When hate does not last: the evolution of brand schema in online communities

Brand schema, a cognitive framework enabling the organization and interpretation of brand-related information (Halkias, 2015), plays an important role in influencing consumer decision-making and emotional brand connections (Bhaduri and Ha-Brookshire, 2017). Online platforms, however, have fundamentally changed how these cognitive frameworks develop and evolve through networked interactions (Kozinets et al., 2021). While existing studies have treated brand schemas as relatively stable cognitive structures (Peracchio and Tybout, 1996), the digital landscape demands a reconsideration of this assumption. The proliferation of online communities has transformed virtual spaces into arenas for collective negotiation of symbolic brand meanings (Hallier Willi et al., 2014). These communities exhibit distinctive engagement patterns, from passionate advocacy to vehement opposition. While positive engagement cultivates brand love (Albert and Merunka, 2013), negative engagement can manifest as anti-brand behaviors (Bryson et al., 2013; Dalakas et al., 2023), including boycotts, negative word-of-mouth and the emergence of anti-brand communities (Brandão and Popoli, 2022).

We define brand-friendly online communities as virtual spaces where members predominantly express positive brand engagement through supportive discourse, brand advocacy and emotional attachment. Brand-unfriendly communities, conversely, are characterized by negative brand engagement manifested through criticism, opposition and anti-brand activism. This dichotomy in online brand engagement raises fundamental questions about the stability and evolution of consumer brand perceptions over time. While researchers have examined related phenomena such as brand recovery from crises (Greyser, 2009) and negative publicity management (Dawar and Pillutla, 2000), the cognitive-social processes facilitating transitions from anti-brand to pro-brand orientations remain insufficiently theorized, leading to our research question:

Through a qualitative investigation involving in-depth interviews with online brand community members in the UK, we illustrate six pathways consumers use to alter their brand beliefs, transitioning from a negative schema to a positive one.

We theorize this transformation by integrating schema theory (Fiske and Taylor, 1991) with the theory of social representations (Farr and Moscovici, 1984). Our research makes several key contributions. By bridging schema theory and social representations, we introduce a holistic framework, Collective Schema Dynamics (CSD), that captures the interplay of cognitive and social processes in brand schema transformation. This framework helps in examining both the cognitive architecture of brand beliefs and their social construction in community contexts. It highlights the interplay between individual cognitive processes and social dynamics, creating a complex environment for the transformation of brand schemas. Existing theoretical frameworks have not fully addressed this complexity. By integrating these elements, this approach overcomes the limitations of previous models that treated individual cognition and social meanings separately, without considering their interdependence. The qualitative data discloses six distinct pathways through which consumers shift from negative to positive brand schemas in online communities. We also propose an extensive taxonomy of online brand engagement, derived through inductive analysis. Additionally, we highlight the inherently dynamic nature of online brand engagement, advancing theoretical and practical understanding of contemporary branding practices.

In the following sections, we present a brief literature review on brand schema and online brand engagement, followed by a detailed explanation of our qualitative methodology and the techniques used for data analysis. We conclude by presenting a theoretical framework that integrates the theories of schema and social representations, providing a dynamic perspective on brand engagement to understand changes in brand schema.

Brand schema, rooted in schema theory, represents internalized cognitive structures consumers develop to organize and process brand information (Halkias, 2015). These schemas encompass all the perceptions, beliefs, associations and attitudes consumers hold about a brand, shaped by their experiences, interactions and exposure to brand stimuli (Puligadda et al., 2012). Various dimensions of brand schema have been explored, including its formation, content, activation and influence on consumer responses and brand-related outcomes (Davvetas and Diamantopoulos, 2016; Meyers-Levy and Tybout, 1989; Sujan and Bettman, 1989). These dimensions operate within a complex ecosystem of individual and contextual factors that shape how consumers mentally represent brands. Individual factors such as past experiences, attitudes, personality traits and cognitive processes, as well as contextual factors such as marketing communications, brand experiences, social interactions and cultural norms also impact brand schema formation and evolution (Brakus et al., 2009; Puligadda et al., 2012). It has also been shown that brand schema transformation can occur through dynamic interactions between different marketplace actors. For instance, Giesler (2012) shows in his analysis of Botox’s evolution, the meanings of branded innovations emerge through progressive sequences of contestations between brand images promoted by innovators and competing “doppelgänger” images promoted by other stakeholders. The latter are “disparaging images and stories about a brand circulated in popular culture by a loosely organized network of consumers, antibrand activists, bloggers, and opinion leaders” (Thompson et al., 2006, p. 50) but they can significantly influence how consumers perceive and relate to brands. When market creation is understood as a social process of legitimation and doppelgänger brand images signal an innovation’s perceived incongruence with prevailing social norms, the evolution of brand schema may be viewed as a chain of brand image battles through which concrete exchange structures between producers and consumers are established.

Contemporary digital platforms have redefined brand image dynamics, with features such as hashtags and shares driving the negotiation of meaning (Kozinets et al., 2021). Brand schemas are now shaped by networked consumer interactions rather than linear exchanges, as consumer-generated content rapidly evolves narratives through viral and algorithmic amplification (Vallaster and von Wallpach, 2013). Online communities play a key role in these brand image contestations, driven by consumer engagement through interactive experiences and community interactions (Kamboj and Rahman, 2017; Evans et al., 2001). The co-creation of brand meanings has shifted from firm-consumer dyads to networked value creation involving multiple stakeholders (Kennedy and Guzmán, 2016). In digital spaces, consumers actively co-create brand identities, motivated by social factors, entertainment value, brand compatibility, communication appeal and brand commitment. This co-creation emerges through reciprocal exchanges, where firms relinquish control and consumers gain agency in shaping brand schemas (Kennedy and Guzmán, 2016). This process reinforces the idea of brand schemas as collectively negotiated rather than unilaterally imposed. The shift from static to dynamic conceptualizations of brand schema becomes particularly evident when examining varying levels of consumer engagement in digital spaces.

Understanding consumers’ varying engagement levels is important for understanding how brand schemas transform in digital spaces. An overall positive engagement fosters brand love, defined as “an intensive emotional connection characterized by intimacy, passion and commitment” (Palusuk et al., 2019, p. 116), resulting in brand loyalty and resistance to negative brand events (Albert and Merunka, 2013; Hollebeek et al., 2014). Conversely, negative engagement can lead to anti-brand behaviors such as avoidance (Ruppel and Einwiller, 2021), switching (Roy et al., 2022) or joining anti-brand communities to share their negative experiences (Casey et al., 2023; Rossolatos, 2021). This negative engagement can lead to brand hate, defined as “the most intense and consequential negative emotion consumers may feel toward brands” (Zarantonello et al., 2016, p. 11). Brand hate manifests through emotions such as anger and contempt (active) as well as fear, disappointment and shame (passive). Existing research highlights various consumer concerns, encompassing both operational issues, such as product quality and ethical production, and strategic challenges, including brand counterfeiting and globalization (Krishnamurthy and Kucuk, 2009; Sharma et al., 2022). These issues often result in consumer–brand disidentification (Dessart et al., 2020; Rodrigues et al., 2021).

However, a critical gap remains in understanding whether dissatisfied consumers can alter their brand schemas when confronted with incongruent new information or a reevaluation of existing knowledge. This issue is particularly relevant in today’s digital landscape, which has fundamentally transformed how consumers access brand information and develop relationships with brands (Swaminathan et al., 2020). Online communities, in particular, facilitate the rapid creation of emotional brand connections through networked affect (Arvidsson and Caliandro, 2016). Additionally, platform architectures significantly shape how consumers encounter, interpret and share brand meanings (Kozinets et al., 2010). These technological mediations introduce new dynamics in the formation and evolution of brand schemas within digital spaces. When consumers are exposed to brand messages that challenge their established schemas, they are likely to engage in heightened cognitive elaboration to resolve this incongruity (Maier and Mafael, 2024). This process could lead to adjustments in their schemas; a phenomenon that remains relatively underexplored.

From a theoretical perspective, the existing literature uses service-dominant logic (Merz et al., 2009), actor–network theory (ANT; Giesler, 2012) and social influence theory (SIT; Carlson et al., 2008) to understand the evolution and negotiation of brand meanings. For instance, service-dominant logic provides a lens for understanding brand meaning co-creation, positioning brands as dynamic social processes where value emerges through complex networks of resource-integrating actors (Merz et al., 2009). Similarly, ANT explains how brand meanings are negotiated and transformed through interactions between human and non-human actors in marketing networks, suggesting that brand evolution occurs through the alignment and realignment of diverse stakeholder interests (Giesler, 2012). Similarly, SIT explains how individual brand perceptions are shaped and modified through social processes and collective meaning-making (Carlson et al., 2008). It emphasizes how social others can affect consumer emotions, opinions and behaviors through both overt and subtle mechanisms (Dahl, 2013).

Each theoretical framework offers a distinct lens through which to examine brand meaning construction, though each carries specific limitations when applied to dynamic online contexts. For instance, service-dominant logic, while acknowledging the co-creative nature of brand meanings, falls short in explaining the cognitive mechanisms through which individual actors process and internalize these co-created meanings (Allen et al., 2008). The framework proposed by Merz et al. (2009) emphasizes resource integration and value creation but does not sufficiently explain how mental structures mediate the interpretation of brand-related experiences. This theoretical blind spot leaves questions about individual meaning-making processes unanswered. Similarly, ANT presents another theoretical limitation. Though Giesler’s (2012) application of ANT explains the negotiation of brand meanings through network interactions, it neglects the role of preexisting mental frameworks in shaping how network actors interpret and respond to these interactions. The theory’s focus on network dynamics overlooks the cognitive foundations (Couldry, 2008) that guide individual actors’ engagement with and understanding of brand-related information. This limitation manifests empirically in studies applying ANT to brand controversies, where researchers often document network reconfigurations without explaining why individual consumers interpret identical brand actions differently based on preexisting cognitive frameworks (Giesler, 2012).

SIT encounters similar constraints. While Carlson et al. (2008) explain how social processes shape brand perceptions, their framework does not adequately address the interplay between individual cognitive structures and collective meaning systems. SIT’s emphasis on social influence mechanisms leaves unexplored the question of how personal mental schemas interact with and are modified by shared social representations (Marková, 2016). An integrated theoretical approach is needed to capture both cognitive and social dimensions of brand schema transformation in contemporary digital environments.

Integrating schema theory (Stein, 1992) with social representations theory (Moscovici, 2001) bridges theoretical gaps by linking individual cognitive processes with collective meaning-making. Schema theory explains how mental frameworks organize and interpret brand-related information, while social representations theory highlights how these individual schemas interact with and are shaped by shared social understanding. Together, they offer a dynamic explanation of how brand meanings evolve through the interplay of personal and collective meaning systems. This integration addresses key shortcomings in existing theoretical approaches while providing a holistic framework for understanding brand meaning evolution.

Rooted in cognitive psychology, schema theory argues that individuals construct mental structures (or schemas) to interpret and understand the world (Stein, 1992). These schemas influence perception, interpretation and memory, shaping how new information is processed and how individuals engage with their environment (Fiske and Taylor, 1991). Despite its utility, schema theory has been critiqued for its reductionist approach, static portrayal of cognitive structures, overemphasis on internal processes and limited real-world applicability (Augoustinos et al., 2014; Lodge et al., 1991). In digital spaces, however, the evolution of brand schemas is influenced by discursive interactions in online communities, where distinct engagement styles – guiding, following, adapting and venturing – shape the negotiation and transformation of brand meanings (Lima and Pessôa, 2023).

While schema theory focuses on individual cognitive processes, social representations theory adopts a broader perspective by linking individual psychology with collective social processes (Augoustinos et al., 2014). Integrating these theories provides a holistic understanding of how brand schemas are transformed in online brand communities. This synthesis addresses criticisms of schema theory’s individualistic focus and highlights the importance of social dynamics in constructing brand meanings. It is important to recognize that critics of integrated theoretical approaches caution against the risk of theoretical ambiguity and diminished explanatory power. Traditional cognitive psychologists argue for the independence of mental processes from social influences to maintain analytical precision (Anderson, 2020). However, such criticisms overlook the inherently social nature of brand-meaning construction, especially in the interactive and collaborative environments of digital communities. The empirical investigation that follows attempts to capture this theoretical integration in practice by examining how consumers navigate and reconstruct brand meanings across different online communities.

To understand changes in brand schema, we conducted an in-depth exploratory study. Data was collected over six months in 2023 from online brand community members in the UK. We obtained ethical approval for this project and followed all ethical guidelines for this research. The research design employed a structured approach to community identification, classification and participant recruitment to capture diverse perspectives on brand engagement. To identify relevant brand communities, we used a systematic and replicable approach using Google as the primary search tool, given its widespread use in online research (Hillis et al., 2012). However, recognizing the biases inherent in search engine algorithms, such as the privileging of popular or search-engine-optimized sites (Mager et al., 2023), we took deliberate steps to mitigate these limitations. For instance, we consulted with brand community experts to identify additional relevant communities not captured through Google. We also pre-defined criteria to ensure that only communities explicitly focused on brand-related discussions and demonstrating recent activity were included in the data. We achieved this by implementing a detailed search protocol (see Siddaway et al., 2019) that utilized two matrices of search terms. The primary matrix included terms such as brand community, brand forum, brand fan group and brand discussion, whereas the secondary matrix featured terms like brand reviews, brand complaints, brand enthusiasts and brand club. To make sure we had relevant data, we focused on posts from the preceding six months.

Furthermore, we operationalized an “active” community based on the following criteria:

• a minimum of three new discussion threads per week;

• at least ten participant responses per thread; and

• a member base comprising at least 100 registered users.

For each search query, we reviewed the first two pages of results, which reflect widely accepted practices in online research to balance comprehensiveness and relevance (Mendelson, 2007). We identified 20 brand communities through this process hosted on platforms such as Reddit, Facebook Groups and Discord. To confirm our sample was representative, we cross-verified these communities using publicly available information. Following community identification, a systematic classification protocol was developed to categorize these spaces according to their predominant brand sentiment and engagement patterns.

Once the relevant communities were identified, we moved to the classification process. The classification protocol follows a three-stage analytical framework grounded in digital ethnographic traditions. The first stage involves systematic content selection, where we examined a carefully selected sample consisting of the 20 most recent discussion threads in each community. For each thread, analysis encompasses the initial post and the subsequent 10 responses, allowing us to capture both conversation initiation patterns and community response dynamics. Moving into the second stage, we then coded sentiment markers throughout these interactions, building a textured understanding of community attitudes. We performed sentiment analysis by using the natural language processing (NLP) capabilities of the NLTK library in Python (version 3.11) (Lee, 2024). The application of NLTK for sentiment analysis is novel but follows established methodological precedents for textual analysis in consumer research and digital marketing (Alzate et al., 2022; Hartmann et al., 2019; Ordenes et al., 2017). Some comparative studies have validated lexicon-based approaches such as those offered by NLTK against human coding (van Atteveldt et al., 2021). It provides a methodical framework for sentiment analysis, empowering researchers to categorize and quantify subjective information embedded within textual data through its pre-trained models, lexical resources and customizable algorithms (Lee, 2024).

The identified markers manifested in various forms of brand engagement. Positive markers materialized through explicit brand advocacy, detailed product recommendations and instances where community members actively defended the brand against criticism. Additional positive indicators included constructive feedback mechanisms and the sharing of positive experiences. Conversely, negative sentiment surfaced through sustained criticism, organized boycott discussions and systematic promotion of alternative brands. Communities exhibiting high volumes of complaint aggregation or warning posts about products and services also signal negative sentiment patterns. We evaluated the presence and frequency of specific sentiment indicators.

In the third stage, we implemented a precise classification rubric. We recognize that the binary categorization into brand-friendly and unfriendly communities may oversimplify the complex spectrum of brand sentiment expressed in online communities. However, this parsimonious reduction to two categories functioned as a theoretical abstraction mechanism. We followed Eisenhardt’s (1989) argument that simplification can help researchers identify patterns that might otherwise remain obscured by excessive granularity. This approach parallels similar methodological decisions in organizational research, where complex phenomena are initially reduced to deductive fundamental categories to establish theoretical foundations (Spiggle, 1994). The parsimony of this classification system offered us a specific advantage for theory generation. It created clear boundaries for comparing and contrasting community characteristics, allowing us to identify distinctive patterns in community evolution and brand relationships. The binary categorization allowed us to examine broad patterns in community evolution while maintaining analytical clarity.

Communities earned a brand-friendly designation when positive sentiment markers exceeded 60% of coded interactions, and when moderation policies actively supported constructive dialogue. The presence of brand evangelists and consistent brand defense by community members further reinforced this classification. In contrast, brand-unfriendly classification applied to communities where negative sentiment markers dominated more than 60% of interactions. These communities often featured organized criticism campaigns, methodical documentation of problems and systematic promotion of alternative brands (see Figure 1 for a graphical illustration). The threshold of 60% was established following iterative content analysis phases and intercoder reliability testing (Cohen’s κ = 0.83), achieving optimal discrimination between community types while minimizing classification ambiguity. The protocol required meticulous documentation through standardized recording procedures (Saldana, 2021). Each analyzed community generated a detailed profile capturing the URL, community name, membership statistics and activity metrics. We documented sentiment analysis results, final classification decisions and any notable exceptions. Every profile included the analysis date and researcher identification, establishing a clear audit trail (Miles et al., 2020). To maintain methodological rigor, regular calibration meetings addressed edge cases and maintained consistency in classification decisions.

Permission to engage with participants was obtained from moderators or website administrators before collecting data. Invitations for online interviews were then posted in these communities, and additional participants were recruited through snowball sampling (Dosek, 2021). We phrased the purpose of our study broadly to elicit detailed descriptions of members’ experiences of brand engagement (Roulston and Choi, 2018). To minimize retrospective sensemaking and social desirability biases, we emphasized the voluntary nature of the interviews, assured anonymity, avoided priming for socially acceptable responses and encouraged concrete examples (Bergen and Labonté, 2020). This systematic approach helped us to collect rich qualitative data from a variety of online brand communities. The interview protocol was developed in consultation with seasoned researchers in consumer behavior. After pre-testing and refining interview questions, we conducted 25 in-depth, semi-structured qualitative interviews with members from both brand-friendly and unfriendly communities (refer to Table 1 for participant details). We purposefully attempted to enlist and prioritize participants who switched from brand-unfriendly to friendly communities. The brands discussed in the interviews span multiple product categories, from FMCG to luxury goods (see Table 1). This heterogeneity is intentional and methodologically sound, as our research examines fundamental processes of schema transformation rather than category-specific engagement patterns. Following Eisenhardt’s (1989) guidelines for theory building, diverse cases can uncover robust theoretical patterns. The emergence of consistent transformation pathways across different product categories strengthens the generalizability of our findings while acknowledging that engagement intensity may vary by category.

The interviews lasted 45–90 min (average 55 min) and were conducted via Zoom conference call feature with video. The length of each interview varied depending on the number of brands discussed. Previous studies suggest that participants are more inclined to provide honest responses online, especially when addressing sensitive topics (Thunberg and Arnell, 2022). The semi-structured format offered an open yet focused dialogue, encouraging discussion. We took detailed notes during the interviews and then transcribed them for analysis.

Our data analysis followed a systematic, inductive approach to concept development and theory building (Gioia et al., 2013). The analysis proceeded through two phases, each with distinct analytical objectives. Phase 1 focused on developing a consumer taxonomy based on brand relationships and online community engagement patterns. We conducted first-order analysis, staying close to informant terms and experiences. Exemplar first-order quotes include “hypocrisy is the only constant” (P3), “they won’t listen to customers like us anyway” (P5), “their products are good but I don’t talk about it online” (P24). Based on the Gioia method, we derived second-order themes and aggregate dimensions capturing participants’ descriptions of their engagement behaviors, emotional connections to brands and interaction patterns in online communities. The resulting data structure of consumer taxonomy is shown in Table 2. To preserve participant anonymity, we masked specific brand names and community identifiers in our analysis.

Having established a taxonomic foundation for understanding consumer engagement patterns, the analysis then examined the specific pathways through which brand schema transformation occurs by using the Gioia method. We began with line-by-line open coding of forgiveness narratives using in vivo codes to maintain fidelity to informant language. Examples include expressions such as “fundamental change in their culture” (P4), “maybe I overreacted” (P7) and “those decisions were necessary” (P11). Through analytical bracketing (Giesler and Thompson, 2016), we identified discrete units of consumer justification for brand forgiveness. To establish analytical trustworthiness, both authors independently coded three interviews and resolved coding discrepancies through detailed discussion. Following the first-order analysis, we engaged in second-order theoretical coding to identify emerging themes and aggregate theoretical dimensions. We consolidated similar first-order quotes into broader second-order themes – for instance, “leadership apology”, “cultural evolution signals” and “rational acceptance”. This process generated 18 themes. We examined the relationship between these themes by using axial coding and identified six aggregate theoretical dimensions underpinning brand forgiveness (see Table 3).

The analysis involved iterative cycling between data, codes and emerging theoretical categories. Regular author meetings helped in the discussion of evolving patterns and refinement of the coding framework. We documented our analytical decisions and theoretical observations in detailed memos. Rather than imposing predetermined categories, we allowed the theoretical structure to emerge inductively from empirical patterns in the data. This inductive approach yielded both a structured taxonomy of engagement forms and a process model of schema transformation that together provided a complex picture of brand–consumer relationships in online contexts.

Our analysis showed two basic dimensions pertaining to online brand communities:

1. membership in a brand-friendly or unfriendly community; and

2. active or passive level of participation in the community.

Content analysis identified that there are six distinct groups of consumers in the interview data, which helped us develop a thematic taxonomy (see Table 2). Content analysis revealed distinct patterns in how consumers engage with brands online, forming identifiable categories based on their relationship orientation and participation intensity. Consumers were classified into the following categories: critical activists, neutral netizens, active evangelists, passive irates, silent spectators and disconnected loyalists. This approach aligns with research showing that online community members can be categorized based on their engagement patterns and brand relationships, ranging from highly active participants to passive observers (Ozuem et al., 2021).

Critical activists actively participate in anti-brand activities, expressing intense negative emotions, as illustrated by the following quotes:

These expressions of brand opposition demonstrate how individual cognitive schemas about brands become reinforced and amplified through collective social representations in online communities. The intense emotional language and shared narratives serve as mechanisms for constructing and maintaining negative brand schemas through social validation.

In contrast, passive irates agree with anti-brand activities ideologically but do not actively participate, as illustrated by these quotes:

Silent spectators are members of online brand communities but do not participate at all, as one participant commented:

On the other hand, neutral netizens participate in online communities but maintain a balanced view of branding, which is apparent in this quote:

This account shows how individual brand schemas can evolve through exposure to contradictory evidence, leading to modifications in collective social representations in online communities. The participant’s schema transformation from negative to balanced occurred through direct experience but was mediated by social sharing.

Active evangelists are passionate about certain brands and speak positively about them online, as evident in these quotes:

The evangelists’ narratives show how positive brand schemas are co-constructed and reinforced through social sharing in online communities. Their enthusiastic brand advocacy stems from both individual cognitive processing of brand experiences and the collective amplification of positive brand meanings.

Finally, disconnected loyalists join firm-hosted brand communities but do not actively participate, as one participant noted:

Figure 2 visualizes the taxonomy of online brand engagement. This taxonomy aligns with the social-cognitive perspective where individual brand schemas are shaped through collective interaction patterns in online communities, suggesting that brand perceptions emerge from both personal cognitive processing and social meaning-making (Fiske and Taylor, 1991; Moscovici, 2001).

This classification system extends established consumer culture frameworks, with “critical activists” exhibiting characteristics analogous to the ideologically-motivated opposition documented by Hollenbeck and Zinkhan (2010), while “active evangelists” parallel the brand missionaries identified in consumer tribe literature (Cova and Cova, 2002). The taxonomy thus integrates previously disparate categorizations while adding fine-grained differentiation for digital contexts.

The interviews disclosed a trend where some consumers transitioned from critical activists to neutral netizens or active evangelists, forgiving brands for strategic or operational transgressions and adopting a balanced or positive perspective. The transformation process from negative to positive brand schema operates through multiple mechanisms that involve both cognitive reappraisal and social reinforcement. This dynamic is evident in the six distinct pathways that emerged from our analysis (see Table 3).

Consumers perceive brands acknowledging wrongdoing and initiating transformative changes in business practices, leading to psychological and attitudinal shifts in brand perception, as illustrated by these quotes:

The participants mentioned above represent critical activists who transformed into active evangelists during their brand engagement journey. However, for these participants, the efficacy of brand repentance is evaluated by the attainment of a better purpose, manifested through tangible change in marketing, customer service and advertising practices and other relevant areas. The restorative brand actions pathway captures how tangible actions and structured communication by the brand to address wrongdoings can start schema change by restoring consumer trust and perceptions. This process illustrates a core tenet of our theory – the transformation of individual brand schemas through socially negotiated interpretations of brand actions. When brands implement concrete changes, it provides material for collective reinterpretation of brand meanings in online communities. While tangible brand actions represent one pathway to schema transformation, consumers may also respond to less concrete but symbolically significant changes in brand positioning.

Consumers perceive an ideological change in brand policy or discourse, sufficient for brand forgiveness despite a lack of practical manifestation, as one participant expressed:

The ideological brand shifts pathway reflects how even perceived ideological or rhetorical changes by the brand, without much concrete action, can trigger schema evolution among some consumers.

Consumers reevaluate the severity of the transgression as minor, leading to a change in brand beliefs, as evidenced by this quote:

Transgression trivialization aptly describes the pathway where consumers reevaluate the severity of the brand’s transgression as trivial, leading to schema adjustment.

Consumers shift from dispositional attribution (placing responsibility on the brand) to situational attribution (assigning responsibility to external circumstances), as one participant noted:

Circumstantial reframing encapsulates the shift from viewing the brand’s actions through a dispositional lens to considering the situational circumstances, reframing schema accordingly.

Consumers adopt a pragmatic perspective, contending that societal or consumer beliefs from the brand outweigh the harm caused by past actions, as evident in these quotes:

The quotes above provide evidence that continuous engagement with a brand can lead a community member to weigh the potential benefits against the potential harm. If the decision tips in favor of the good, it has the capacity to trigger a change in beliefs toward brand forgiveness. Hence, this pathway captures the utilitarian weighing of a brand’s societal/personal benefits against its perceived transgressions when transforming schema.

Consumers perceive that the brand has faced sufficient punishment for its transgressions, warranting forgiveness and reconciliation, as one participant stated:

Perceived justice pathway summarizes how consumers may forgive a brand and evolve their schema once they feel the brand has faced sufficient punishment or consequences for its actions.

These six pathways show varying temporal dynamics in brand schema transformation. While some pathways (e.g. restorative brand actions) can facilitate relatively rapid cognitive restructuring when brands make concrete amends, others (e.g. ideological brand shifts, circumstantial reframing) represent more gradual processes requiring extended periods of reinterpretation and negotiation. The data suggests that pathway effectiveness varies according to transgression type. Product-related failures responded most positively to restorative actions, whereas value-based transgressions (e.g. perceived corporate hypocrisy) required ideological shifts or benefit-harm recalculations to facilitate forgiveness. This variation indicates the importance of matching recovery strategies to specific types of brand relationship ruptures. These pathways rarely operate in isolation; our data indicates sequential progression in many cases, with “transgression trivialization” and “circumstantial reframing” typically preceding “benefit-harm calculus” as consumers gradually reconceptualize their brand relationship. The pathways also show interactive effects, where partial forgiveness through one mechanism facilitates subsequent schema transformation through complementary processes.

Our analysis illustrates the bidirectional interplay between cognitive and social mechanisms in brand schema transformation. In the cognitive-to-social direction, individual schema modifications often precede community-level changes. For instance, in the “circumstantial reframing” pathway, consumers first cognitively reattribute brand actions from dispositional to situational factors before sharing these reinterpretations with community members (“gradually I realized, and some of the chat members agreed with me”). Similarly, the “transgression trivialization” pathway begins with personal cognitive reassessment of transgression severity, which may subsequently reshape community discourse. Conversely, the social-to-cognitive direction manifests when collective representations influence individual schema reconstruction. The “perceived justice” pathway reveals how socially validated assessments of brand punishment (“I told everyone on the chat group”) can modify individual schema. Likewise, the “benefit-harm calculus” pathway often originates in community discussions that prompt individual utilitarian evaluations (“I was following the online discussion board […] what I think about it is […]”). These directional patterns enrich the explanatory power of the CSD framework by showing the recursive processes through which brand schemas evolve in networked environments.

Through this qualitative investigation, we sought to understand the dynamics underlying the transformation of brand schema over time. Our results illustrate the interplay between individual cognition and social influences in shaping brand schema transformation. For instance, the “critical activists” show how individual brand schema can be collectively reinforced in certain online communities through processes such as descriptive coding. Conversely, the “neutral netizens” provide an example of how social interactions in these communities can reshape individual brand perceptions, leading to balanced views. This integration highlights brand schema as a dynamic, co-constructed phenomenon (Siano et al., 2022) shaped by both individual cognitive processes and social negotiation in online brand communities. We also observed what Giesler (2012) describes as “brand image contestations” – progressive sequences where competing meanings promoted by different stakeholders shape how brands are perceived and legitimized. Each pathway represents a distinct process through which consumers resolve contradictions between their existing negative schema and emerging positive brand meanings.

Similarly, the “restorative brand actions” pathway shows how brands’ tangible actions to address wrongdoings can initiate changes in consumers’ individual brand schemas. However, these actions are often interpreted and evaluated through the lens of the larger online community. The quotes for this pathway show how the initial negative brand schema was collectively reinforced in the anti-brand community, but the brand’s subsequent actions led to a shift in the individual’s cognitive representation of the brand. Similarly, the “benefit-harm calculus” pathway provides an example of how social representations can shape individual brand schemas. The quotes show participant’s positive interpretation of the brand’s messaging diverges from the dominant social representation in the community, leading to a favorable individual brand schema. The pathways identified in this study expand upon earlier work examining how consumers collectively negotiate brand meanings in oppositional contexts (Hollenbeck and Zinkhan, 2010). While anti-brand communities indicate how negative brand schemas can be reinforced through social interaction, our findings show how similar social processes can drive the evolution from negative to positive brand schemas. Moreover, our work extends research on brand hate (Zarantonello et al., 2016) by demonstrating how negative brand emotions can evolve into neutral or positive states through structured pathways. While previous work has mapped the components of brand hate, we show the mechanisms through which such negative emotional states can be transformed.

These examples illustrate the complex interplay between individual cognition and social influences in shaping brand schema transformation. By acknowledging both the cognitive and social dimensions of this phenomenon, we can develop a refined understanding of consumer-brand interactions. Consequently, we developed a new theory, CSD theory, that encapsulates these elements.

By synthesizing schema theory (Stein, 1992) with social representations theory (Moscovici, 2001), CSD explains how individual cognitive structures interact with collective meaning-making processes in digital spaces (see Figure 3 for a graphical illustration). Through interactions, discussions and shared experiences in these spaces, consumers co-create (Siano et al., 2022) and renegotiate the social representations of brands, shaping and reshaping individual brand schemas in the process (Sloan et al., 2015). The six transformation pathways identified through our analysis show specific mechanisms through which this theoretical integration manifests empirically, building upon prior work on social brand engagement (Hollebeek et al., 2014) and consumer–brand relationships (Fournier, 1998). These mechanisms also extend previous frameworks of brand community dynamics (Muniz and O’Guinn, 2001).

CSD theory has significant implications for theoretical development and practical brand management. It offers a comprehensive framework for understanding the factors shaping consumer perceptions and brand-related decisions. This framework integrates brand schema theory (Puligadda et al., 2012) and social representation theory (Moscovici, 2001), addressing the traditional separation of individual and collective levels in branding literature. By bridging these perspectives, it shows how brand knowledge is constructed, negotiated and utilized. Individual brand schemas are influenced by dominant social representations in consumers’ online networks. As shared meanings, beliefs and values are internalized, they shape perceptions and responses to brand stimuli. However, this influence is bidirectional: individual schemas, shaped by personal experiences, can challenge and transform existing social representations. Through interaction and discourse, these modified schemas may influence collective representations in online communities.

This dynamic interplay helps explain how brand knowledge evolves at both individual and community levels. Empirical findings map pathways of schema-collective interaction, offering a structured approach to analyzing brand meaning co-creation (Vallaster and von Wallpach, 2013). The CSD theory explains processes through which individual cognitive structures interact with social representations to transform brand perceptions. Unlike existing models of brand perception that emphasize stability, CSD foregrounds the inherently dynamic nature of brand schema, particularly in digitally mediated environments. Beyond schema transformation, CSD provides explanatory mechanisms for additional branding phenomena including the formation of brand communities around contested meanings, the evolution of consumer–brand relationships through digital touchpoints and the fragmentation of brand interpretations across market segments – confirming its broader theoretical utility. This theoretical contribution moves beyond simplified sender-receiver models of brand communication to recognize the complex interplay between individual interpretation and collective meaning negotiation in digital spaces. The theory suggests that transformation pathways operate through both cognitive mechanisms (e.g. reappraisal, recategorization) and social processes (e.g. community validation, narrative reconstruction). These dual mechanisms explain how consumers reconcile contradictory brand information by integrating individual experiences with collectively constructed brand narratives. This integration also highlights the role of power dynamics, social influence and communication in shaping brand schemas and social representations. Dominant groups often reinforce certain representations (Denegri-Knott et al., 2006), embedding them in consumer schemas. Conversely, marginalized groups may create counter-representations (Harju and Huovinen, 2015), driving schema modifications and collective change over time.

Our findings disclose the dynamic and evolving nature of online brand schemas, challenging the traditional view of their static stability (Peracchio and Tybout, 1996). Building on research highlighting how brand meanings are negotiated in anti-brand communities (Hollenbeck and Zinkhan, 2010), we show that similar social learning mechanisms can drive positive brand schema evolution. In online communities, this evolution occurs through discursive interactions, where content transitions from virtual to actualized states as members collectively engage. This temporal aspect of online discourse helps reframe past negative brand experiences through new interpretative lenses (Lima and Pessôa, 2023). It manifest distinctly across transformation pathways: “restorative brand actions” and “perceived justice” often catalyze rapid schema reconfiguration, while “ideological brand shifts” and “benefit-harm calculus” typically involve gradual evolution of interpretative frameworks over extended periods, reflecting different temporal characteristics of cognitive-social processes.

We extend Giesler’s (2012) work on brand-mediated legitimation by demonstrating that brand schema transformation involves individual cognitive processes and collective meaning-making. While Giesler examined brand management of competing images over time, we explore the psychological and social mechanisms by which consumers reconcile opposing brand beliefs. Six pathways illustrate the roles of multiplicity, brand engagement experiences and temporality in shaping brand schema transformation. Multiplicity captures the diverse and interconnected pathways through which brand engagement occurs, reflecting the complexity of consumer-brand interactions (Giesler and Thompson, 2016; Holt, 2002). Brand engagement experiences, such as “perceived justice,” showcase how recognizing the consequences of brand hate can shift perceptions, while “restorative brand actions” highlight how tangible brand efforts to address wrongdoing prompt schema changes in a community context (Muniz and O’Guinn, 2001).

Temporality underpins this dynamism, as brand narratives help participants connect past experiences with present interpretations, encouraging reassessments and positive reframing (Lima and Pessôa, 2023). For instance, the “ideological brand shifts” pathway reflects how consumers interpret ideological changes favorably despite limited practical outcomes, while the “benefit-harm calculus” pathway explains how past actions justify present branding efforts. This temporal integration of past experiences and future expectations show the evolving nature of brand schemas (Brakus et al., 2009). Beyond documenting schema transformation processes, our research also contributes an empirically grounded classification system for understanding consumer engagement in online brand communities.

This taxonomy, derived inductively from interviews, provides a realistic and holistic understanding of consumer–brand interactions compared to deductive and a priori conceptual typologies (Bailey, 1994). This taxonomy, grounded in empirical data, can better predict and capture the heterogeneity of consumer beliefs and engagement forms in online brand communities. This is in line with previous research that points to the complexity of consumer–brand engagement in a virtual brand community (Lima et al., 2019). This taxonomy builds upon Vallaster and von Wallpach’s (2013) work on multi-stakeholder brand meaning co-creation by showing how different engagement styles shape collective brand interpretations. While Vallaster and von Wallpach focus on stakeholder interactions, our taxonomy indicates how individual cognitive schemas evolve through these collective meaning-making processes. Similar to how brand hate can manifest through both active and passive components (Zarantonello et al., 2016), our taxonomy indicates that brand schema transformation occurs through both active engagement (critical activists, active evangelists) and passive participation (silent spectators, disconnected loyalists).

Drawing from the study’s theoretical and empirical findings, brand managers should actively monitor and participate in both brand-friendly and brand-unfriendly online communities. The six identified pathways of brand schema transformation provide managers with specific strategies for addressing negative brand perceptions. For example, when faced with brand criticism, managers can implement “restorative brand actions” through tangible changes in business practices, followed by clear communication of these changes to the community. For implementing “ideological brand shifts,” managers should articulate revised brand values through multiple communication channels, use respected third-party endorsements to establish credibility and display authentic commitment through consistent messaging across all brand touchpoints. This approach acknowledges that abstract ideological repositioning requires sustained symbolic reinforcement to transform consumer perceptions. Brand managers should also recognize that consumer transitions from negative to positive brand perceptions often occur through incremental steps rather than immediate shifts. This gradual nature of schema transformation suggests that recovery strategies should be patient and persistent, allowing time for consumers to process and internalize positive brand changes.

The taxonomy of online brand engagement offers brand managers a framework for segmenting and targeting different consumer groups in virtual communities. Managers can develop tailored engagement strategies for each identified segment – from active evangelists who require support and reinforcement in their brand advocacy, to critical activists who need careful attention and strategic response to their concerns. The study establishes that even highly negative brand relationships can transform positively through appropriate management intervention. Brand managers should therefore establish systematic monitoring mechanisms for tracking schema evolution in online communities, allowing them to identify early signs of both negative sentiment formation and opportunities for positive transformation. This approach requires dedicated resources for community management and real-time response capabilities. Ultimately, effective management of brand schema in digital environments requires recognition of both cognitive and social dimensions of consumer brand relationships as outlined in our proposed theoretical framework.

This study contributes to a better understanding of branding by highlighting the dynamic and evolving nature of online brand engagement and brand schema transformation. Nonetheless, we need to acknowledge some limitations. One key issue is the absence of qualitative longitudinal data, which prevents establishing causal sequences in the observed changes to brand schema. While our primary aim was to identify pathways for schema transformation rather than to determine their temporal order, future research could address this gap by using longitudinal methods to explore the causal mechanisms underlying these changes. Adopting Giesler's (2012) longitudinal approach, for instance, could provide valuable knowledge of how transformation pathways unfold over extended periods. This will help us uncover additional dynamics in the evolution of brand schemas. Mixed-method approaches that combine digital ethnography with traditional longitudinal techniques could further improve this exploration. For example, researchers could integrate digital trace data analysis with periodic in-depth interviews to track brand schema evolution across various touchpoints and timeframes.

Another limitation arises from the reliance on Google-derived data. Search algorithms inherently introduce systematic biases (Mager et al., 2023) by prioritizing certain types of content based on proprietary ranking systems. These algorithmic biases can distort the representation of brand perceptions and engagement patterns. Other researchers should triangulate Google-derived data with alternative data sources such as social listening platforms, ethnographic observations and traditional survey methods to construct a complete picture of brand perceptions while accounting for algorithmic mediation of digital traces. Finally, the study’s heterogeneous sample composition, while offering breadth, introduces methodological complexities. Future studies could address this by designing industry-specific investigations that account for sectoral variations while preserving sufficient sample diversity to generate robust theory. Such focused approaches would allow granular analyses of brand schema transformation in specific market contexts. For instance, researchers might investigate if luxury brands show distinct patterns of schema transformation compared to convenience goods. Despite these limitations, our theoretical integration of schema theory and social representations provides a foundation for future research on brand meaning co-creation in increasingly complex digital environments.

As brands navigate increasingly complex digital environments, understanding the malleability of brand schemas becomes essential. Our research suggests that even deeply entrenched negative brand perceptions can transform through structured pathways of reconciliation and reframing. This finding carries significant implications for brand management strategies, especially in crisis situations or when addressing consumer opposition. Brand meaning evolution in digital spaces represents a rich arena for future inquiry. As the latest technology, such as generative AI, continues reshaping consumer-brand interactions (Cheng and Jiang, 2022), examining how platform architectures, algorithmic mediation and emerging forms of digital engagement influence brand schema transformation will become increasingly important.

Funding: CPCE research fund, College of professional and continuing education, Hong Kong Polytechnic University, Hong Kong, China.