failure and Aggressive/Dominance Seeking Behaviors
Unlike other types of failures (e.g., relationship failures), achievement failures inherently undermine competence 1 (Deci and Ryan 2000; Sheldon et al. 2001), which is the belief that one can produce desired effects in a surrounding environment (White 1959). Competence is an innate psychological need manifested through success and reduced through failures, such that when achievement failures occur, competence is simultaneously undermined (Shepherd and Cardon 2009; Teixeira, Silva, and Pieters 2012). Research further shows that low competence due to failures can induce hostility (Berkowitz 1989; Çelik et al. 2013; Przybylski et al. 2014), referring to feelings of anger, antipathy, and irritability (Allen and Anderson 2017; Anderson and Bushman 2002). For example, failures in academic settings (Taylor, Davis-Kean, and Malanchuk 2007) or when playing complex video games (Przybylski et al. 2014) often elicit hostility.
Hostility is rooted in neurobiology (Siever 2008), and its main downstream outcome is the pursuance of aggressive and dominant behaviors as an evolutionary means for enhancing chances of survival and success (Cabral, De Souza Tavares, and De Almeida 2016; Ekman and Cordaro 2011; Lindsay and Anderson 2000). Aggression and dominance are associated with qualities such as assertiveness (Buss and Perry 1992), competitiveness (Adam and Brett 2015), strength (Tibubos, Schnell, and Rohrmann 2013), and threat (Reed, DeScioli, and Pinker 2014). For example, aggressive facial expressions are interpreted as signals of threat and lead opponent negotiators to concede (Sinaceur and Tiedens 2006). These qualities are the primary signals of one's ability to produce desired effects in an environment (i.e., competence). Thus, while aggressive and dominant behaviors resulting from hostility do not provide a resolution to the failure, they can serve as compensatory means of satisfying an unmet need for competence. Extending this logic to consumers, those who experience achievement failures (personally or vicariously via ads) are likely to experience elevated hostility and engage in symbolic behavior that is aggressive and dominant to satisfy their unmet need for competence. Thus, brands and products that symbolize aggression and dominance may become appealing to consumers following failure experiences. Today, symbolic representations of aggression and dominance are found in masculine brands.
Masculinity as a Symbol of Aggression
Aggression is commonly associated with masculinity. For example, men are more likely than women to engage in physical aggression, and this difference is attributed to the effects of testosterone and brain development within gender-differentiating environments that valorize aggression in men (Eliot 2021). Men with high masculinity scores tend to be more aggressive than those with low masculinity scores (Weisbuch, Beal, and O’Neal 1999). This also applies to branding,
2
where brands with high aggression scores are likely to be considered more masculine. By definition, masculine brands are those featuring strong associations with masculine traits—adventurousness, aggressiveness, bravery, daring, dominance, and sturdiness—and weak associations with feminine traits—tenderness, fragility, grace, sensitiveness, and sweetness (Grohmann 2009). Masculine brands are often contrasted with those that are feminine and neutral/undifferentiated. Feminine brands are those featuring strong associations with feminine traits and weak associations with masculine traits, and neutral brands are those with weak associations with both masculine and feminine traits (Grohmann 2009). Accordingly, we suggest that when consumers experience achievement failures, their heightened hostility manifests in aggressive and dominance-seeking behaviors that increase preference for masculine brands (i.e., those that symbolize aggression, strength, and dominance). Thus, we predict:
Moderating Effects of Responsibility and Stability
When people fail in achieving desired outcomes, they tend to try to understand why they performed poorly. The resulting attributions stimulate multiple postfailure responses, including ascribing possible causes (Gendolla and Koller 2001). Researchers have proposed dimensions for classifying the causes and comparing causal explanations (Weiner 2000). Among different attributional dimensions (e.g., stability, responsibility, controllability, locus), responsibility and stability have been identified as the two most relevant and applicable dimensions for understanding consumers’ experiences (Tsiros, Mittal, and Ross 2004; Weiner 2000). Responsibility refers to the assignment of blame and differentiates between the causes that reside within and are controllable by a person (e.g., effort: “I failed in my exam because I didn’t study enough”) and the ones that reside in a person's environment and are not controllable by the person (e.g., task difficulty: “I failed in my exam because the exam was difficult”) (Tsiros, Mittal, and Ross 2004). Stability refers to one's anticipation of failure reoccurrence in the future and differentiates between the causes that are perceived as unstable (e.g., chance: “I failed in my exam because of bad luck”) and the ones that are perceived as stable (e.g., ability: “I failed in my exam because I am not intelligent”) (Weiner 2018).
Failures for which people take responsibility are particularly likely to challenge self-worth and induce psychological responses, whereas failures that people do not attribute to themselves are unlikely to influence self-related emotions (Weiner 2000, 2018). We expect responsibility to moderate the effect of failure on evoking hostility and preference for masculine brands. Specifically, when consumers experience failures (vs. successes) for which they take responsibility, they are more likely to feel incompetent and thus more likely to experience hostility and desire masculine brands. Therefore, we predict:
Further, we propose that achievement failure, responsibility, and stability interact (a three-way interaction) to influence preferences for masculine brands. Prior research shows that valence expectancy theory (Vroom 1964) can be applied to predict how attributional dimensions affect consumers’ responses to failures (Tsiros, Mittal, and Ross 2004). The valence expectancy theory (Vroom 1964) suggests that the motive for taking action is jointly determined by valence—the desirability of the action's outcome—and expectancy—the subjective probability that the action leads to the outcome (M = V × E). Analogously, Tsiros, Mittal, and Ross (2004) suggest that the evaluative factors of failure and responsibility interact to determine the valence of a target outcome. Separately, the probability factor of stability determines the expectancy component of the model. According to valence expectancy theory, and as demonstrated by Tsiros, Mittal, and Ross, stability/expectancy moderates the interactive effect of failure and responsibility on an individual's response to failure. In the same vein, we predict a three-way pattern of interaction among failure, responsibility, and stability.
Once valence (failure × responsibility) is formed, stability determines the probability associated with that valence (Tsiros, Mittal, and Ross 2004). Thus, the moderating effect of responsibility on the relationship between failure and preference for masculine brands further depends on stability. Consumers who attribute failures to unstable causes may maintain hope that the desired outcome can be achieved (Weiner 2018). In contrast, failures that are attributed to stable causes may lower self-esteem and promote passive responses such as withdrawal, a decline in pursuing actions that are directed at achieving the desired outcome, denial, escapism, or self-aggrandizement in other contexts (Abramson, Seligman, and Teasdale 1978; Hirschman 1992). For example, addicts who continuously fail to control their compulsive consumption and experience loss of control are more likely to die by suicide rather than engage in efforts to reestablish control (Hirschman 1992). Thus, stability can alter the moderating effect of responsibility on the relationship between failure and preference for masculine brands. Consumers who attribute failures to stable causes have a decreased expectation of regaining competence through any means; thus, they are less likely to experience approach-oriented emotions such as hostility and are less likely to pursue masculine brands relative to when they attribute failure to unstable causes. We thus predict that stability will moderate the interactive effect of achievement failure and responsibility on hostility and the desire for masculine brands (see the conceptual model in Figure 1). We propose:
Conceptual Model.
Overview of Studies
We conducted eight studies to test these hypotheses. Studies 1a–1c provide an initial demonstration of the effect of achievement failure (failure vs. success) on preference for masculine brands. Study 1a shows that vicarious experiences of failure elicit preferences for masculine liquor brands in college football fans. Study 1b shows that consumers’ experiences of failure alter their patterns of shopping on Amazon by increasing preferences for masculine brands. Study 1c shows that consumers are more likely to click on Facebook ads that incorporate failure (vs. success) messaging when the focal brand is masculine (but not feminine or neutral). Study 1d demonstrates that feminine brands that employ aggressive branding cues can also benefit from failure messaging. Studies 2a and 2b establish that personal and vicarious experiences of failure created via radio ads lead to hostility, which enhances preferences for masculine brands.
Study 3 tests the managerially relevant moderators of stability and responsibility in the context of testimonial ads. Finally, Study 4 demonstrates that masculine consumption can help individuals recover from achievement failures. See Table 1 for an overview of studies and results.
Overview of Studies and Findings.
Notes: Means with different superscripts are significantly different from each other at the p < .05 level. Comparisons underlying superscript differences are conducted within failure and success conditions in Study 1a, between corresponding failure and success conditions (cells in the same row) in Studies 1b, 1c,1d, 2a, 2b, and 3, and within failure and control conditions (all cells compared with masculine product) in Study 4.
Study 1a: Vicarious Experiences of Failure Alter Liquor Consumption in Sports Fans
In Study 1a, we tested whether experiences of failure (vs. success) lead to preferences for more masculine brands (H1) in a field setting, specifically in the context of sports. Research shows that sports fans vicariously experience failure and success based on the achievement failures of their favorite teams, and that their perceptions of their own competence and capability (in tasks unrelated to sports) decline as a result of their team's loss (Hirt et al. 1992; Sloan 1979). Drawing on this research, we examined the sales of liquor brands on each game day during the 2021 college football season in a geographic area characterized by a high degree of identification with the local university football team. We expected fans to experience failure and hostility as a result of the football team's losses and therefore show preferences for more masculine brands on the days that the football team lost relative to the days in which the team won.
Method
Liquor sales data
We obtained liquor sales data from a U.S. state agency for a geographic county in which the state's flagship university is located (in the focal state, liquor is only distributed through the state agency and sold via agency stores and stores that are in contract with the state agency) on 12 football game days (seven wins and five losses) played across the Fall 2021 college football season. The New York Times's National Collegiate Athletics Association college football fan map confirmed that citizens in this specific county identify heavily with the focal team (Giratikanon et al. 2014).
Brand gender associations
We pretested a set of 145 bestselling brands sold in the state-run and in-contract liquor stores during the football season in terms of consumers’ perceptions of their masculinity and femininity. Following Grohmann (2009), we used separate scales for measuring brands’ masculinity and femininity scores. Brand masculinity was measured by how adventurous, aggressive, daring, dominant, brave, and sturdy the brand was perceived (on seven-point scales anchored by “not at all” and “very much”). Brand femininity was measured as the extent to which the brand expresses tender feelings, is fragile, graceful, sensitive, sweet, and tender (on seven-point scales anchored by “not at all” and “very much”). These two scales yielded continuous masculinity and femininity scores for each brand.
Control variables
We used each brand's average unit price, unit volume, alcohol proof, opponents’ relative strength (calculated as the ratio of the opponents’ College Football News 2021 preseason rankings relative to the local university's ranking; College Football News 2021), and the day of the week the game was played as control variables.
Analysis
To test the prediction that vicarious experiences of failure as a result of a football team's loss (vs. win) drive fans’ preferences for more masculine (but not feminine) brands, we conducted a moderation analysis (PROCESS Model 1, Hayes 2018). We entered the game outcome (1 = loss, 0 = win) as the independent variable, unit sales of all brands as the dependent variable, brand masculinity scores as the moderator, average unit price ($), volume (mL), proof, the relative strength of the opponent, and the day of the week the game was played (0 = Thursday, 1 = Friday, 2 = Saturday) as control variables.
Results and Discussion
Consistent with our hypothesis, a two-way interaction between game outcome and brand masculinity emerged (B = 14.09, SE = 5.50, p = .011, 95% CI: [3.2994, 24.8729]). Probing the interaction using the Johnson–Neyman technique, we find that game loss (vs. win) had a significant negative impact on sales of less masculine brands (masculinity score ≤ 3.27, B = −14.54, SE = 7.42, p = .050) and a positive impact on the sales of highly masculine brands (masculinity score ≥ 5.83, B = 21.53, SE = 10.98, p = .050) (Figure 2; see Web Appendix B for full results). We repeated this analysis with the brand femininity score as the moderator, and the interaction between game outcome and brand femininity was not significant (p = .897).
Regression Coefficient of Performance Outcome Predicting Scales (Study 1a).
The results of this study show that consumers’ vicarious experience of failure (vs. success) affects their preferences for more masculine brands. On days that the football team lost and fans vicariously experienced failure, they purchased more masculine brands of alcohol. However, brand masculinity was not a significant predictor of sales on the days that the football team won, and brand femininity had no impact on sales on either occasion. Further, the finding holds after controlling for the unit price associated with each brand and other potential confounding variables. One limitation of the data in this study is that we were not able to separate the data based on the hour of sales (before and after games). However, the sales of liquor after games were included, as the state-licensed stores are generally open until midnight. These results offer preliminary, real-world evidence that failure in achievement contexts increases preferences for brands with masculine associations (H1). Given the correlational nature of these findings, we test the causality of these variables in subsequent studies.
Study 1b: Personal Failures Alter Patterns of Brand Preferences on Amazon
Study 1a provided evidence for the effect of vicarious experiences of failure on preferences for brands with masculine associations. In Study 1b, we replicated this effect by examining how personal experiences of failure and success influence consumers’ preferences for brands with masculine associations. To enhance realism, we used a noninterventional design and explored how naturally occurring experiences of failure and success in consumers’ daily lives altered their brand choices while shopping on Amazon. We compared the masculinity and femininity ratings of brands that consumers bought on Amazon before and after failure and success incidents to demonstrate that failures (but not successes) in achievement contexts lead to preferences for more masculine brands.
Method
Screener study
We set up a screener study to create a pool of frequent Amazon shoppers—those who self-identified as shopping three times or more per week on Amazon—who had recently experienced an incident of failure or success. U.S.-based participants from Prolific Academic (N = 5,459) participated in the screener study. They were asked how frequently they shopped on Amazon and then were randomly assigned to a condition in a between-subjects design (success vs. failure) in which they were asked whether they had experienced failure or success in an achievement context (which was defined for them) during the past two weeks for which they could accurately recall the date. Participants who met these criteria were invited to take part in the main study.
Main study
Two hundred forty-seven qualified participants from the screener pool took part in the main study. They provided a brief description of their failure or success experience and indicated to what extent they felt they had failed (on nine-point scales anchored by “not at all” and “very much”). Then, participants provided screenshots of their purchases on Amazon immediately before and after their failure or success. Finally, they responded to demographic questions. We removed 45 participants because the dates of order in their uploaded screenshots were chronologically incompatible with the stated dates of failure, or they had not provided screenshots of their Amazon purchases, leaving 202 participants (55% female, 45% male; Mage = 35.08 years).
Brand gender associations
Two independent coders calculated brand gender scores for each brand in the provided screenshots. Brand gender scores quantify the degree to which each brand was perceived as masculine or feminine based on brand design (name, logo, font, and color) and accompanying product (Lieven et al. 2015; Lieven 2018) (see Web Appendix C for full instructions). Raters assigned masculinity (femininity) scores to each brand at the brand design and product levels (on nine-point scales anchored by “not at all,” and “a great deal”). We used an intraclass correlation measure as a measure of agreement due to the continuous nature of the ratings (Hughes and Garrett 1990). Intercoder reliability was acceptable and ranged from .71–.84. The coders met to reconcile conflicting codes. The overall masculinity (femininity) rating for each brand was calculated as the average of reconciled brand design and product masculinity (femininity) ratings; αs of overall masculinity (femininity) for products purchased before and after the success or failure incident ranged between .76 and .88.
Results and Discussion
One-way analysis of variance (ANOVA) revealed that participants were significantly more likely to view themselves as failing in the failure (M = 7.41, SD = 1.80) versus success (M = 1.58, SD = 1.46; F(1, 201) = 647.83, p < .001, η2 = .763) conditions, confirming the success of the manipulation.
To explore how achievement failure (failure vs. success) and time of purchase (before vs. after the incident) predicted masculinity of purchased brands, we conducted a two-way mixed ANOVA, with the time of purchase as the within-subject factor and achievement failure as the between-subjects factor. See Figure 3 for the overall results.
Effects of Failure and Success on Preferences for Masculine Brands (Study 1b).
Pertaining to our hypothesis, there was a significant interaction between the time of purchase and achievement failure on masculinity of the purchased brand (F(1, 200) = 10.83, p = .001,
We repeated this analysis to explore whether achievement failure (success vs. failure) and time of purchase (before vs. after the incident) predicted the femininity score of purchased brands. There was no significant interaction between achievement failure and time of purchase (p = .260) on the femininity of the purchased brands.
Given the heterogeneity of consumers’ failures across a range of different achievement contexts (e.g., education, sports), this study illustrates the robustness of the effect of failures on preferences for masculine brands. The results provide support for H1 and demonstrate that failures (but not successes) increase preferences for brands that are more masculine, as indicated by objective coders. However, these incidents appear to have no significant impact on boosting preferences for options that are perceived as more feminine.
Study 1c: Failure Messaging Affects Responses to Facebook Ads
In Study 1c, we replicated the main effect (H1) found in Studies 1a and 1b by exploring how marketers can successfully employ failure messaging to promote masculine brands. In addition, we tested whether failure (vs. success) messaging is effective in promoting masculine brands alone or if it also enhances preference for feminine and neutral brands. In the interest of ecological and internal validity, we launched six Facebook ads with success or failure messaging for masculine, feminine, or neutral brands over a period of four days and examined their corresponding effects on click-through rate.
Method
Stimuli
Prior research suggests that slender and round fonts convey brand femininity, whereas heavier and angular fonts convey brand masculinity (Lieven et al. 2015; Shaikh, Chaparro, and Fox 2006). We identified a neutral fictitious brand name (“NIM”) and manipulated its overall gender association through employing masculine (Lemon/Milk typeface), feminine (Kristen ITC typeface), and neutral (Calibri typeface) font. These brand designs were paired with the same image of a hiking destination (the focal product). A pretest confirmed that the gender association manipulations were successful (see Web Appendix D for the brand designs and pretest results). Next, we paired the masculine, feminine, and neutral brand designs with either a failure or a success message. The failure message read, “Not performing well at your job? Need a release from the pressures?,” and the success message read, “Performing well at your job? Need a release from the pressures?,” resulting in six ads. A pretest confirmed that the ads did not differ in terms of being realistic, engaging, and true to life, and that the success and failure manipulations were successful (see Web Appendix D for ads and manipulation checks).
Participants, procedure, and design
To conduct the main study, we followed the guideline outlined by Orazi and Johnston (2020) on conducting field studies using Facebook's split test function. We ran the study on Facebook for four days and targeted U.S. male and female users aged 18 and older, with a total budget of $480. The ads were viewed a total of 54,297 times over four days by 28,413 male and 25,884 female Facebook users. The masculine brand/success messaging ad was displayed 7,672 times and received 93 clicks, and the masculine brand/failure messaging was displayed 7,454 times and received 358 clicks. The feminine brand/success messaging ad was displayed 9,095 times and received 55 clicks, and the feminine brand/failure messaging was displayed 11,015 times and received 81 clicks. The neutral brand/success messaging ad was displayed 9,887 times and received 77 clicks, and the neutral brand/failure messaging was displayed 9,174 times and received 67 clicks.
Results and Discussion
Following the guidelines provided by Orazi and Johnston (2020), we created a simulated data set to conduct logistic regression and examined the interactions between failure messaging and brand gender. Our simulated data set included three variables: coding for gender associations of brands (0 = masculine, 1 = feminine, 2 = neutral), coding for failure messaging (0 = success, 1 = failure), and coding for response to the ad (0 = did not click, 1 = clicked). Due to the dichotomous nature of the dependent variable (i.e., response to ad), we conducted a binary logistic regression (PROCESS Model 1, Hayes 2018; indicator coding for brand gender associations) with messaging as the independent variable, brand gender as the moderator, and response to ad as the dependent variable. See Figure 4 for the overall pattern of findings.
Probability of Clicking on Facebook Ad (Study 1c).
Consistent with our hypothesis, two-way interactions between failure messaging and brand gender associations emerged (failure and neutral vs. masculine interaction: B = −1.48, Z = −7.22, p < .001; failure and feminine vs. masculine interaction: B = −1.22, Z = −5.77, p < .001). When the brand presented was masculine, participants were significantly more likely to click on the failure (vs. success) messaging ad (B = 1.41, Z = 12.03, p < .001), whereas there were no significant differences between the click-through rates of failure vs. success messaging ads when the brand was neutral (B = −.06, Z = −.39, p = .699) or feminine (B = .20, Z = 1.12, p = .261). Finally, the results revealed a main effect of failure messaging such that the failure message was more likely to generate clicks than the success message (B = 1.41, Z = 12.03, p < .001). There was also a main effect of brand gender, such that the masculine brand generated more clicks relative to the feminine brand (B = −.70, Z = −4.11, p < .001) and relative to the neutral brand (B = −.45, Z = −2.89, p = .004).
Study 1c provided ecologically valid support for our prediction that failure (vs. success) messaging is more effective in promoting masculine brands (H1). Importantly, this effect did not emerge for neutral or feminine brands, reducing the plausibility that failure outcome messaging increases consumption in general (Mandel et al. 2017). Having demonstrated the effect of failure on preference for masculine brands across a range of contexts, participants, and study designs, we next focus on examining the mechanism underlying this effect.
Study 1d: The Advantage of Failure Messaging for Feminine Brands Depends on Aggressive Branding Cues
We theorized that achievement failures increase hostility and, thus, preferences for choices that symbolize aggression, primarily, masculine brands. In Study 1d, we aimed to provide additional support for our theorizing by exploring whether making feminine brands seem more aggressive would lead to an increase in consumers’ preferences for these brands when they experience failure (vs. success). Although aggression is conceptually linked to masculinity rather than femininity, brand masculinity and femininity are independent dimensions (Grohmann 2009); thus, feminine brands can be more or less aggressive. Leveraging research that has found an implicit association between the color red and perceptions of aggressiveness (Geng, Hong, and Zhou 2021), we explore how failure messaging combined with an implicit cue of aggression in brand design (i.e., those employing the color red) can boost consumers’ preferences for feminine brands.
Method
189 Prolific Academic participants (50% female, 50% male; Mage = 32.69 years) were randomly assigned to a condition in a 2 (failure messaging: failure, success) × 2 (logo color: red, blue) design. We developed testimonial ads for a fictitious craft shop named Hestia, which a pretest confirmed was perceived as a feminine brand name (N = 29; Mmasculinity = 4.15, SD = 1.39; mean compared with scale midpoint: t(28) = −3.29, p = .003; Mfemininity = 5.47, SD = .84; mean compared with scale midpoint: t(28) = 3.01, p = .005). In the ads, participants read a college student's testimonial about how wearing an infinity bracelet bought from Hestia Crafts reminds the student to learn from failures (vs. successes) and move forward (see Web Appendix E). In the ads, the brand logo was displayed once in red (hue: 0, saturation: 240, lightness: 120) and once in blue (hue: 160, saturation: 240, lightness: 120), as prior research shows that these specific colors evoke perceptions of aggression and agreeableness, respectively (Geng, Hong, and Zhou 2021). Participants expressed their interest in buying products from Hestia Crafts (“How interested are you in buying a product from this brand?”; nine-point scale anchored at “not interested” and “very interested”). They evaluated the ads in terms of engagement and mental simulation (see Web Appendix J for scales), and the extent to which the testimonial made them recall a failure in an achievement context that made them feel incompetent. Finally, they rated the brand in terms of aggression (“How would you describe this brand?”; nine-point scale anchored at “not aggressive” and “very aggressive”).
Results and Discussion
Manipulation checks
A series of ANOVAs revealed that the ads were equally engaging (p = .516) and realistic (p = .878) and created similar levels of mental simulation (p = .301). Participants in the failure condition were more likely to recall an incident of failure in an achievement context that made them feel incompetent (M = 3.40, SD = 2.50) than participants in the success condition (M = 2.09, SD = 1.74; F(1, 187) = 17.38, p < .001,
Preference for the feminine brand
We conducted a 2 (ad messaging: failure, success) × 2 (logo color: red, blue) ANOVA on preference for the feminine brand. The main effects of failure messaging (F(1, 185) = 3.69, p = .056) and logo color (F(1, 185) = 2.82, p = .095) were marginally significant. The interaction between the failure messaging and logo color was significant (F(1, 185) = 7.83, p = .006,
Preference for Feminine Brands (Study 1d).
The results of this study provide support for our theorizing regarding the effects of failure on preference for choices that symbolize aggression. Specifically, this study shows that feminine brands that are perceived as more aggressive are more likely to benefit from failure (vs. success) messaging. These results reduce the plausibility that the results of prior studies are driven by attributes other than aggression that are associated with masculine brands, as we replicated the findings in the context of feminine brands that were positioned as more aggressive. These results also align with prior work suggesting that brand perceptions are malleable (Lieven et al. 2015): by changing the coloring scheme associated with a brand, feminine as well as masculine brands can be positioned as aggressive.
Study 2a: Test of Underlying Mechanism with Personal Failure Experiences
We conducted Study 2a to examine the mechanism underlying the effect of achievement failure on preferences for masculine brands through an experiment in which the independent variable (achievement failure) is manipulated and the hypothesized mediator (hostility) is measured. Further, we examined potential alternative mediators of the hypothesized effect. Previous research shows that aversive experiences can lead to self-threat (Wang and Roedder John 2019), desire to restore control (Durante and Laran 2016), negative affect (Luce, Bettman, and Payne 1997), and comfort-seeking (Troisi and Gabriel 2011). Given the aversive nature of failures, we test these alternative explanations alongside hostility as the underlying mechanism.
Method
Participants, procedure, and design
238 U.S.-based participants from the Prolific Academic research panel (Mage = 42.02 years; 51% female, 49% male) were randomly assigned to one of the two conditions (achievement failure: failure, success). While assigning participants to the conditions, we used a randomized block design (Kirk 2012), as it ensures equal proportions of genders in each condition; thus, it eliminates any potential confounding effect of gender and allows us to focus on predicted effects. To create failure and success experiences, we used an anagram task that has been used in prior research (Chartrand et al. 2010; Leander and Chartrand 2017). In the success condition, participants completed 15 four-letter anagrams (e.g., lead → deal) one at a time. Participants were told that people typically solved each anagram in about 60 seconds, but they could take as long as they wanted to complete it. In the failure condition, participants completed 15 six-letter anagrams (e.g., antler → learnt) and were given only 15 seconds before the next word was presented.
Next, participants completed a separate evaluation task in which they were presented with an ad for a backpack with a fictitious masculine brand, Wilderness (Brough et al. 2016; see Web Appendix F), and were asked, “How likely are you to buy this brand?” on a nine-point scale anchored by “definitely would not buy it” and “definitely would buy it.” Then, participants were presented with a series of mediator measures in randomized order: hostility (six items: “To what extent do you feel this way in the moment?” “angry,” “hostile,” “irritable,” “scornful,” “disgusted,” and “loathing”; α = .97; Przybylski et al. 2014; Watson and Clark 1994), self-threat (three items; e.g., “I feel my identity is at risk”; α = .75; Wang and Roedder John 2019), comfort-seeking (three items; e.g., “This brand provides me with a sense of comfort”; α = .92), negative affect (ten items: “ashamed,” “jittery,” “distressed,” “afraid,” “upset,” “guilty,” “irritable,” “hostile,” “nervous,” and “scared”; α = .88; Luce, Bettman, and Payne 1997), and sense of control (three items; e.g., “Things are out of my control”; α = .91; Durante and Laran 2016), all on nine-point scales. In addition, participants completed questions measuring the effectiveness of the failure manipulation (perceived success: “To what extent were you successful in completing this activity”; perceived competence, using six items adapted from Ryan, Mims, and Koestner [1983]; e.g., “I think I am pretty good at solving anagrams”; α = .98) on nine-point scales anchored by “not at all” and “very much.” See Web Appendix J for all of the measures.
A confirmatory factor analysis showed that the mediator (hostility) and manipulation check (competence) measures represent distinct constructs. In the model, competence and hostility items loaded highly and significantly on their relevant factors as expected, and revealed good model fit (comparative fit index = .98; Tucker–Lewis index = .98; root mean square error of approximation = .07; χ2(53) = 123.34, p < .01). Further, the two-factor model fit significantly better than the one-factor model (χ2(54) = 2,113.83, p < .001; Δχ2(1) = 1,989.96, p < .001), demonstrating distinct constructs.
Results and Discussion
Manipulation check
Participants in the failure condition indicated significantly less success in completing the anagram task (M = 2.66, SD = 1.68) and felt less competent (M = 2.36, SD = 1.65) than participants in the success condition (experienced success: M = 7.02, SD = 1.83; F(1, 236) = 885.00, p < .001,
Preference for the masculine brand
An ANOVA examining the effect of the manipulation revealed a significant effect of achievement failure such that participants in the failure condition were more interested in the masculine brand (M = 6.38, SD = 1.87) than participants in the success condition (M = 5.52, SD = 2.02; F(1, 236) = 11.58, p < .001,
Mediation analysis
We used PROCESS Model 4 (Hayes 2018), examining whether hostility (M) mediates the effect of failure (X) on preference for a masculine brand (Y) and whether alternative mediators (self-threat, brand comfort, negative affect, and sense of control) also explain the effect. Results indicated that hostility mediated the effect of failure on preference for the masculine brand (B = .53, SE = .14, 95% CI: [.2740, .8354]), but none of the other mediators did (self-threat: B = .01, SE = .04, 95% CI: [−.0756, .0922]; brand comfort: B = −.23, SE = .12, 95% CI: [−.4812, .0053]; negative affect: B = .01, SE = .06, 95% CI: [−.1348, .1318]; sense of control: B = .004, SE = .02, 95% CI: [−.0346, .0493]; see Web Appendix F for full results). The results of this study provide support for H2 by demonstrating that consumers’ hostility mediates the effect of failure on preference for a masculine brand. Further, alternative explanations—self-threat, comfort-seeking, negative affect, and sense of control—were not significant mediators in the analysis, suggesting they do not explain the effects. Notably, hostility is a specific dimension of negative affect that refers to feelings of anger and irritability (Allen and Anderson 2017; Anderson and Bushman 2002). Hostility is uniquely associated with failure and masculinity, and thus masculine preferences, whereas the other five affective states that compose negative affect (fear, sadness, guilt, shyness, and fatigue) are not related conceptually or empirically to failure or masculinity.
Study 2b: Test of Underlying Mechanism with Failure Messaging
Study 2b provides an additional test of the proposed mediation process. The independent variable (achievement failure) is manipulated through testimonial radio ads, and the hypothesized mediator (hostility) and purchase intentions are measured. Prior research on narrative persuasion (Escalas 2007; Hung and Wyer 2009) shows that people mentally simulate scenarios presented in narrative format ads such that they internalize the goals and motivations of spokespeople and vicariously experience their emotional states. Thus, we expect testimonial ads to be particularly effective in eliciting mental simulations of failure and success experiences.
Method
Stimuli
A pretest confirmed that Wilderness was perceived as a masculine brand (Brough et al. 2016; see Web Appendix G). We hired a male voice actor to create testimonial radio ads for Wilderness coffee. In the ads, participants listened to an employee's recount of his performance in a work presentation (success, failure), followed by a description of the Wilderness coffee brand. Both ads were 40 seconds long, and the actor was instructed to keep his vocal tone as similar as possible while reading the ads. See Web Appendix G for the ad scripts.
Participants, procedure, and design
U.S.-based participants (N = 214) from the Prolific Academic research panel (Mage = 42.02 years; 51% female, 49% male) randomly listened to one of the two ads (failure, success). Next, participants were asked, “How likely are you to buy Wilderness coffee brand?” on a nine-point scale anchored by “definitely wouldn’t buy it” and “definitely would buy it.” Then, participants were presented with the hostility measure (six items: “While listening to the radio ad to what extent did you feel:” “angry,” “hostile,” “irritable,” “scornful,” “disgusted,” and “loathing”; α = .97; Przybylski et al. 2014; Watson and Clark 1994) on nine-point scales. Participants completed failure manipulation check measures (“How would you describe the speaker's performance at his presentation?” on a nine-point scale anchored by “much better than expected” and “much worse than expected”; adapted from Tsiros, Mittal, and Ross [2004]) and rated the ads in terms of their realism (“This ad is realistic”; measured on a nine-point scale anchored by “not at all” and “a great deal”) and mental simulation of the stories (see Web Appendix J for items).
Results and Discussion
Manipulation and confound check
Participants viewed the main character's performance as better than expected in the success (M = 3.14, SD = 1.90) versus failure (M = 5.64, SD = 2.75; F(1, 212) = 60.99, p < .001,
Preference for the masculine brand
An ANOVA revealed a significant effect of achievement failure such that participants in the failure condition were more interested in the masculine brand (M = 5.16, SD = 2.19) than participants in the success condition (M = 4.24, SD = 2.16; F(1, 212) = 9.72, p = .002,
Mediation analysis
We used PROCESS Model 4 (Hayes 2018) to examine whether hostility (M) mediates the effect of failure (X) on preference for a masculine brand (Y), which yielded a significant indirect effect (B = .45, SE = .17, 95% CI: [.1701, .8174]; see Web Appendix G for complete results). The results of Study 2b thus replicate the finding of Study 2a, such that hostility underlies the effect of failure (vs. success) on preference for masculine brands. Study 2b also illustrates the robustness of this effect by demonstrating it in the context of testimonial ads designed to evoke a vicarious experience of failure. The manipulation employed in this study provides an easily actionable template that marketers can use to leverage the effects of failure in promoting masculine brands: consumer testimonials, a common advertising format. Our results are consistent with literature showing that audiences mentally simulate story events (Escalas 2007; Hung and Wyer 2009), even when they focus on failure.
Study 3: When Failure Messaging Boost Preferences for Masculine Brands
We conducted Study 3 to test how achievement failure, failure cause stability, and responsibility for the failure interact to affect preference for masculine brands. We used testimonial ads to manipulate achievement failures, stability, and responsibility. These two attributional dimensions are practically and theoretically important, given that failure varies naturally on these dimensions. Further, ads that aim to influence consumers’ preferences through incorporating failure outcome messaging may not produce the desired effect if they attribute the outcome to causes for which consumers are not responsible (H3) and to stable causes (H4).
Method
Stimuli
We developed testimonial ads for a fictitious Tai Chi gym named Tiger Tai Chi, which a pretest confirmed was perceived as a masculine brand (N = 78; Mmasculinity = 6.05, SD = 1.54; mean compared with scale midpoint: t(77) = 6.00, p < .001; Mfemininity = 3.93, SD = 2.44; mean compared with scale midpoint t(77) = −3.86, p < .001). In the testimonial ads, participants read a college student's (“Casey Wilson”) recount of their performance in their professional program (success, failure), which they attribute to one of four causes (ability, luck, difficulty, effort). Casey then describes hearing about and trying Tiger Tai Chi gym and its transformational effect on their life. The manipulations of achievement failure, responsibility, and stability were adapted from Tsiros, Mittal, and Ross (2004), in which responsibility and stability of achievement failure (failure and success) were manipulated through attribution of the outcome to ability (personal responsibility, stable outcome), effort (personal responsibility, unstable outcome), task difficulty (no personal responsibility, stable outcome), and luck (no personal responsibility, unstable outcome). Pretests confirmed that the ads were similar in terms of being realistic, engaging, and true to life. Further, respondents reported similar levels of engagement in the story while reading different ads, confirming that they mentally simulated the story in the testimonial ads. See Web Appendix H for stimuli and pretests.
Participants, procedure, and design
Nine hundred sixty-five U.S.-based college students recruited from the Prolific Academic research panel (Mage = 24.65 years; 50% female, 50% male) were randomly assigned to view a testimonial ad in a 2 (achievement failure: failure, success) × 2 (responsibility: personally responsible, not personally responsible) × 2 (stability: stable, unstable) randomized block design.
Next, participants were asked about their interest in taking part in a class at the Tiger Tai Chi gym on nine-point scales anchored by “not at all” and “very much,” which served as the dependent measure. Participants completed manipulation check items adapted from Tsiros, Mittal, and Ross (2004), measuring achievement failure (“How would you describe Casey Wilson's performance at the physical therapy program?”; nine-point scale anchored by “much better than expected” and “much worse than expected”), outcome responsibility (“Casey Wilson was personally responsible for their recent level of performance”; nine-point scale anchored by “strongly disagree” and “strongly agree”), and stability (“The cause of Casey Wilson's performance is stable”; nine-point scale anchored by “strongly disagree” and “strongly agree”).
Results and Discussion
Manipulation checks
A series of one-way ANOVAs revealed that the manipulations of outcome, responsibility, and stability were successful. Participants viewed the main character's performance as better than expected in the success (M = 4.30, SD = 1.84) versus failure (M = 6.00, SD = 2.07; F(1, 963) = 180.18, p < .001,
Preference for the masculine brand
We conducted a 2 (achievement failure: failure, success) × 2 (responsibility: personally responsible, not personally responsible) × 2 (stability: stable, unstable) ANOVA on masculine brand preference. The expected three-way interaction between achievement failure, cause stability, and cause responsibility emerged (F(1, 957) = 7.32, p = .007,
Effect of Achievement Failure, Stability, and Responsibility Manipulations on Masculine Brand Preference (Study 3).
Given our hypotheses, we decomposed this interaction by treating stability as the higher-order moderator of the interaction between achievement failure and responsibility. We first considered results when the cause was unstable. The moderating effect of responsibility on the achievement failure-masculine brand preference relationship was significant (F(1, 478) = 33.51, p < .001), such that the effect of failure (vs. success) on preferences for masculine brands was significant when the consumer was responsible (i.e., effort: F(1, 478) = 53.38, p < .001) but nonsignificant when consumer was not responsible (i.e., luck: F(1, 478) = .82, p = .366).
We next considered results when the cause was stable. The moderating effect of responsibility on the achievement failure–preference relationship was marginally significant (F(1, 479) = 3.19, p = .075) such that the effect of failure (vs. success) on preferences for masculine brands was significant when the consumer was responsible (i.e., ability: F(1, 479) = 5.02, p = .025) but nonsignificant when the consumer was not responsible (i.e., task difficulty: F(1, 479) = .09, p = .758) (see Web Appendix H for the complete table of results). To provide evidence of robustness, we replicated the interaction with the continuous manipulation check measures (i.e., participants’ perceptions of the achievement failure, responsibility, and stability) using PROCESS Model 3 (Hayes 2018), and a three-way interaction emerged (B = −.01, t = −2.95, p = .003) confirming the ANOVA results (see Web Appendix H for additional information).
The result of this study provides support for H3 and H4 by demonstrating that the effects of achievement failure on consumers’ preference for a masculine brand depend on the attributional dimensions of stability (i.e., whether it is stable or unstable) and responsibility (i.e., whether one is or is not responsible for the failure). The pattern of results is consistent with the predicted three-way interaction: The effect of failure (vs. success) on preference for a masculine brand is bolstered when failures are attributed to causes for which people take responsibility, and this moderating effect is stronger when failures are attributed to unstable causes but weaker when attributed to stable causes.
Study 4: Masculine-Themed Consumption Can Restore Perceived Sense of Competence
The goal of Study 4 is to provide further evidence for our theorizing by exploring whether masculine-themed consumption can increase consumers’ perceptions of their competence, thereby enabling them to overcome their sense of failure. In addition, while prior studies manipulated brand gender association through brand design elements, this study manipulates brand gender by altering the gender association of the accompanying product. Prior research shows that gender associations of the product that accompanies a brand (alongside the brand design) determine perceptions of brand gender; brands that are paired with more masculine products are perceived as more masculine, whereas brands that are paired with more feminine products are perceived as more feminine (Lieven 2018). Following prior literature suggesting that darker tones and lower brightness, dim reflectiveness, and firmer textures create perceptions of masculinity (Van Tilburg et al. 2015), we altered gender associations of products by manipulating color and texture. We explored how consumers’ postfailure interactions with these products affected their perceptions of their own competence. We expected an increase in consumers’ perceived competence when they interacted with more masculine (vs. feminine and neutral) products following failure.
Method
Stimuli
We chose three stress balls (orange and soft, green and semifirm, and black and firm) to represent products with varying degrees of masculinity (see Web Appendix I). In a lab pretest, student respondents rated stress balls in terms of masculinity and femininity (on a nine-point scales anchored by “not at all masculine [feminine]” and “very masculine [feminine]”; Allison et al. 1980). The degree of masculinity and femininity was determined by comparing their masculinity and femininity scores with the midpoint of the scale. The orange and soft stress ball was perceived as feminine (N = 55; Mmasculinity = 4.29, SD = 1.84; t(54) = −2.85, p = .006; Mfemininity = 5.78, SD = 1.67; t(54) = 3.46, p = .001), the green and semifirm stress ball was perceived as neutral (N = 50; Mmasculinity = 4.86, SD = 1.62; t(49) = −.61, p = .543; Mfemininity = 4.60, SD = 1.74, t(49) = −.41, p = .684), and the black and firm stress ball was perceived as masculine (N = 52; Mmasculinity = 5.88, SD = 1.82; t(51) = 3.50, p = .001; Mfemininity = 4.60, SD = 1.74; t(51) = −1.67, p = .100). 4
Participants, design, and procedure
A total of 450 student volunteers were recruited from a research lab in a public business school in the south-central United States in exchange for a $5 Starbucks gift card. Twenty-three (5.1%) respondents failed to finish the survey, leaving 427 students (49% female, 51% male; Mage = 22.29 years). Participants were assigned to one of the conditions in a 2 (achievement failure: control, failure) × 4 (product gender: no product, feminine product [orange and soft ball], neutral product [green and semifirm ball], masculine product [black and firm ball]) design. To safeguard against social threats to internal validity due to mere interactions with products, data for the conditions that required interactions with products and the data for the condition that did not require product interaction were collected in separate and randomly assigned lab time slots.
First, participants responded to failure and control recall tasks. In the control condition, participants were asked to recall their past day: “In this part of the experiment, you are asked to write a paragraph (5–10 sentences) describing your life events over the past 24 hrs. Some of the things you might want to mention are your activities and people you met or spent time with.” In the achievement failure condition, participants were asked to recall a memory of an achievement failure: “In this part of the experiment, you are asked to write a paragraph (5–10 sentences) describing an event in which you failed in performing a task (e.g., job or educational contexts). Some of the things you might want to mention are what happened and how the event made you feel” (adapted from Horowitz et al. 2001). Next, participants who were not assigned to product consumption conditions were asked to sit quietly while the software recorded their responses to the earlier task (four minutes). Participants who were assigned to consumption conditions were told that they were taking part in a research study commissioned by the marketing department of a manufacturing company. These participants randomly received one of the three stress balls and were instructed to interact with the product, look at it, feel it, squeeze it, and slowly toss it from one hand to the other (Kimport and Robbins 2012) for four minutes. Then, participants expressed their sense of competence (“competent,” “confident,” “intelligent,” “capable,” and “skillful”) on nine-point scales anchored by “not at all” and “very much” (adapted from Cuddy, Fiske, and Glick [2008]; α = .93).
Results and Discussion
We conducted an ANOVA with the achievement failure manipulation and product gender as independent variables and sense of competence as the dependent variable. See Figure 7 for the overall results. The interaction between product gender and achievement failure on competence was significant (F(3, 419) = 24.26, p < .001,
The effect of gendered consumption on perceptions of competence.
Pairwise comparisons showed that, in the failure condition, participants who interacted with the masculine product reported a higher sense of competence compared with participants who interacted with the neutral product (p = .015), feminine product (p < .001), and no product (p < .001). Further, participants who interacted with the neutral product expressed a higher sense of competence compared with participants who interacted with the feminine product (p < .001) and no product (p < .001). However, participants who interacted with the feminine product did not report any boost in the sense of competence compared with participants who did not interact with any products (p = .719).
This study shows that consumers experience an increase in their perceived competence when they interact with masculine products (but not neutral or feminine products) following achievement failure. This study builds on our prior studies, which demonstrated that people prefer masculine-branded products following failure as a means to restore competence, to show that actual consumption of masculine products does effectively do what consumers seek. This outcome is beneficial from a customer satisfaction and consumer well-being perspective.
General Discussion
The success story has long held a place in marketers’ promotional arsenal as a way to highlight the efficacy of their products and build positive associations with brands. Given the host of favorable consequences created by the glow of success, it might seem counterintuitive to deliberately evoke thoughts of failure in advertisements. Across eight studies in several consumption domains conducted in the lab and in the field, we demonstrate how and when marketers can successfully employ failure in ads to promote brands. We show that personal or vicarious experiences of failure enacted via ads increase preferences for brands with masculine associations (Studies 1a–1c). In support of this theorizing, feminine brands that employ aggressive branding cues can also benefit from failure messaging (Study 1d). Further, we demonstrate that hostility underlies these effects and rule out alternative explanations (self-threat, brand comfort, negative affect, and sense of control) (Studies 2a and 2b). We identify moderating effects whereby attributing failure to internal and unstable causes boosts the effect of failure on preferences for brands with masculine associations (Study 3). Finally, we show that individuals who experience failure can restore lost competence through masculine-themed consumption (Study 4).
Theoretical Contributions
The current work makes several theoretical contributions, particularly in the areas of advertising and brand choice. Previous research has mainly focused on the favorable effects of success messaging and aspirational marketing on promoting brands (Lutz, MacKenzie, and Belch 1983; Shimp 1981) and warned against using ads that evoke negative feelings (Burke and Edell 1989; Jain, Mathur, and Maheswaran 2009). One exception to this is ads in which the featured products address a consumer need, adopting a problem–solution format. We move beyond this limited context to examine how failure ads can influence responses to products or brands that are not necessarily positioned as a solution to the failure depicted in the ad or experienced by the consumer. Specifically, we show that ads that employ failure messaging or vicarious experiences of failure can stimulate the pursuance of masculine brands as a means to resolve feelings of incompetence that result from failure, and this pursuit is not necessarily tied to the function of the featured product.
Furthermore, we extend research on the downstream effects of achievement failures on consumption (Gao, Wheeler, and Shiv 2009). Previous research has primarily demonstrated that such failures lead to preferences for brands and products that directly address the source of incompetence or closely symbolize competence (e.g., “Brain Fitness Software,” “Brain Fountain Pen”). We build on this work to show that the experiences of failure can boost preferences for brands that are associated with masculinity because they evoke hostility and desire for options that symbolize aggression. Further, we identify stability and responsibility as managerially relevant moderators of this effect by showing that the effect of failure on boosting preferences for masculine brands is stronger when people do (vs. do not) take responsibility for their failures. Additionally, this moderating effect is bolstered when failures are attributed to unstable causes but weaker when they are attributed to stable causes.
Considerable literature has investigated gender associations of brands and products and generally has identified consumers’ gender as the driving factor for these options (e.g., Brough et al. 2016; Lieven et al. 2015). We add to this literature by showing that consumers may pursue brands with masculinity associations for reasons beyond gender identity. Specifically, we demonstrate that consumers, regardless of their gender, show stronger preferences for brands with masculine associations when they experience failure in achievement contexts. In addition, we show that postfailure consumption of masculine options increases consumers’ perceptions of their own competence. Finally, while prior work on aggression in marketing examines aggressive behavior directly as an indicator of aggression (Kristofferson et al. 2017), our focus on the prebehavioral state of hostility is aligned with the broader emotion–action sequence that characterizes aggression (Anderson and Bushman 2002).
Managerial Implications
One of the major objectives of an advertisement is to promote brands, and marketers repeatedly rely on aspirational marketing to do so. A better understanding of failure (vs. success) messaging can have considerable value in extending knowledge of the persuasiveness of marketing appeals. First, this research provides managerial implications for helping consumers move on after a failure through tailored advertisements and consumption, creating value for consumers who experience achievement failure. Second, we show that, contrary to common practices, failure messaging is likely to be more effective than success messaging in promoting masculine brands. Specifically, we demonstrate how marketing managers can use testimonial ads and effectively incorporate attributional dimensions of failure to boost preferences for masculine brands. Third, we show that the most effective forms of failure messaging feature failures that are attributed to unstable (i.e., temporary) causes that reside within consumers, which can guide testimonial ad design. Finally, based on our theorizing, the effect of failure messaging on consumers’ brand interest is contingent on the associations between the brand and aggression; thus, feminine brands that cue aggression through strategies such as incorporating aggressive design elements (e.g., the color red; Geng, Hong, and Zhou 2021) can successfully employ failure messaging to promote the brands.
Consumer Well-Being and Public Policy Implications
While people strive to perform well in achievement contexts, the actual outcomes of their efforts are far from perfect; salespeople may fail to meet sales quotas, students may fail to gain admission to academic programs, athletes may fail to win a match, doctors may fail to save patients’ lives, and academics may fail to publish their manuscripts. Prolonged cases of achievement failure can lead to adverse symptoms such as chronic stress, depression, and grief, and in extreme cases, cause people to consider self-harm and suicide (Ucbasaran et al. 2013). Thus, addressing the feelings of incompetence resulting from failures is critical to maintaining consumer well-being. In an ideal world, consumers would seek direct resolutions to address the source of their failures; however, such resolutions are not always achievable or immediately possible. This research indicates that consumers can manage the psychological discomfort resulting from failure through consuming masculine brands.
Future Research
The current research provides new lenses for understanding the effect of failures on consumers’ preferences, and by connecting masculine brands to achievement failure, this work offers multiple avenues for future research. While the focus of our work is on failure in achievement contexts, and our pilot study showed that such failure is more likely to affect one’s sense of competence than failure in a relationship context, we contend that relationship failures may also affect preferences for masculine brands. For example, research shows that anxious attachment and distress can elicit hostility after a relationship breakup (Dutton and Winstead 2006). Future research could explore conditions under which relationship failures lead to a preference for masculine brands. Future research can also investigate whether different types of failure, ad format, and the nature of the appeal may alter the present findings.
Further, we demonstrated that failure promotes the symbolic consumption of masculine-themed options. One future area of research is understanding the effects of individual traits on promoting other forms of compensatory consumption, such as escapism and direct resolution in response to failure (Mandel et al. 2017), in contrast to preferences for masculine brands. For example, people high in self-denial tend to use selective attention as a means to deal with problems (Fleishman 1984). Thus, consumers’ self-denial may moderate the strategies that they use to cope with failure, as consumers high in self-denial may adopt escapism while dealing with failure. Relatedly, previous research suggests that entity theorists use brands to signal their traits and engage in self-enhancement, whereas incremental theorists believe one's qualities can improve only through more effortful learning and actual development of their abilities (Lee and Bolton 2020). Future research can examine how entity (vs. incremental) theorists respond to failure messaging. In summary, our research takes an important step toward understanding the effects of failure messaging on shaping consumer preferences for masculine-themed products and brands. We further explain why these effects manifest and identify key managerially relevant moderators that bolster these effects. We hope that future research will explore these insights further to fully understand the ramifications of failure on consumer decision making and behavior.
Supplemental Material
sj-pdf-1-mrj-10.1177_00222437231181078 - Supplemental material for Success Is Not Final; Failure Is Not Fatal: How Failure Versus Success Messaging Leads to Preference for Masculine Brands
Supplemental material, sj-pdf-1-mrj-10.1177_00222437231181078 for Success Is Not Final; Failure Is Not Fatal: How Failure Versus Success Messaging Leads to Preference for Masculine Brands by Niusha Jones, Blair Kidwell and Anne Hamby in Journal of Marketing Research
Footnotes
Acknowledgments
The authors thank the JMR review team for their helpful comments on earlier versions of this manuscript, and Kirk Smith and Idaho State Liquor Division for their support of this research.
Coeditor
Vikas Mittal
Associate Editor
Karen Page Winterich
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
