Analysis of consumer boycott behavior from the perspective of brand hate: A mixed-methods approach

ABSTRACT There is a growing interest in understanding negative interactions between brands and consumers. Consumers have become increasingly sensitive to brand-related controversies (Roy et al., 2022). With digitalization and especially the spread of social media, the accelerating speed of information dissemination (Mehta et al., 2020) can transform corporate missteps into large-scale public relations crises in a very short time (Sheth, 2020). The negligence or disregard of organizational decisions' and activities' effects on consumers may precipitate adverse outcomes, including the erosion of consumer confidence, the deterioration of corporate reputation, and the formation of negative brand attitudes among consumer segments (Keskin et al., 2024). According to Woolverton and Dimitri (2010), businesses should make decisions that consider societal interests, recognizing the environment as a critical factor. Indeed, consumers can love a brand for its high quality while simultaneously hating it for failing to provide good working conditions to its employees (Prestini and Sebastiani, 2021). Hate functions as a force triggering both emotional and behavioral responses (Fetscherin, 2019). Brand hate carries a deeper emotional intensity than simple dislike, causing consumers to establish a strong negative bond with the brand. In this context, hate emerges as a consequence of negative situations in consumer-brand relationships. Factors triggering brand hate are the fundamental drivers of specific emotions in consumers, which in turn direct them toward certain behaviors (Mushtaq et al., 2024). The consequences of brand hate encompass a wide spectrum including negative word-of-mouth communication, online and direct complaints, brand switching, brand avoidance, seeking revenge (Haase et al., 2022), and boycott behavior (Bryson et al., 2013; Nguyen, 2021). Brand hate research has gained increasing academic attention in recent years. Quantitative methods have been predominantly adopted in these studies, while qualitative and mixed methods have been used relatively infrequently. The systematic literature review conducted by Mushtaq et al. (2024) revealed that brand hate is an emerging research area requiring more comprehensive studies due to its applied nature. In response to the growing need for mixed methods in the literature, this thesis study examines brand crises, which play a triggering role in the emergence of brand hate, and boycott intention, which is one of its consequences, through the integration of quantitative and qualitative methods. This study employs a mixed-method approach: an experimental design using a hypothetical brand for quantitative analysis, followed by a qualitative examination of real consumer reactions to a recent brand crisis. This dual methodological approach allowed for understanding the relationships between these variables both under controlled experimental conditions and within real market dynamics. Indeed, as stated by Assoud and Berbou (2023), increasing the use of secondary data in research is important to overcome the limitations of primary data. Online streaming platforms, which have widespread use in the digital ecosystem, were chosen as the application area of the research. In the systematic literature review on brand hate conducted by Rai et al. (2023), it was found that existing studies focus on traditional sectors such as food, retail, fashion, and banking. However, examining different businesses and sectors was suggested to reach more comprehensive and generalizable findings on brand hate. As emphasized by Kesse et al. (2021), the antecedents and consequences of brand hate can vary according to brands, sectors, and cultural contexts. When examining the relevant literature, it was observed that no comprehensive research has been conducted on digital streaming platforms. This situation makes it meaningful for the thesis to base both the quantitative and qualitative phases on online streaming platforms as a sector. The theoretical framework of the research is based on the process whereby negative emotions first emerge in consumers and subsequently shape various attitudes and behaviors. Within the scope of the quantitative study, the S-O-R (stimulus-organism-response) model serves as the theoretical foundation to examine the relationships between brand crises as environmental stimuli, brand hate as the organism, and boycott intention as the behavioral response. This model explains that environmental factors to which individuals are exposed cause changes in their internal states, and these changes transform into behavioral responses. According to Bagozzi (1986), when consumer behavior is depicted through the S-O-R system, stimuli consist of both marketing mix variables and other environmental inputs. The organism refers to the internal processes and structures that occur between external stimuli and the resulting final actions, reactions, or responses. These intervening processes and structures consist of perceptual, physiological, emotional, and cognitive activities (Bagozzi 1986, p. 46, cited in Chang et al., 2011). Response refers to the final consumer behaviors, including approach or avoidance (Chang et al., 2011). Within the scope of quantitative research, brand crises were determined as the stimulus, brand hate as the organism, and boycott intention as the response. Research findings by Hegner et al. (2018) reveal that different types of crises create differentiated effects on consumer reactions. Brand crises are classified as moral crises that arise from violations of ethical values and product (performance) crises that result from disruptions in technical/physical elements (Hegner et al., 2018). Both crises related to corporate ethics and performance can negatively affect consumers' purchasing tendencies (Baghi and Gabrielli, 2019). Consumers react negatively when confronted with serious corporate misconduct and evaluate the potential costs and benefits of participating in a boycott (Klein et al., 2004). Therefore, the moderating role of consumer boycott motivations is examined in the research model. Two types of boycott motivation are defined in the literature: instrumental and expressive. In instrumental boycotts, goals are clearly stated and structured (Hendarto et al., 2018). This type of boycott aims to create pressure to make the target company change a controversial policy. Expressive boycotts, on the other hand, are a general form of emotionally charged protest that conveys consumers' dissatisfaction with the actions of the target company (Klein et al., 2001; Hendarto et al., 2018). Neuroticism and agreeableness, among consumer personality traits, were also included in the experimental research as moderating variables. Banerjee (2016) suggested that consumer personality influences customers' purchasing decisions. Research in psychology shows that personality traits influence negative behaviors such as avoidance or revenge (McCullough et al., 2001). While individuals with high moral anger exhibit higher boycott intention, especially in ethical violation situations (Smith, 1990, cited in Kübler et al., 2020), individuals open to forgiveness are less likely to avoid the brand or make negative criticisms and more likely to repurchase (Fetscherin and Sampedro, 2019). Therefore, based on recommendations for detailed examination of consumer personality traits in future brand hate research (Chahal and Dolkar, 2020; Taqi et al., 2024), it was deemed appropriate to include neuroticism and agreeableness, among the Big Five personality traits, as moderating variables in the model. Indeed, in real-world contexts, variables are generally conditionally related to each other, and moderating variables are needed to explore the nature of these relationships. In the methodological design of the quantitative research, a single-factor experimental design approach was adopted, creating two different experimental groups where the brand crisis type (moral (ethical) crisis vs. performance crisis) was manipulated. Participants were presented with different news scenarios where crisis types were manipulated, and subsequently asked to evaluate scale items prepared in accordance with the research hypotheses. In testing the research hypotheses, the relationships between the dependent and independent variables were first analyzed with an independent samples t-test. The analysis results showed that the independent variable, brand crisis type (moral x performance), has a positive and significant effect on the dependent variable, boycott intention. Detailed examination revealed that moral crises lead to a higher level of boycott intention compared to performance crises. After confirming the effect of brand crisis type on boycott intention through independent samples t-test analysis, Models 4, 5, and 7 in the SPSS Hayes Process Macro were used to examine the effects of the mediating and moderating variables included in the research model. The findings revealed that moral crises, compared to performance crises, cause both higher levels of boycott intention and more brand hate. Additionally, the mediating role of brand hate in this relationship was statistically confirmed. However, the fact that the relationship between crisis type and boycott intention maintained its significance with the inclusion of brand hate in the model indicates that this mediation effect is partial. It should also be emphasized that the explanatory power of the partial mediation role of brand hate is quite high (R2 = 58%). This finding indicates that a brand crisis largely transforms into boycott intention through the mediation of brand hate. The inclusion of moderating effects in the analysis did not lead to any change in the basic relationships between dependent, independent, and mediating variables. Neuroticism, among consumer personality traits, was found to have a significant moderating effect. It was established that neurotic individuals develop a higher level of hate and exhibit stronger boycott intentions, especially in moral crisis scenarios. When examining the effect variation based on neuroticism levels, it was found that individuals with high neuroticism show stronger hate and boycott intention in moral crisis scenarios compared to individuals with low neuroticism. Conversely, analyses conducted for agreeableness revealed that this variable does not have a significant moderating effect. In other words, no moderator effect of agreeableness was detected in the indirect relationship where brand hate has a mediating role between brand crisis and consumer boycott intention. Similarly, in the analyses for consumer boycott motivations, no significant moderator effect was found for either instrumental or expressive motivations. Although the results show that boycott motivations have a direct significant effect on boycott intention, the moderating effect tested in accordance with the research hypotheses was statistically insignificant. Following recommendations by Keskin et al. (2024) regarding the use of secondary data in brand crisis analysis and calls for mixed methods in brand hate research (Sakulsinlapakorn and Zhang, 2019; Yadav and Chakrabarti 2022; Taqi et al., 2024, Chahal and Dolkar, 2024), it was concluded that complementing quantitative research findings with qualitative analysis would provide valuable insights. The quantitative findings showing that moral crises cause higher levels of hate and boycott behavior compared to performance crises, supported by existing literature (Kübler et al., 2020), necessitated the examination of a real ethical brand crisis case. Following the experimental research using a hypothetical brand, the crisis experienced by Disney+ in Turkey regarding the "Atatürk" series was selected as a case study to examine in-depth consumer reactions under real-world conditions. The boycott movement that began after it was learned that the platform canceled the broadcast of the series (whose filming was completed and trailers were released) in August 2023 constitutes a moral crisis, as consumers perceived it as a disrespectful and racist attitude toward an important societal figure. Literature indicates that in response to unethical or unfair corporate actions, consumers can experience intense negative emotions such as anger, which effectively predicts negative behavioral responses, including conscious boycotting of the firm's products and services (Makarem and Jae, 2016). At this stage, focusing on consumers' emotional intensity was deemed important. The methodological limitations of the quantitative research did not allow for an in-depth analysis of consumer emotional intensity. According to Zhang and Laroche (2020), brand hate is a multi-layered emotion experienced at different intensity levels, and consumers can experience this hate at both low and high intensities. Küçük (2019) developed a hierarchical model explaining brand hate based on Sternberg's (2003) psychological theories. This model examines both one-dimensional and multi-dimensional brand hate according to hate intensity. According to Sternberg's (2003) theory, the combination of different emotions can transform into a single emotion like hate. Within this model, "cold hate" (representing a mild level of aversion) first progresses to the "cool hate" stage (an intermediate level), then reaches its most intense form, the "hot hate" stage (characterized by high emotional intensity). Cool hate is associated only with brand switching, while brand sabotage is exclusively associated with hot hate. According to Fetscherin (2019), hot hate consists solely of anger and is characterized by an intense feeling of anger in response to threat. Hate at this level includes consumers' willingness to make material sacrifices to harm the brand. Within the qualitative research, the aim was to observe which hate levels interact with which motivations to form boycott intention. Accordingly, both the themes of boycott motivations were determined and the relationships between motivations and hate levels were analyzed. Another focus of the qualitative research was determining the antecedents of brand hate. The presence of data covering the pre-crisis period and consumers' posts reflecting their previous experiences/opinions triggered during the crisis have made the identification of these themes particularly important. Given the literature indicating that outcomes differ for each level of brand hate, clarifying relationships between emotion levels and brand hate antecedents was also pursued. In the qualitative research phase, a total of 5729 data points were collected from Instagram, Twitter (X), and Ekşi Sözlük through netnographic analysis examining consumer-generated content on social media platforms. Considering the evolution of platform target audiences and consumer profile differentiation (Kwon and Park, 2023), determining whether consumer hate levels also differ among platforms constituted another research question. First, manual sentiment analysis was performed, revealing different hate levels across all three platforms. Cool hate was more prevalent on Ekşi Sözlük, cold hate on Twitter (X), and hot hate on Instagram. In the content analysis, consumer boycott motivations were grouped into three categories: instrumental, expressive, and mixed motivation. Instrumental motivations consist of four sub-categories: "call to action," "creating awareness", "desire to make a difference", and "directing to alternatives". Expressive motivations comprise five sub-categories: "self-protection", "self-defense", "self-enhancement", "expressing support", and "desire to threaten/punish the target brand". Since some consumer messages contained elements of both motivation types, these were coded as mixed motivation. However, these motivations were only included in the frequency tables section of the findings, and mixed motivations were not exemplified in detail in the theme explanation section. While previous studies have explored instrumental and expressive boycott motivations, this research introduces additional categories, contributing novel insights to the literature. For example, Makarem and Jae (2016) examined the reasons and motivations underlying boycott behaviors in Twitter posts. Content analysis of 1422 tweets revealed that instrumental motivations were predominant, attributed to boycotts generally aiming to achieve specific goals. The findings of this thesis, however, provide an original contribution to the literature by revealing the dominant role of expressive boycott motivations in moral crises, enriching the conceptual framework of expressive boycott motivations with newly added themes. After examining consumer boycott motivations, themes related to brand hate antecedents were determined. These comprise five main themes: "brand-related", "product/service-related", "individual-related factors", "situational factors", and "indirect factors" along with various sub-themes. Brand-related factors include "country of origin", "brand management process", "poor customer service quality", and "negative brand personality". Product/service-related factors were identified as "poor performance", "content inadequacy", and "deficit-value". Individual factors emerged as "ideological incompatibility", "ethical/moral factors", "dissatisfaction", "symbolic incompatibility", and "perceived betrayal". Situational factors were identified as "gossip/rumors", "lack of information", "social environment", and "presence of competitors". Finally, indirect factors were determined as "sponsorship activities", "celebrity hostility", "political party hostility", and "sub-brand hate". Among the examined social media platforms, expressive boycott motivations were dominant on Ekşi Sözlük. Consumers' hate emotion levels were coded as no hate (neutral), cold hate, cool hate, and hot hate, and it was observed that individuals conveyed boycott messages with expressive boycott motivations across all four emotional states, including the neutral level. Individuals acting with instrumental boycott motivation showed a neutral emotional level at 37,5% and cold hate level at 31,78%. Consumers exhibiting the highest level of hate (hot hate) demonstrated expressive boycott motivation in 58,70% of cases. On Instagram and X, while instrumental motivations dominated at the neutral emotional level (56,35%; 78,99%), expressive boycott motivations gradually increased as emotional levels intensified. The relationship between hate levels and brand hate antecedents was also examined, with findings showing that the most intense hate level was related to brand-related factors across all three platforms. When all research findings are evaluated together, they align with existing literature regarding moral crises generating more negative attitudes and behaviors compared to performance crises (Kübler et al., 2020; Sameeni et al., 2023). Additionally, the stimulating effect of neuroticism on hate and boycott behavior has been confirmed. Although agreeableness was predicted to reduce hate and boycott intention, this effect was found to be insignificant. While the moderating role of boycott motivations in the relationship between crisis type and boycott intention was not significant in the quantitative research, analyses of post-crisis consumer-generated data reveal that expressive boycott motivations predominate in moral crisis situations. The variation in emotional intensity across social media platforms also confirms social media platform schema theory (Kwon and Park, 2023). This research has several important limitations. The use of a fictional brand in the quantitative study, where two different crisis types (moral vs. performance-based) were manipulated, limited the external validity of the research (Wang et al., 2021). In the qualitative research conducted to balance this limitation, only a single moral crisis case was examined. Future research comparing both moral-related and performance-related crises through real cases will enhance the generalizability of findings. Additionally, the limitation of the research to online streaming platforms and the netnographic analysis covering only a 10-week period represent significant constraints. Future research should examine crises across different sectors longitudinally and investigate the long-term effects of consumer boycott behaviors. Keywords: Brand Hate, Boycott Intention, Brand Crises, Consumer Personality Traits, Consumer Boycott Motivations