I am ashamed of my brand-self! Consumer-brand identi ﬁ cation as a moderator of emotional reactions following symbol-laden brand failures

Purpose – The study of shame has a long tradition in intra-and inter-personal psychology. This paper aims to investigate whether consumers can experience brand shame after self-relevant consumption incidents. Speci ﬁ cally, this research proposes that consumers follow a complex shame-inducing process in the aftermath of unpleasant experiences involving their favorite brand. The moderating role of relational tie strength between consumers and their favorite brand existing prior to symbolic failures is examined. Design/methodology/approach – A scenario-based, online survey ( n = 660) among consumers who have recently experienced a self-relevant failure with their favorite brand was conducted. Con ﬁ rmatory factor analysis ensured the reliability and validity of the measurement model. For testing the conceptual model, data was analyzed by means of a moderated mediation analysis. The proposed model was tested against, among others, common method bias and alternative models. The ﬁ ndings were cross-validated with a scenario-based online experiment ( n = 1,616). Findings – Results show that brand shame is a key mediator between customer dissatisfaction and brand anger when self-relevant, symbolic failures happen. Moreover, strong consumer-brand identi ﬁ cation triggers brand-detrimental effects. It is shown to in ﬂ uence the connection between consumers ’ inward-(i.e. brand shame) and resulting outward-directed (i.e. brand anger) negative emotions on brands, which lead to consumer vengeance. Originality/value – To the best of the authors ’ knowledge, this research is the ﬁ rst to introduce the concept of situational brand shame to the literature on favorite brands. Furthermore, it shows that consumer-brand identi ﬁ cation moderates the direct and indirect (via brand shame) unfavorable effects of failure-induced dissatisfaction on brand anger. This research adds insights to the investigation of the “ love-becomes-hate ” effect arising after self-relevant failures involving consumers ’ most preferred brand


Introduction
The tenet of relationship marketing is fostering close ties between consumers and brands.Extant research shows that consumers who are closely connected to a brand react less detrimentally to negative word-of-mouth (NWOM) (Lisjak et al., 2012).Such a forgiving reaction stems from a strong overlap between brand-connected consumers' identity and their favorite brand (Belk, 1988).This makes brand-connected consumers regard a risk to the brand as a risk to their self.To maintain a positive self-view and to reduce dissonance, such The current issue and full text archive of this journal is available on Emerald Insight at: https://www.emerald.com/insight/1061-0421.htm consumers are inclined to ignore others' brand criticism (Cheng et al., 2012).
However, research on service failures contrasts this "forgiveness hypothesis" by showing that strong-relationship consumers are those who respond most unfavorably to negative brand actions (Wolter et al., 2019).Specifically, they show a "love-becomes-hate" effect, which means that brandconnected consumers feel especially disappointed and harmed after a brand failure (i.e. a situation when the performance of a brand's service/product falls below a customer's expectations).Such an event leads them to believe that their favorite brand did not fulfill its part of the relational contract.This stimulates feelings of betrayal and retaliation (Gr egoire and Fisher, 2008;Zhang et al., 2021).
An important difference between this finding and the "forgiveness hypothesis" is the variation in the self-relevance of the two types of transgressions: when a brand failure directly and personally affects the bonded customer, such an event becomes highly self-relevant.Such a failure regularly interferes with consumers' personal goals and thus elicits a more adverse reactionthat is, increased angeras compared to those individuals less affected by the brand's transgression (e.g.observers of negative publicity).Especially for consumers with strong relational bonds to their favorite brand, such an incident bears a strong symbolic meaning as it emits that they have relied on the wrong brand to construct and communicate their consumer self (Tangney and Fischer, 1995).
In this research, we add a new aspect to the explanation of the "love-becomes-hate" effect by investigating consumer reactions to symbol-laden brand failures of their most favorite brand.In more detail, we define symbol-laden brand failures as situations of high personal relevance when a brand falls below consumers' expectation regarding psychological, identity-related benefits (Ringberg et al., 2007) and a favorite brand as a specific brand that consumers have preferred over others prior to a critical service incident.Extant literature sheds some light on various aspects of consumers' interactions with their favorite brand, such as consumer co-creation (Hajli et al., 2017).However, research remains rather silent about consumers' reactions to self-relevant brand actions, which interfere with their personal goals and negatively affect consumers' sense of self.A notable exception is the work of Davis and Dacin (2022) on consumers' motivation to maintain brand attachment who refer to "imagebased transgressions" (i.e. a situation when a brand represents an undesired identity for consumers, leading to incongruity in their identification with the brand) as a comparable concept.Consequently, we investigate the following research question: RQ. How can one explain consumers' intensified negative brand reactions after self-relevant failures of their favorite brand and what is the role of relational bonds as a boundary condition?
To answer this question, this research introduces a novel conceptual model (see Figure 1) based on literature that includes dissatisfaction, brand shame, brand anger and consumer-brand identification (CBI), which is the "consumer's perceived state of oneness with a brand" (Stokburger-Sauer et al., 2012, p. 407).It assumes that following a self-relevant failure involving their favorite brand, consumers typically experience dissatisfaction, which is related to brand anger.This research elaborates on this established relationship by investigating the mediating role of brand shame and by examining the moderating role of CBI.
Brand shame is a self-conscious social emotion of selfcondemnation which typically results when a person interprets a dissatisfactory trigger event as personally important but incongruent with his/her identity goals (Swee et al., 2021).It is assumed that consumers feel ashamed after learning that their most favorite brand has exploited the consumer-brand relationship and has personally let them down by providing an inferior treatment, while the opposite is expected (Tan et al., 2021).In their urge to overcome this unpleasant feeling and to minimize dissonance, consumers try to protect their selves at the cost of their favorite brand by developing brand-directed anger.Shame's mediating role on the dissatisfaction !anger link allows the decomposition of the dissatisfaction effects into both an inward-directed (i.e.brand shame) and a subsequent outwarddirected negative emotion (i.e.brand anger).
It is further proposed that this negative reaction even intensifies when CBI and consequently the perceived selfrelevance of the failure increase.Before the failure, for Figure 1 Conceptual model individuals with high CBI, the favorite brand has heightened symbolic meaning and is regarded as a trusted partner to construct their self-identity.However, to them a brand failure is not only a functional performance problem, but more importantly, a symbol that the brand is unable to satisfy identity-based self-motives (Hedrick et al., 2007).Basically, it is an identity-related failure that makes bonded consumers question their favorite brand's dedication to their relationship and foster their perception that they have relied too long on an untrustworthy partner who now opportunistically exploits consumers' waiver to use competitive brands.In contrast, individuals with low CBI are theorized to regard failures as more self-neutral (Wolter et al., 2019).There is some evidence that consumers in a strong relationship have higher expectations and become more annoyed when experiencing self-relevant failures or other forms of inferior brand actions than less brand-connected consumers (Gr egoire and Fisher, 2008;Wolter et al., 2019).This suggests that they can have stronger negative emotions and more intense spillover effects following a brand failure symbolizing their favorite brand's lack of effort to continue the relationship and the nonexistence of shared values.
This research provides both theoretical and empirical evidence (by means of two studies) that consumers can experience shame after self-relevant failures which consumers face with their favorite brand.In certain situations, brand shame potentially harms brands by eliciting anger, which has been earlier shown to trigger consumer retaliation (Septianto et al., 2020).This research contributes to the ongoing discussion on the explanation of the dissatisfaction !anger link (Yang et al., 2022) and the existing confusion concerning the role of relational bonds (Jabeen et al., 2022).The framework warns marketers that seemingly small but selfrelevant failures can have detrimental brand effects, but also suggest some countermeasures.The following section conceptualizes the model's constructs and elaborates on their relationships.

Conceptual background and hypotheses
2.1 Brand-failure-induced negative emotions 2.1.1Customer dissatisfaction In marketing literature, definitions of dissatisfaction are typically influenced by two schools of thought: the first comes from the expectancy-disconfirmation model, claiming that (dis)satisfaction results from a comparison of consumers' expectations and post-purchase outcomes (Oliver, 1977).If the outcome is judged to be worse than expected, "negative disconfirmation" arises.Researchers belonging to the second school (Shaver et al., 1987) claim that dissatisfaction is a relatively undifferentiated distress emotion caused by a disadvantageous event (Tan et al., 2021).Synthesizing both perspectives, this research conceptualizes customer dissatisfaction as the transient negative feeling that customers experience after a failure involving their favorite brand when their expectations remain unmet.

Brand anger
Anger is a strong outward-directed negative emotion, most frequently experienced after frustrating events, such as brand failures (Zhang et al., 2021).Central elements of anger are customers' belief that they have been wronged unjustifiably, that others are responsible for the failure, and that it is focused on specific targets that undermined consumption goals (Wen-Hai et al., 2019).In contrast to dissatisfaction, anger is a more specific emotion that can lead to intimidating, attacking or even eliminating the source of harm (Yi and Baumgartner, 2004).This research refers to brand anger as consumers' negative emotion of hostility targeting their favorite brand as the inferred source of the experienced injustice.
Extant literature suggests that individuals react with a primary, negative emotion to unpleasant events, which can be labeled as dissatisfaction.As individuals try to make causal inferences about negative events, cognitions of appraisal enter the emotion process and elicit more specific discrete emotions, such as anger.Anger is likely to arise after self-relevant brand failures, as consumers automatically make external attributions to protect their self and blame the involved brand (Michael et al., 2007).The literature strongly supports dissatisfaction as a prerequisite for anger (Chen et al., 2021).Therefore: H1.Following a self-relevant failure of consumers' favorite brand, the more dissatisfied consumers are, the greater their brand anger.

Brand shame
Conceptualizing brand shame.Another negative emotion that arises when a person interprets an event as self-relevantbut incongruent with his/her identity goalsis shame (Tracy and Robins, 2004).Shame is a negative self-conscious social emotion that includes self-condemnation and arises when a person perceives himself/herself as having failed one's standards, goals or moral values (Sabini et al., 2001).These negative judgments are characterized by an evaluation of the self as inherently flawed or inadequate (Carden et al., 2018).
Ashamed people tend to focus on the deficiency of their entire self and regard themselves as responsible for triggering negative events (Tracy and Robins, 2004).According to Niedenthal et al. (1994), shame focuses on a perceived transgression of one's self (e.g."I am a bad consumer having preferred this betraying brand!").Consequently, shame is very harmful for self-esteem; that is, it is a highly negative social emotion leading to intense coping strategies.Several other researchers studying the nature and consequences of shame in various fields share this view (as detailed in Supplementary Appendix 1).It illustrates that shame can arise in a wide range of contexts and suggests that shame can trigger a collapse of interpersonal relationships.
This research argues that in the case of self-relevant failures involving a favorite brand, people "typically experience a mixture of emotions in response to (. ..) events" (Tangney et al., 1996(Tangney et al., , p. 1263)).In this specific situation, shame is one of these emotions.Some failures have been shown to trigger consumers' shame (Sugathan et al., 2017).Here we maintain that brand shame arises when consumers feel that their consumer-self is inferior because of being linked to aformerly highly self-relevant but nowfaulty brand.Brand shame stems from consumers' belief that they were earlier misguided and have constantly chosen the wrong brand, which has finally turned out to be incongruent with their self-concept and values.Thus, brand shame develops when consumers experience a personal selection failure and regard their selves as responsible for this flaw (Tangney and Fischer, 1995).Shame is likely to occur after a perceived failure involving the (earlier) preferred brand.Such a failure has a strong personal relevance to the consumer as it is a symbol-laden incident that signals him/her that the brand seems to have no interest to continue its relationship to its most valuable consumers and does not share the same values anymore.Consequently, the consumer feels betrayed and is no longer able to satisfy his/her psychological need for self-enhancement.This can elicit harsh self-criticism for relying too long on the wrong favorite brand and inwarddirected aggressiveness as a reaction to this self-inflicted, acute problem (Sugathan et al., 2017).
Clarifying the mediating role of brand shame.Negative emotions caused by dissatisfaction block consumers' goals or demand accomplishments.In this research, it is assumed that brand shameat least partiallymediates the unfavorable effect of customer dissatisfaction on brand anger.Disappointed consumers have been shown to sometimes experience shame (Sun, 2018).Brand shame might arise as a specific negative emotion after consumers who have preferred this brand over others prior to the failure think that they must bear the responsibility of having been a fan of a faulty brand.This corresponds also to the "shame-on-me" effect (Platt and Freyd, 2015), suggesting that betrayed customers feel a threat to their self-worth.In turn, brand fans question their "consumer-self" because of constantly favoring a brand that ultimately betrays them despite their loyalty.In such a situation, consumers face a highly uncomfortable identity threat and a harm to their selfconcept.They must overcome this dissonance by applying forms of self-affirmation to restore the disordered self-concept and engage in patterns for achieving this goal (Cheng et al., 2012).
In the case of a self-relevant, symbolic failure, blaming others for a failure is one likely strategy to overcome the imbalance of the self (Tangney and Dearing, 2003).Consequently, one can argue that consumers who experience brand shame are likely to make the involved brand responsible for their inability to achieve their personal goals.Even when these individuals know that their initial attribution of blame is irrational and unjust, they are likely to develop the others-directed negative emotion (Heaven et al., 2010) of brand anger, which elicits aggressive behaviors (e.g.bad-mouthing the source of anger).Earlier research has shown that self-conscious emotions, causing frustration of self-relevant goals, are likely to develop toward the perceived end of a self-relevant brand relationship or after its disentanglement (Mills et al., 2007).Shame has been shown to have an intensifying effect on anti-brand emotions (e.g.hating the brand) (see Supplementary Appendix 1).Hence, it follows: H2.Following a self-relevant failure of consumers' favorite brand, the more dissatisfied consumers are, the greater their brand shame.
H3.Following a self-relevant failure of consumers' favorite brand, brand shame mediates the impact of dissatisfaction on brand anger such that the greater consumers' brand shame, the greater their brand anger.

Consumer-brand identification theory
Literature has typically approached consumers' identification with a brand from two perspectives.According to the sociological viewpoint, brands should be best understood as a vehicle for communicating a consumer's identity (Fetscherin et al., 2016).Sociology-based definitions regularly characterize CBI as the perception of sameness between a certain brandas it conveys a symbolic meaningand the consumer (Tuškej et al., 2013).
The second school is the social psychological perspective.It relies on social identity theory to conceptualize CBI.According to this theory, social identitywhich is an essential part of a person's self-conceptis based on the social entities to which a person belongs (Tajfel and Turner, 1985).The theory claims that individuals define themselves and other persons by a categorization into social groups.This idea was later transferred to the marketing context by suggesting that consumers identify with attractive companies because they satisfyto some extenttheir permanent need for selfdefinition (Bhattacharya and Sen, 2003).More recently, research has provided evidence that consumers regularly (cognitively) identify with brands (Stokburger-Sauer et al., 2012;Wolter et al., 2019).Brands convey a certain symbolic meaning, which typically manifests itself in consumers' perceptions of strong value congruity (Tuškej et al., 2013).The resulting self-categorization gives rise to a psychological ownership of the brand because a consumer's identity now extends to the brand as part of his/her social identity (Bhattacharya and Sen, 2003).As brands have a strong symbolic meaning, consumers can use them to create and maintain a positive self-concept, i.e. achieving self-relevant goals (Johnson et al., 2011).This is also emphasized in psychology-based definitions of CBI, which regularly describe it as the extent to which the brand expresses and enhances a consumer's identity.Carlson et al. (2008) refer to the personal identification with a brand as the degree to which an overlap exists between a person's self-schema and the schema he/she holds for the brand.This research proposes that CBI particularly covers this overlap and is a cognitive indicator for self-relevant relationships, which implicate some important aspects of a consumer's self-concept or self-image (Büyükda g and Kitapci, 2021; Johnson et al., 2011).
In cases where a consumer has such a personal intertwining with a specific brand, events related to the brand have a large significance (Lisjak et al., 2012), especially when they are negative.A self-relevant failure involving the consumer's most important brand is a major disorder of a customer-brand relationship, triggering a dissolution of the mental oneness.Losing a relationship that embodies a great deal of self-concept is a grave threat to the consumer's identity.Negative service incidents terminate consumers' ability to satisfy their identitybased motives of self-uniqueness and self-enhancement by means of a brand (Hedrick et al., 2007).They can harm the self, particularly when they are interpreted as a person's own failure in making faulty connections.This ultimately leads to strong psychological pain (Cheng et al., 2012).Service problems can cause strong feelings of emotional distress (Wolter et al., 2019) and can lead to maladaptive behaviors such as revenge, which may trigger brand hate (Bayarassou et al., 2021).

Role of consumer-brand identification for negative emotions
This article's key proposition is that CBI moderates the considered direct and indirect effect of dissatisfaction on brand anger, such that an increase in identification leads to more detrimental emotional reactions of consumers.This claim is supported by earlier interdisciplinary research (see Supplementary Appendix 1), suggesting that shame regularly elicits unfavorable consequences for close relationship partners.For consumers high in CBI, brands are such an intimate partner.It is assumed that this escalating effect becomes evident with three complementary mechanisms: (i) an enhancement of the direct unfavorable effect of dissatisfaction on brand anger; (ii) an increase of the unfavorable effect of dissatisfaction on brand shame; and (iii) an amplification of the unfavorable impact of brand shame on brand anger (see Figure 1).

2.3.1
The moderating role of consumer-brand identification on the dissatisfaction fi brand anger effect Social exchange theory claims that, in any exchange, both parties face costs and rewards.While costs refer to the negative consequences (e.g.loss of money), rewards are the positive consequences (e.g.receiving adequate services).The exchange process is governed by certain norms, such as the norm of reciprocity, which states that individuals ought to react similarly to the way they have been treated (Cropanzano and Mitchell, 2005).For instance, when consumers experience negative actions from brands, for instance, broken promises, they tend to show negative affective and cognitive reactions in return.After a self-relevant negative incident, individuals high in CBI are supposed to be angrier than brand-unconnected consumers.Not only is naming the outward-directed emotion of anger as a key determinant of postfailure destructive habits intuitive, but it is also consistent with extensive academic research (Gr egoire and Fisher, 2008).
For individuals high in CBI, events that happen with a brand have a larger significance (Lisjak et al., 2012).CBI creates a form of situational involvement that exacerbates the impact of a poor, dissatisfactory brand performance on consumers.Individuals with a strong brand connection have the high expectation of being treated well by the brand (Bambauer-Sachse and Rabeson, 2015).For them, the occurrence of a failure causes a higher mismatch between the brand's expected performance and the perception of the actual performance.Individuals high in CBI experience a higher imbalance in their exchange relationship given their higher expectations (i.e.being better served than the average customer) (Wolter et al., 2019).
CBI biases individuals to perceive that the incurred costs (i.e.failure) are larger than they objectively are (i.e.resulting in a higher negative disconfirmation).This may be because, for them, the close relationship also involves heightened personal costs: They have made their self-vulnerable by trusting a faulty or retrospectively inappropriate brand.Individuals high in CBI are, therefore, more disappointed and ego-involved, which translates into a more intense feeling of aggression toward the brand compared to individuals low in CBI.Hence: H4.Following a self-relevant failure of consumers' favorite brand, the greater consumers' consumer-brand identification is, the greater the impact of dissatisfaction on brand anger.

The moderating role of consumer-brand identification on the dissatisfaction fi brand shame effect
In a similar vein, the unfavorable effect of dissatisfaction on brand shame intensifies for consumers with high CBI.This prediction is grounded on identification theory (see above) and here particularly on social identity theory.Accordingly, if a person's self and brand concepts are connected, a brand's performance affects the individual's self-concept.Brands have been shown to be important instruments for consumers to construct a positive self-concept (Chaplin and John, 2005).In line with earlier research (Cheng et al., 2012), this study claims that consumers with a high CBIand hence with a strong connection between their self-concepts and a specific brandparticularly perceive a tight coupling between their favorite brand's performance (i.e.being a reliable relationship partner by fulfilling a service correctly) and their own performance (i.e.choosing a reliable brand).Psychological research demonstrates that individuals feel threatened when they fail to perform in an expected way (Tesser, 2000).Consequently, it is theorized thatrelative to low onesindividuals with high CBI respond to a brand failure in a similar way as if they had committed a personal failure.The more dissatisfying a service failure is (and hence the greater the brand's perceived deviance from personal values), the more consumers experience a strong threat to their positive self-concept.This translates into a more intense level of brand shame as a self-conscious negative emotion (i.e.internalization of dissatisfaction).Therefore: H5.Following a self-relevant failure of consumers' favorite brand, the greater consumers' consumer-brand identification is, the greater the impact of dissatisfaction on brand shame.

2.3.3
The moderating role of consumer-brand identification on the brand shame fi brand anger effect Continuing this reasoning, the unfavorable effect of brand shame on brand anger is likely intensified by increasing CBI levels.This proposition is based on the literature on dissonance reduction (McGrath, 2017, for a review), self-affirmation (Sherman and Cohen, 2006) and shame regulation (Norder et al., 2022).First, when a brand fails, consumers experience a crack in the cognitive fit between them and the brand, as well as ideological concerns due to an infringement of joint values (Whitler et al., 2021).This causes both a severe threat to their self-evaluation and a cognitive dissonance.Consequently, individuals are highly motivated to eliminate psychological discomfort by restoring a positive self-concept.Shame, an affective sign of cognitive inconsistency is assumed to trigger dissonance reduction coping strategies, including the denial of responsibility and the willingness to engage in defensive actions (Zeelenberg and Pieters, 2004).In this respect, people have been shown to extend their ego-defensive motivation to people, places, events and entitieslike brandsassociated with their self-concept (Davvetas and Diamantopoulos, 2017).
Second, this research assumes that in the event of a service failure, individuals with high CBI reflect on values that are personally relevant to them (e.g.endorsing a brand).The failure clearly reveals the discrepancy between dissatisfied consumers' important values and the brand's violation of these values.Consequently, the consumer engages in self-affirming activities that reduce forgiveness toward the brand.Individuals with high CBI have an intense interest in maintaining an overall positive concept of self-integrity.Individuals tend to construe their selves in line with their long-term values, rather than their short-term personal or brand-related failures (Schmeichel and Vohs, 2009).
Third, individuals with high CBI have a particularly strong drive to restore their overall positive self-view and to reduce shame.According to social identity theory, maintaining a favorable self is of utmost importance.Being aggressive toward the brand is perceived as an appropriate way to cope with this self-esteem harming situation (Michael et al., 2007).It is common for self-conscious emotionslike shameto be caused by a frustration of self-relevant goals (Mills et al., 2007).Hence, they are likely to emerge with a self-relevant service failure.Destructive emotional coping is a natural reaction.This is particularly true for individuals with high CBI whoas compared to individuals with a low identificationperceive a stronger restriction of their personal goals and consequently feel more ashamed and develop aggressive feelings toward the brand.Anger helps them to separate their selves from the brand.Therefore: H6.Following a self-relevant failure of consumers favorite brand, the greater consumers' consumer-brand identification is, the greater the impact of brand shame on brand anger.

Method
Main study.To assess the hypotheses, two studies were conducted.In the main study, data was collected from a sample of 660 adult consumers active on social media (64% female; M age = 28.3,SD age = 9.4) by means of a scenario-based online survey and a standardized questionnaire (in English).
We applied convenience sampling with a snowball technique by triggering contacts with students and their acquaintances via social media.Data collection took place in Central Europe, with most participants being from Austria (23%).Data cleansing (e.g.biased response patterns, satisficing and language proficiency) reduced the original sample of 685 respondents beforehand.The applied survey procedure was adopted from Davvetas and Diamantopoulos (2017).The first part asked every participant to spontaneously indicate their favorite brand in a freely chosen product category.Purification of responses identified and corrected potential biasing patterns.
The respondents named over 350 different brands (Apple was mentioned most frequently by 15%) in 90 product categories (including technology, fashion, cars, food and personal care products).Following the indication of the favorite brand, participants completed a series of questions that included CBI with their favorite brand (Stokburger-Sauer et al., 2012) and additional descriptives of their brand relationship status (e.g.relationship length).
The second part presented a short scenario asking subjects to imagine that they recently bought their favorite brand, but soon after that, they recognized a rather small but symbolic brand failure (see Supplementary Appendix 2 for details).The flexible framing allowed approaching a general audience rather than restricting to a sample of consumers who share the same favorite brand.Utilities offered by one's favorite brand are highly diverse.But purchasing such brands has in common a price that has to be paid in exchange.This made it possible to frame a scenario of subjects' (different) favorite brands and at the same time sharing the same kind of brand failure (the validation study eliminates potential brand-specific effects by presenting subjects with a given brand).Hence, all subjects' purchasing expectations were not met and they experienced the same kind of service failure.Finally, participants completed the second part of the questionnaire with well-established measures from academic research for dissatisfaction (DIS), brand shame (SHA), brand anger (ANG) (Supplementary Appendix 3 presents the wording of the items), as well as two items assessing the perceived scenario believability (i.e."The scenario was realistic.";"I could easily put myself in the situation described in the scenario.").The questionnaire also included some additional items to distract the respondents to safeguard against potential common method effects.The sequence of all items was randomized, and seven-point response formats (ranging from 1 = "I totally disagree" to 7 = "I totally agree") were used throughout.The survey ended with demographic questions.
Validation study.To cross-validate our findings, a scenariobased online experiment was conducted in which CBI was manipulated (low vs high) (see Supplementary Appendix 2).Here data from 1,616 participants (57% female; M age = 27.1,SD age = 8.6) was collected by the same technique as in the main study (see Supplementary Appendix 4 for sample characteristics).In the between-subjects design, participants were asked to read through a scenario describing a symbolic brand failure of their favorite restaurant (i.e. a service failure series occurring during a visit of their preferred restaurant) before responding to a standardized online questionnaire, including the same constructs as the main study.The participants rated all items on five-point scales, whereby larger values indicated higher agreement.

Measurement model evaluation
Confirmatory factor analysis (CFA) investigated the psychometric properties of the latent variables' measures.The measurement model fitted the data well (main study: x 2 = 97.05,df = 57, RMSEA = 0.03, CFI = 0.99, SRMR = 0.02; validation study: x 2 = 845.46,df = 38, RMSEA = 0.12, CFI = 0.93, SRMR = 0.07).Further analysis established both validity and reliability as high Cronbach's alpha coefficients [ranging from 0.87 (0.67) to 0.90 (0.96)], satisfactory item-to-construct standardized loadings [ranging from 0.72 (0.61) to 0.89 (0.95)] and appropriate composite reliabilities [ranging from 0.83 (0.68) to 0.89 (0.96)], as well as good average variance extracted values [ranging from 0.61 (0.42) to 0.76 (0.88)], almost always exceeded the conventional thresholds.Convergent validity was confirmed, as the average variance extracted of all constructs in the main and validation study exceeded the 0.50 threshold, except for dissatisfaction in the validation study.Since dissatisfaction has a composite reliability >0.60, its convergent validity is still adequate (Lam, 2012).Overall, the data supported discriminant validity (Fornell and Larcker, 1981).Supplementary Appendices 3 and 5 provide a summary of the psychometric properties by displaying means and standard deviations of the constructs, as well as the correlations between constructs.
As all measures are self-reported, common method bias (CMB) (Podsakoff et al., 2003) could be a potential issue.Different implemented procedures safeguarded against this problem (MacKenzie and Podsakoff, 2012), e.g.easing participants' recall task to a service/product failure of their choice, using a clear temporal separation between this retrospective measurement and scales capturing emotional responses, enhancing participants' motivation with careful instructions and minimizing scales' repetitiveness.Furthermore, statistical methods investigated potential CMB.In more detail, CFA evaluated three alternative measurement models of the four latent variables: (i) a one-factor model (all items are conceptualized to load on a single common factor); (ii) a four-factor model (all items loading on their respective factors); and (iii) a five-factor model, expanding the four-factor model by an unmeasured latent method factor which was constrained to load equally on all items.
For the main study, the one-factor model fitted the data poorly (x 2 = 2,003.71,df = 78, RMSEA = 0.21, CFI = 0.67, AGFI = 0.45).When comparing the original four-factor model (see above) with its five-factor alternative (x 2 = 91.63,df = 58, RMSEA = 0.03, CFI = 0.99, AGFI = 0.97), no substantial improvement in model fit upon inclusion of an unmeasured latent method factor was found.Furthermore, no substantial differences emerged for any of the central constructs when comparing standardized regression weights in both models.In sum, these results, which similarly emerged in the validation study, indicate that CMB was not a significant issue in the data.

Results main study
Scenario plausibility.Empirical realism, i.e. that participants regarded a failure with their favorite brand as imaginable, was proven since the respondents could really put themselves in the situation described in the scenario: the mean score of the two check variables (r = 0.69, p 0.001) was significantly higher than the scale's mean (M = 4.17, SD = 1.73, t = 9.99, p 0.001).
Main results.Conditional process analysis tested the moderated-mediation model for brand emotions (see Figure 1), that is, Model 59 of the SPSS PROCESS macro (Hayes, 2018) with 10,000 bootstrap-samples.This approach allowed the bias-corrected bootstrapped confidence intervals for the hypothesized mediation effect to be obtained and to further probe the interactions at different levels of the moderator.In this model, DIS was the independent variable, SHA the mediator, ANG the dependent variable and CBI the moderator.Several covariates contributed to explain variance: failure severity, prior failure experiences, relationship length, product category, age, gender, and country of residence.
Explanatory power and overall statistical significance of the two regression analyses were satisfactory (Table 1, rows "R 2 , F").The results showed thatin support of H1 -DIS had a significant positive effect on ANG (b 1 = 0.47).Further, DIS had a significant positive effect on SHA (b 2 = 0.69), supporting H2.Affirming H3, the effect of SHA !ANG (b 3 = 0.30), and the indirect effect of DIS !SHA !ANG (b 2 •b 3 = 0.21) were significant with postulated signs.Thus, brand shame partially mediated the relationship between consumers' dissatisfaction and brand anger (see Table 1).
In addition, to demonstrate that anger triggers brand-aversive intentions, we conducted further analyses.They showed that the more angry individuals get, the more they were inclined to voice NWOM (i.More importantly, however, the study provided evidence for CBI's moderating role: Specifically, the results revealed a significant positive interaction effect of DIS Â CBI on ANG (b 4 = 0.04), supporting H4; of DIS Â CBI on SHA (b 5 = 0.07), supporting H5; and finally of SHA Â CBI on ANG (b 6 = 0.06), supporting H6.Therefore, CBI exerted a reinforcing effect on the relationships between consumers' unfavorable postfailure emotions (see Table 1).
Additional results.After having verified all research hypotheses, the interactions were subsequently probed at one standard deviation above/below the mean level of CBI (see Table 1, rows "conditional direct effects").The results revealed that the conditional direct effects of DIS on ANG and on SHA, as well as of SHA on ANG, remained significant and had the hypothesized direction across the varying moderator levels.Even more important was the finding that the magnitude of these effects was significantly conditioned by the levels of CBI.This meant that, while moving from low (ÀSD) to high (1SD) CBI levels, the effects of DIS on ANG, of DIS on SHA and of SHA on ANG increased in magnitude (see Table 1).This finding is also demonstrated in Supplementary Appendix 6 for the DIS !SHA effect, as the slopes got steeper when CBI increases.In further support of H3, this pattern also holds for the conditional indirect effect of DIS on ANG via SHA, and as a consequence on the conditional total effects of DIS on ANG (see lower part of Table 1).Supplementary Appendix 7 corroborates the results further by testing rival model specifications.

Results validation study
Empirical realism of the validation study was again confirmed since the mean score of the two check variables was significantly higher than the scale's mean (M = 3.88, SD = 0.89, t = 40.03,p 0.001).The direct effects proclaimed in H1-H3 were successfully cross-validated with the scenario-based online experiment: DIS !ANG (b = 0.80, p 0.01), DIS !SHA (b = 0.64, p 0.01) and SHA !ANG (b = 0.04, p 0.05).Again, the detrimental effect of anger on NWOM increase was shown (ANG !NWOM) (b = 0.45, p 0.001).Furthermore, separate regression analyses for lowly vs highly identified consumers provided empirical support for the moderation hypotheses.More specifically, two Chow tests (Chow, 1960) confirmed that all regression coefficients were significantly different: Among individuals with high CBI, the coefficients were consistently higher as compared to lowly-identified individuals (see Supplementary Appendix 8).Thus, higher CBI levels intensify the negative spillover effects as proposed in H4-H6.

General discussion
Despite the large quantity of research on brand failures (Torres et al., 2021), knowledge on the joint role of consumer emotions and strong customer-brand relationships during those unpleasant experiences remains limited so far.Such situations are predisposed to elicit feelings typical for 'brand shame', like parting from social norms and closeness to the unattractive self (Tangney and Fischer, 1995).Shame has been shown to arise in different contexts, ranging from intrapersonal processes to the consumer context (see Supplementary Appendix 1).However, shame in the context of failures of a consumer's favorite brand (i.e. a brand that was earlier highly preferred over others) has been scarcely studied.In the rare cases in which shame has been investigated after unpleasant service experiences, the goal was typically to explain what drives shame (Sugathan et al., 2017).Nevertheless, this research extends this view by arguing that consumers follow a complex emotional-cognitive process after brand failures (i.e.situations when a product/service performance falls below customers' expectations).In more detail, brand shame (i.e. an inwarddirected negative emotion) should be considered as a key mediator between failure-induced dissatisfaction (i.e. a relatively diffuse negative emotion) and brand anger (i.e. an outward-directed negative emotion).In literature, the direct link between customer dissatisfaction and anger is quite well established (Chen et al., 2021;Yang et al., 2022) and can also be supported by this study.However, research on the mechanisms behind this relationship is still sparsebut nevertheless regularly called for (Hejdenberg and Andrews, 2011).This research builds on Lewis' (1971) original shamerage theory that posits that dissatisfaction-induced shame can instigate a specific hostile type of anger described as humiliated fury, which has later been shown to engage individuals in several brand-aversive habits, such as boycotting and NWOM (Septianto et al., 2020).Our cross-validated research demonstrates the role of shame as a driver of anger.Here, this study also connects to the evolving body of literature focusing on a more fine-grained view on consumers' various negative emotions toward earlier preferred brands (Haase et al., 2022).
Most importantly, based on literature showing that bonded consumers are more annoyed and demanding after brand failures (Wolter et al., 2019), this research proposes and empirically confirms that CBIas a cognitive indicator of prefailure relationship status in terms of the self-relevance of the brand to the consumerintensifies the direct and indirect (via brand shame) unfavorable effect of failure-induced dissatisfaction on brand anger.The hostility-inducing power of close ties is consistent with the "love-becomes-hate" effect (Gr egoire and Fisher, 2008).The research at hand confirms recent insights (Weitzl and Einwiller, 2020) that CBI regularly intensifies loyal consumers' harsh reactions (e.g.anger) following unmet expectations involving their earlier favorite brand.

Theoretical implications
Consequently, at least two theoretical implications emerge.First, this research contributes to the service failure literature.Here, extant research has mainly focused on the emotional, cognitive and behavioral consequences of negative service incidents of regular brands (Wang et al., 2023).For instance, Sameeni et al. (2022) show that such negative experiences can elicit customer regret.At the same time, this literature stream demonstrates researchers' sincere interest in the effects of different types of failures, such as logistical vs nonlogistical incidents (Roy et al., 2022).The research at hand, however, is one of the first to focus on symbol-laden failures of customers' most favorite branda topic which is hardly covered by earlier studies (Davvetas and Diamantopoulos, 2017, for one of the exceptional contributions).Nevertheless, studying the causes which may erode consumers' relationships with their most preferred brand is critical.Our results reveal that shame plays a vital role.So far, literature on negative emotions after dissatisfactory consumption incidents has mainly studied outward-directed emotions, most notably anger (Septianto et al., 2020).
In contrast, work on inward-directed emotions is scarce.Here, guilt has been the dominant subject of examination in literature.Guilt typically focuses on a perceived transgression of one's behavior (e.g."I made a bad purchase") (Tangney and Dearing, 2003) and makes individuals blame a certain behaviorbut not their entire selffor the negative events (Tangney et al., 2007).Some literature suggests that consumers' guilt after service failures reduces complaining and NWOM (Soscia, 2007).However, shame and guilt represent distinct emotional experiences (Tangney et al., 2007).This research empirically shows that consumers can experience brand shame after selfrelevant failures.Consumers feel ashamed because they regard themselves as responsible for choosing a "betraying" brand that is involved in an incident perceived as self-relevant and incompatible with their personal norms, standards and selfgoals (Tan et al., 2021).As such, brand shame focuses on a perceived transgression of one's self (e.g."I am a bad consumer") (Tangney and Dearing, 2003).
This study is the first to indicateconsistent with literature from social psychology (Sun, 2018) that consumer dissatisfaction might induce brand shame.Facing an identity threat caused by a brand failure, strongly harms consumers' self-concept.Consumers tend to overcome this cognitive dissonance by applying forms of self-affirmation (Cheng et al., 2012).Given that blaming others for a failure has been shown to have an ego-protective function (Tangney and Dearing, 2003), this research highlights that after self-relevant failures, consumers who experience brand shame make the brand responsible and are likely to develop the others-directed negative emotion of brand anger (Heaven et al., 2010).This study demonstrates that brand shame has a reinforcing effect on brand hostility because consumers protect their ego after their most preferred brand has hurt them.
Second, this research adds to the literature on CBI.Many consumers regard brands as relational entities with which they establish close bonds.In this regard, CBI acts as a central cognitive indicator of prefailure relationship status in terms of the self-relevance of the brand to the consumer.Close bonds can have brand-protective effects in some failure situations (e.g. when consumers attribute a problem to a controllable cause; Torres et al., 2021), but can trigger negative outcomes under other conditions.Earlier research demonstrates that loyal consumers often react more negatively than their unbonded counterparts (Zhang et al., 2021) and that this negative reaction is conditional on several circumstances (e.g.failure history; Hutzinger and Weitzl, 2023).Extant literature on relational bonds suggests that following brand transgressions, the perceived threat of the individual's self can trigger downstream effects (Davis and Dacin, 2022;Sameeni et al., 2022).Yet, no research has studied the role of close cognitive bonds after symbolic failures of consumers' favorite brand and its unique emotional aftermath.Results imply that in situations, which demonstrate a favorite brand's inability to provide psychological benefits, CBI intensifies the direct and indirect (via brand shame) effect of failure-induced dissatisfaction on brand anger.Therefore, these results add to the literature showing that consumers' strong cognitive bonds with the favorite brand can lead to highly negative reactions after brand failures, denoted as the "love-becomes-hate" effect (Wolter et al., 2019).This research is the first to introduce the concept of brand shame to study the "love-becomes-hate" effect and to investigate its dependency on CBI.In turn, CBI is an important cognitive moderator in explaining the transition from inwardto outward-directed negative emotions after brand failures.

Managerial implications
For marketing practitioners, this study also conveys important managerial implications.First, for marketers, it is vital to note that a strong cognitive connection between consumers and their favorite brandsin addition to favorable consequences, such as increased customer satisfaction and purchase intention (Dash et al., 2021) can severely backfire in the event of brand failures.Great customer-brand relationships can become very toxic from one moment to the next.The more consumers identify with a brand, the more negative their emotional reactions might be in case of negatively perceived product/ service incidents.This research shows that dissatisfied customers not only internalize the shame they perceive toward their connection with an erroneous brand right after a failure.Soon after that, they externalize their negative emotions by building anger toward the brand.The important thing to realize is that this anger is even amplified the more consumers identify with the brand.Consumers' anger has been shown to harm brands dramatically through acts of NWOM.
This does not imply that marketers should refrain from building strong relationships, but that they should reduce anger-inducing circumstances, such as limiting the perceived frequency of brand failures.Here, for example, recent research shows that bonded consumers are quite forgiving when they have experienced no or only few as compared to multiple failures (Hutzinger and Weitzl, 2023).Individuals may have an incentive to forget negative experiences (i.e.self-deception) to maintain self-esteem (Michael et al., 2007).Marketers can apply the mechanism that recollection of positive events is easier than recollection of negative ones by emphasizing prior personal positive experiences of the consumer (i.e.celebrating the joint brand history in personalized recovery responses) and creating new ones (e.g. an outstanding compensation), which stimulates a positive mood (Li, 2013).
Second, marketers should realize that, after problems involving their favorite brand, customers develop brand shame, which significantly harms the brand.Moreover, after initially blaming themselves for the unpleasant brand experience, it is important to note that customers take out their negative emotions on the brand.Thus, marketers should be knowledgeable about potential strategies to reduce consumers' brand shame.Psychological literature (Yelsma et al., 2002;Norder et al., 2022) discusses several coping styles how individuals manage shame.In the aftermath of a failure, marketers could, for example, cultivate self-compassion (i.e.fostering a sense of empathy for oneself), a sense of humility or self-acceptance with proactive recovery efforts, which may reduce feelings of defectiveness and unworthiness, particularly among highly identified individuals: (e.g."You are still a smart consumer!A single failure does not prove you wrong that you made a wise brand decision in the past!").
Self-forgiving could be achieved, among others, by reducing the relevance of the negative experience to construct the self.Marketers could here emphasize brand characteristics that are still valuable for the construction of the self-image: (e.g."We are very sorry for this problem!We can assure you that we haven't let you down and it's still our utmost goal to satisfy your needs.We had so many good times together.You always knew that our restaurant is the best in town for connoisseurs deserving an outstanding culinary experience.";Exemplary countermeasures refer to the scenario presented in the validation study).
Furthermore, the strategies of withdrawal and avoidance can be used.Through applying withdrawal, individuals try to distance themselves from, or at least minimize contact with, the shame-inducing situation.Avoidance helps individuals to reduce awareness of shame by either denying it or replacing it with excitement and joy in most cases.Marketers could adapt these shame-reducing strategies to the context of brands.By fostering withdrawal, consumers' minds can be shifted away from the unpleasant situation by assuring them that the brand failure was not such a strong infringement of values.Marketers can emphasize the multitude of shared values left between them and the consumers.Recovery statements like: Your valued criticism shows that you and our restaurant still share the same passion for food.We are grateful that you have reminded us about our everexisting core value: striving for the best!A cold dish does stand in contrast with our dedication.We still share the same values.
Might be helpful in getting rid of (particularly highly identified) consumers' doubts.Marketers should refrain from digging deeper into the underlying shame-inducing brand failure and rather provide a positive outlook into a failure-free future governed by common norms.By fostering avoidance, consumers who suffer from a brand failure can be assured that there is no need to feel shame.Marketers can state that failures are not only an important aspect of personal life but also a basic stimulus for modern organizations to be truly innovative.
In addition, marketers can deliver shame-and angerreducing humor in their recovery messages to especially individuals with a high CBI to increase their fun and excitement.Experiencing positive emotions might help dissatisfied customers let go of their brand shame and adopt a more favorable emotional stance toward the brand.Particularly self-deprecating humor (i.e.amusing consumers by joking at the brand's own expense), which is perceived as particularly funny, is adequate to help consumers to overcome negative experiences (Kobel and Groeppel-Klein, 2021).A response after a restaurant failure can read as follows: "When you choose another dish, please bear in mind that we are all out of snailswe needed them for the service today, as you might have noticed."Thus, marketers are well advised to consider service recovery after brand failures primarily as a joint identification-, relationship-rebuilding and favorable mood-inducing activity, as compared to a pure financial reparation (Yang and Mundel, 2022).Here, the whole response strategy should be personalized and helpful to restore the consumer's self-image and minimize shame by eliciting self-forgiveness.

Limitations and future research
While the findings provide several critical implications for theory and practice, some limitations pave the road for future research.First, this research explores brand shame as a crucial mediator between brand fans' dissatisfaction and brand anger using a convenience sample.It must be acknowledged that this role is maybe only valid in the context of favorite brands' failures.Cross-validation studies using more rigorous sampling techniques should be conducted to confirm this finding, but also further boundary conditions (e.g.failure attributions) impacting the magnitude of induced customer shame should be assessed.Here, particularly the impact of different types of service failures (e.g.unfriendly staff, malfunctioning product) could be interesting.Furthermore, other inward-directed emotions, such as guilt, might play a potentially different role.Additionally, the moderating effect of CBI on these emotions needs to be explored in future research.
Second, this research focuses on internal shame, which is felt when individuals perceive that their deep-seated flaws are exposed to themselves.In addition, when these flaws are also exposed to others, individuals feel external shame (Matos et al., 2015).At the same time, however, external shame might reduce brand-damaging behaviors as consumers might want to hide their humiliation, which needs to be verified by future research.On a related note, additional personality variables tend to play a role.Results might depend on the degree individuals consider it important to protect their positive social image in relationships (Chan et al., 2009).Thus, this should be examined for externally shameful brand failures in the future.
Third, this work did not consider service recovery efforts (i.e.webcare), which have been shown to impact complainants' cognitions and behaviors (Weitzl et al., 2018).Classic recovery strategiesmost notably, defensive (e.g.denying responsibility) and accommodative webcare (e.g.apologizing)have been extensively studied.Inducing fun and excitement in affected consumers might be a promising option to counter negative feelings.Initial results about humor in webcare responses reveal that it has a positive impact on consumers' excitement (Shin and Larson, 2020).Further exploring self-restoring recovery options in the future is a worthwhile endeavor.

Table 1
Moderated-mediation analysis (main study)Notes: a Nonstandardized regression coefficients; b Mean centered for products; c Detailed results on the covariates included are available upon request from the corresponding author; d CBI evaluated at average level and average level 6 one standard deviation; e H3, H4, H5 and H6 refer to the corresponding hypotheses investigated; f Bootstrap confidence intervals (for a Type I error of 0.05; 10,000 bootstrap samples) Source: Authors' own work