Channel-switching behaviour and customer deviance Open Access

In recent years, more than 10 billion USD of financial losses have been attributed to customer fraud (PWC, 2022). Since then, negatively deviating customers have been on the rise (Mayr and Teller, 2023) and are damaging retailers. Not only that, such behaviours threaten bystander-customers' wellbeing (Grybś-Kabocik, 2016) and their shopping experience (Fombelle et al., 2020), driving them out of stores. Not surprisingly, bystander-customers switch services due to toxic clients (Cai et al., 2018).

Despite this, current research corroborates their effects on employees (Mayr and Teller, 2023) but leaves the impact on bystander-customers mostly aside. As a phenomenon prevalent in the service industry, research addresses customer-to-customer interactions (Nicholls, 2010) and contagion effects limited to customer misbehaviour (Schaefers et al., 2016; Srivastava et al., 2022). This is concerning, as negative customer deviance (NCD), a behaviour which describes subtle forms of social norm infringements, does not only unfold within the customer-customer dyad. Moreover, it goes beyond self-centred misbehaviour and its contagion effects on bystander-customers within the service sector.

While research effectively addresses the need for NCD research, it leaves unexplained how bystander-customers are affected by NCD in retail stores, and how retail-channel-switching intentions are formed as a consequence. This is even more concerning as, since the recent crisis, consumers have been switching retail channels, in the form of both brand switching and across-brand switching at an unprecedented rate (Charm et al., 2020). Moreover, within the vast availability of retail channels, retailers face increasing pressure in terms of profitability and return on capital (Catena et al., 2023). Thus, to diminish the financial risk posed by NCD, retailers need to understand how NCD affects bystander-customers and specifically its impact on retail-channel- switching.

Accordingly, we aim to reveal NCD effects on bystander-customers in stores and to examine its correspondence with the formation of switching intentions from store-based to online shopping. Therefore, an NCD effect chain in store-based retailing, from the view of bystander-customers, is revealed.

This is one of the first studies to examine the effects of NCD in retailing in correspondence with store-based to online retail-channel-switching. The findings indicate that NCD affects retail-channel-switching intentions directly and indirectly, as this effect precedes mediation in the form of shopper ill-being, shopping experience dissatisfaction, affective decommitment, negative attitude, and the influence of moderating variables. Based on that, strategies for retailers are suggested to help diminish and control NCD-induced threats to bystander-customers.

We contribute to current research by (1) examining multiple dimensions of bystander-customers’ NCD perceptions in retail stores, (2) explaining NCD’s impact on those customers and their retail-channel-switching intentions, and (3) providing practical implications for retail management.

Customer deviance comprises negative and positive forms of customer behaviour according to normative theory (Mertens et al., 2016). While NCD refers to badly behaving customers (Mayr and Teller, 2023), positive deviances (Mayr and Teller, 2024) entail pro-social (Kim and Kim, 2024) and customer citizenship behaviour (Le et al., 2024). However, we focus on NCD due to its detrimental impact on retailers and the risk it poses to abandon bystander-customers from stores.

As a research field which is dominated by psychological and service marketing research, only a few studies investigate NCD within retailing (Dootson et al., 2023), testing the severity of one specific NCD dimension, opportunistic deviance, in relation to service robots. Current research effectively captures various forms of NCD relating to abusive behaviour, verbal aggression, and fraud (DeCelles et al., 2019; Okan and Elmadag, 2020), customer misbehaviour (Chaouali et al., 2022), dysfunctional behaviour (Huang et al., 2022), disruptive behaviour (Cai et al., 2018), and customer incivility (Hur et al., 2022). While each of these terms addresses one aspect of NCD in a specific context, such as for example incivility or misbehaviour, entailing product or property damages, broken policies, impatience, and frustration (Srivastava et al., 2022), a whole range of other forms of NCD specifically directed at bystander-customers (e.g. harassment or arguments with other customers) and within the context of store-based retailing are scarcely tackled within the literature presently. While NCD entails norm violations targeting bystander-customers, frontline employees and store assets, other forms of negative customers behaviour captured within research, such as misbehaviour (Chaouali et al., 2022), is characterised by an inappropriate behaviour which concerns mainly customers’ self-centredness. Since NCD is not exclusively and intentionally associated with egocentrism, such behaviours are rather subtle. We therefore expand prior research by addressing multiple subtle forms of NCD and their effects.

Overall, within retailing, NCD research corroborates frontline employees’ perspectives and NCD’s effects on them (Mayr and Teller, 2023) but rarely discusses bystander-customers’ NCD encounters (e.g. Fisk et al., 2010) or their reactions to them. The effects of forms of NCD on bystander-customers draw attention to services marketing research as it confirms that switching services can occur as a consequence of encountering toxic clients in restaurants (Cai et al., 2018). The effects of NCD on bystander-customers within the context of retailing and their switching behaviour are unspecified, as within retailing, research links switching behaviour either to channel attributes (Obeng et al., 2016) or to customers' characteristics (Van Nguyen et al., 2022), leaving it unclear how multiple forms of NCD in retail stores impact bystander-customers' emotions and their attitudes.

To close these research gaps, we test NCD empirically, taking the viewpoints of bystander-customers within retail stores, linking multiple NCD dimensions to bystander-customers' affective states, attitudes, and retail-channel-switching intentions, which is crucial for developing strategies to mitigate NCD’s negative outcomes.

As a psychological behavioural theory, reactance theory (Brehm and Brehm, 1981) explains the emotional and behavioural responses of people in case their freedoms or choices are restricted. In the case of NCD, bystander-customers are restricted in their way to act when they encounter such behaviours while shopping in stores. Accordingly, psychological reactance arises as the freedom to act autonomously in a desired way is threatened (Steindl et al., 2015). Consequently, individuals who feel threatened grasp at restoring this lost freedom (Matarazzo and Diamantopoulos, 2022) and thus, retail-channel-switching intentions are likely to be formed if bystander-customers encounter NCD while shopping in stores. Reactance theory suggests that it takes the form of cognitive and affective responses, translating into affective, attitudinal, and behavioural actions in response to a threat (Dillard and Shen, 2005), such as NCD for bystander-customers. Within this notion, reactance is affective when a person reacts in the form of feelings, attitudinal when they form a specific attitude, and behavioural if their behaviour is adjusted due to a threat. The theory elucidates the emotional and attitudinal processes that arise in response to perceived threats, including the sequence of effects leading to the restoration of an original state. In this context, we utilize this theory to explore how bystander-customers perceive NCD as a threat to their freedom to enjoy a positive shopping experience. This perceived threat triggers reactance in the form of reduced commitment and intentions to switch retail channels as bystander-customers seek to regain their sense of autonomy and restore their shopping satisfaction.

NCD, characterised by negatively socially deviating customer behaviours within retail settings (Mayr and Teller, 2023), is proposed to significantly undermine the retail experience. Previous research supports that unethical customer behaviour fosters negative attitudes towards the shopping experience, potentially leading to the avoidance of shopping in stores as a matter of customer expectation disconfirmation (Mayr et al., 2022). Given that shopping experience satisfaction is crucial for customers' contentment with their stores and future return intentions (Klaus and Maklan, 2013), we suggest that NCD encounters adversely affect the shopping experience. Not only that bystander-customers’ shopping experience is diminished, but they also encounter emotional distress and ill-being as a result of witnessing aggressive or unethical conduct (Grybś-Kabocik, 2016; Okan and Elmadag, 2020). As exit behaviours and service provider switching is prevalent among bystander-customers witnessing customer rage (Dorsey et al., 2016; Cai et al., 2018), similar effects concerning NCD and retail-channel-switching are expected. Hence, we infer:

Shopper ill-being corresponds with disharmonious consumption experience and is mainly examined within the hospitality sector (Abbes et al., 2019). Customers who engage in store based shopping typically anticipate a positive experience, encompassing cognitive, affective, social, and physical responses (Bustamante and Rubio, 2017). In alignment with reactance theory (Brehm and Brehm, 1981), NCD disrupts enjoyable experiences in stores for bystander-customers, thereby triggering ill-being and a negative impact on their shopping experience satisfaction and affective commitment. As a critical factor which denotes a favourable emotional attachment to a store (Dean, 2007) and fosters long term loyalty to a shopping channel (Maggioni et al., 2019), affective commitment and the effect of ill-being on it is crucial for retailers. Research in the field of marketing suggests that ill-being has a negative impact on long-term relationships (Sirgy and Lee, 2008), as it compels customers to diminish their allegiance to the channel in question. Thus, we propose:

Research underlines the pivotal role of shopping experience satisfaction in determining customers' overall satisfaction with their in-store shopping encounters (Klaus and Maklan, 2013) and suggests that the quality of customers' experiences is significantly linked to affective commitment to a brand (Khan et al., 2020). It implies an interplay of shopping experience satisfaction and customers’ emotional investments in the retailer.

Furthermore, the attitude towards the affective experience at the store (Chun et al., 2017), that is the level of enjoyment, fun, and overall satisfaction derived from a shopping experience, is influenced by customers’ critical cognitive evaluations of their experiences (Anshu et al., 2022) and they manifest in the development of a loyal attitude towards the retailer (Lucia-Palacios et al., 2018). This underscores the importance of customer satisfaction in nurturing long-term affective commitment. Therefore, we propose:

Emotions, motivations, and attitudes serve as precursors to the decisions customers make and ultimately lead to behavioural outcomes (Schiffman et al., 2013), which is important for understanding retail-channel-switching behaviours. In retail environments, emotions and individual traits significantly influence customers' decision-making processes positively (Martinelli et al., 2021). Research has shown that attitudinal factors (beliefs, behavioural tendencies) and affective processing play a critical role for shaping consumers’ decision-making process (Puccinelli et al., 2009) and they influence customer loyalty positively through affective commitment (Evanschitzky et al., 2006). Based on that, we propose:

Building upon the findings of Van Nguyen et al. (2022), who identified social influence, customer characteristics, and risk perception as key drivers of retail-channel-switching intentions, we argue that the emotional and psychological factors that shape customers’ attitudes also play a crucial role in the formation of switching intentions. Additionally, studies display, that such behavioural intentions are positively influenced by individuals' attitudes (Martinelli et al., 2021) which implies a correspondence between attitudinal factors concerning the shopping experience and customers' intentions to switch shopping channels. Therefore, we conclude:

According to previous research, which examines emotional and attitudinal dynamics in retail contexts, customers’ emotions, satisfaction and behavioural intentions to repatronise are impacted by store characteristics and in-store interactions (Terblanche, 2018). Negative behavioural responses correspond with angry customers (Bougie et al., 2003) and omnichannel retail-channel- switching is influenced by social groups’ behaviour (Van Nguyen et al., 2022).

Similarly, and in alignment with our proposed research model and reactance theory (Brehm and Brehm, 1981), NCD in stores represents a threat, provoking an affective response (shopper ill-being), triggers an attitude formation (shopping experience dissatisfaction), translates into an affective reactance (affective decommitment), engenders an attitudinal reactance (negative attitude towards the affective experience), and results in a behavioural reactance (intention to switch from the store-based retailer to online shopping). Thus, we assume:

As reactance is impacted by (1) personal and (2) social factors (Miron and Brehm, 2006), we navigate through a series of moderators (see Figure 1):

• (1)

  Customer purchase engagement

Customer purchase engagement encapsulates the multifaceted dimensions of customers' conscious attention, their enthused participation, and their social connection while shopping in stores (Vivek et al., 2009). Building on findings from organizational research, prior research has demonstrated the moderating role of stakeholders' engagement in the dynamics between relationship of negative psychological climate in one’s surroundings (comparable to NCD) and individual wellbeing (Shuck and Reio, 2014), with the impact being stronger for those who are engaged. In parallel, within digital services, research anticipates stronger effects between service satisfaction and customers' service usage continuance, as a sign of commitment, among those who are more engaged (Thakur, 2019). Despite the notable contextual divergences between these studies and our own setting, we expect customer purchase engagement to moderate affective and attitudinal effects in response to NCD in retail stores. Thus, we expect:

Trust in the retailer refers to customers' willingness to rely on a retailer in which they have confidence (Moorman et al., 1993). Prior research indicates that trust plays a moderating role in various retail contexts. For instance, trust has been found to moderate, in the context of NCD, the effects of employees’ ill-being on their job satisfaction (Mayr and Teller, 2023). Within online shopping, research reveals that trust moderates the impact of social influence on online shopper behaviour, including commitment (Davis et al., 2021). This suggests that customers who have a high level of trust in a retailer are likely to be more positively engaged and less susceptible to the effects of negative influences on their shopping behaviour. Given these insights, it is reasonable to expect a moderating influence of trust between NCD and ill-being as well as satisfaction and commitment. Thus, we propose:

Customers (n = 1,008) at least 16 years represent the sample for the web-based survey used in this research, obtained through representative quota sampling in Austria, a typical western European retail market in terms of high store intensity and an average online retail penetration. We selected in-store and online shoppers as only those who use both channels are eligible for switching. Among the customers, 50% were females, 49.9% males and 0.01% diverse, and the majority was between 25 and 64 years old, with each age group making up 18–20% of the total sample. More than 50% were in employment, 9% were self-employed, and 23% were retired. The data were collected with the help of an online panel provider which we have used for previous studies. The measurements were retrieved from academic literature (see Table 1).

As the study design consisted of web-based surveys with self-administration, common method variance (CMV) (Tehseen et al., 2017) could have been a problem. In order to avoid CMV, certain remedies were imposed. Procedural remedies: The web-based survey was designed to avoid any biases due to proximity effects (Weijters et al., 2009) or item wording effects (Harris and Bladen, 1994). Therefore, the order of the measurement items was counterbalanced (Tehseen et al., 2017). To control the effects of item wording, item characteristics, item ambiguity, and double-barrelled questions (MacKenzie and Podsakoff, 2012), only validated measurements were used (Podsakoff et al., 2003). Additionally, we included an attention check and pre-tested the survey. Statistical remedies: In terms of statistical remedies used to uncover any biases, we tested all construct correlations by applying correlation matrix procedures (Tehseen et al., 2017). We used an unrelated marker variable (Podsakoff et al., 2003) and we conducted Harman's one-factor analysis (Podsakoff and Organ, 1986). As the correlation scores for all variables were <0.90 (Bagozzi and 1988), and as partialling out the marker variable (Tehseen et al., 2017) produced scores below 0.3 (Lindell and Whitney, 2001), we deduced that CMV had been detected. Harman's one-factor test indicated that a single factor explained 25.9%. We therefore assume that this research is unbiased.

Structural equation modelling: To test the research model, partial least squares structural equation modelling (PLS-SEM) was used as it is suitable for predicting and comparing complex relationships (Hair et al., 2020). This component-based approach conceptualises latent variables as weighted composites (Cho and Choi, 2020) and tests the strengths and directions of path correlations based on statistical significance (Tenenhaus et al., 2005). Measurement model assessment: Following methodological recommendations (Anderson and Gerbing, 1988; Hair et al., 2020), we tested the convergent and divergent validity before assessing the measurement model. Confirmatory composite analysis (Hair et al., 2020) indicates highly significant (p < 0.001) factor loadings higher than 0.70 (Hulland, 1999). Reliability testing illustrates excellent results, with all latent constructs higher than 0.70 (Bland and Altman, 1997). Composite reliability (CR) scores coincide with the Fornell Larcker criterion (Fornell and Larcker, 1981) as all values are above 0.70. Further, the average variance extracted (AVE) demonstrates values higher than 0.50 (Bagozzi and Yi, 1988) (see Table 1). Following this, we assessed the discriminant validity by comparing the squared AVE values with the correlations (Fornell and Larcker, 1981) (see appendix) and by performing the heterotrait monotrait test (HTMT) (Ab Hamid et al., 2017). The HTMT displays satisfactory discriminant validity as all values are below 0.90 (see Table 2).

The direct effects reveal that all hypotheses are supported by the SEM, following Chin (1998) (see Figure 2). The results confirm the impact of NCD on shopping experience satisfaction (H1a) (p < 0.001; t = 5.592) and on shoppers' sense of ill-being (H1b) (p < 0.05; t = 2.872), and between NCD and bystander-customers' intentions to switch to online retail channels (H1c) (p < 0.001; t = 7.507). Shopper ill-being directly affects the shopping experience satisfaction of bystander-customers (H2a) (p < 0.001; t = 3.556) and their affective commitment (H2b) (p < 0.001; t = 4.247). The strongest effects are found between shopping experience satisfaction and affective commitment (H3a) (p < 0.001; t = 27.853) and for that between shopping experience satisfaction and bystander-customers' attitude towards the experience in the store (H3b) (p < 0.001; t = 22.981). A strong impact is evident between affective commitment and attitude towards the affective experience (H4) (p < 0.001; t = 8.327). The findings reveal that the attitude towards the experience affects retail-channel-switching intentions directly (H5) (p < 0.001; t = 5.852).

We confirm the affective-attitudinal response-reactance chain between bystander-customers' NCD perceptions and their intentions to switch from retail stores to online shopping as a result of shoppers' ill-being, their shopping experience dissatisfaction, their affective decommitment, and their negative attitude towards the affective experience (H6) (p < 0.05; t = 2.001). This effect is fully mediated even without the presence of affective commitment and attitude towards the affective experience, as shopping experience satisfaction has a direct effect on retail-channel-switching intentions, which indicates the importance of attitudinal components for switching intentions.

To establish the moderation effects, we applied multi-group analysis (MGA) (Henseler, 2012). Moderation effects were deduced if the p-value of the path coefficients within the group comparison was smaller than 0.05 or larger than 0.95 (Hair et al., 2020). Before employing MGA, we demonstrated metric invariance following the steps outlined by Putnick and Bornstein (2016). Accordingly, we ran multiple PLS-SEM analyses to assess group differences in the measurement model. As no differences were revealed, metric invariance was concluded.

The findings reveal significant results for all moderation effects of customer purchase engagement. The MGA displays that with decreased purchase engagement, the impact of NCD on shopper ill-being is stronger (p = −0.380**; CEG_low, 0.297; CEG_high, −0.083) (H7a). This indicates that purchase engagement surprisingly impacts negative emotional states positively. Additionally, with increased bystander-customer’s engagement, the effect of shopping experience satisfaction on their commitment strengthens (p = −0.118*; CEG_low, 0.528; CEG_high, 0.646) (H7b). This suggests a correspondence between customers’ purchase engagement and their satisfaction and emotional bonding with the retailer, with the latter being more pronounced when engagement levels are higher.

We confirm all the moderation effects of trust in the retailer. With high trust in the retailer, the effect of NCD on bystander-customers' shopper ill-being is equally high (p = −0.425*; TIR_low, −0.284; TIR_high, 0.143) (H8a), whereas with low trust, this effect is reversed, which means that shopper ill-being is not affected by NCD if bystander-customers’ trust in the retailers is low, which is rather counterintuitive.

Affective commitment is also stronger when there is trust, supporting the moderation by trust in the retailer of the effect of shopping experience satisfaction on affective commitment (p = −0.284*; TIR_low, 0.313; TIR_high, 0.597) (H8b). The findings underline the importance of trust in the retailer when it comes to NCD's negative effects on affective and attitudinal dimensions, displaying stronger effects for those customers who highly trust retailers.

Our research contributes theoretically to reactance theory (Brehm and Brehm, 1981) as we display the application of this theory in the context of NCD. Our findings reveal that bystander-customers exposed to NCD (Fombelle et al., 2020; Mayr et al., 2022) while shopping in stores correspond with their switching retail channel behaviour (Jebarajakirthy et al., 2021) by shopping online instead, which is crucial for store based retailers. We expand on channel-switching research by highlighting its correspondence to non-cognitive dimensions (NCD perceptions) and emotional factors. This focus contrasts with existing studies that mainly emphasise attitudes (Van Nguyen et al., 2022) or best-deal-seeking behaviour (Chiu et al., 2011). Thus, we fill a critical research gap in current research and we provide a more comprehensive understanding of the complexities of channel switching behaviour.

We significantly advance research on customer deviance by shedding light on the impact of different NCD dimensions on bystander-customers – findings that have been overlooked in previous research (Fombelle et al., 2020; Mayr and Teller, 2023). We expand this research area as we test NCD effects in a retail context contrary to prior research which centres around service marketing. By uncovering multiple NCD dimensions beyond customer misbehaviour (Chaouali et al., 2022), abusive behaviour, verbal aggression, and fraud (Okan and Elmadag, 2020; Zhang et al., 2023), we successfully broaden the view of negative customer behaviour. Furthermore, we demonstrate the flip side of the narrative of positive deviances (Mayr and Teller, 2024) as outlined in normative theory (Mertens et al., 2016). Our findings display NCD effects beyond mere contagion (Schaefers et al., 2016) as we reveal complex dynamics of NCD and its linkage to retail-channel-switching intentions relating to emotional and attitudinal shifts within bystander-customers. Contrary to existing research (Dorsey et al., 2016; Nicholls, 2010), these NCD effects not only relate to customer-customer interactions, but also to customer-employee and customer-firm interactions.

In contrast to prior research which corroborates the customer experience evaluation in relation to the product, the delivery of the product, and the perceived convenience (Anshu et al., 2022; Verhoef et al., 2009), we powerfully underscore the importance of affective and attitudinal parameters that shape the evaluation process that drives behavioural intentions. These findings provide novel insights for managerial strategies within the retailer-customer dyad.

Our findings indicate the necessity of combatting NCD through proactive actions before it occurs. Frontline employees as boundary spanners within the customer-retailer dyad, need to be equipped with skills to monitor and manage bystander-customers’ dissatisfaction and empower them to find suitable and prompt solutions for these customers. Therefore, retailers need to train and support employees to detect NCD, and to handle such incidents effectively by responding immediately, by means of a monitoring approach, and eventually with the help of surveillance technology. Further, retailers should consider a compensation scheme for bystander-customers (e.g. in the form of goodwill gestures), which would contain the negative outcomes for highly affected bystander-customers.

As affective commitment is key to retail-channel-switching, retailers are advised to foster a strong emotional bonding with their customers, for example by imposing loyalty programs that not only reward repeat purchases but also encourage customer feedback and interaction with the retailer. Such programs can enhance commitment and mitigate negative feelings associated with shopping in stores due to NCD encounters.

As wellbeing prompts the whole NCD response-reactance chain, retailers must prioritise a sense of wellbeing during the shopping experience, by monitoring the store atmosphere, making sure frontline employees are caring and friendly, and fostering proactive assistance in order to build strong ties with their customers and counteract the attitudinal reactance. Retailers must therefore invest in creating a positive and engaging shopping environment and they should reinforce desirable behaviours by actively showcasing examples of expected behaviours.

Considering moderating influences, retailers should keep customers engaged and foster their trust through encouraging feedback, personalised interactions, follow ups after purchases, and recovery measures such as taking responsibility, apologising to and empathising with bystander-customers, offering solutions, and setting boundaries through retail policies and by documenting NCD incidents. Retailers could also introduce participatory marketing strategies that encourage customer involvement by fostering in-store events and hands-on demonstrations to keep their customers’ engaged.

As NCD directly affects bystander-customers’ wellbeing, satisfaction, and switching intentions, retailers are advised to penalise severely badly behaving customers in the form of a warning, temporary suspension, or the termination of services in the case of persistent, repeat offenders, as a matter of retail policy. As a strategic actionable step, we recommend the establishment of retail policies against NCD, technological surveillance to recognise NCD and enable fast counteractions, and omni-, multi-, or cross-channel solutions that offer shopping alternatives to highly affected bystander-customers.

As does all research, this research has limitations. The research analyses NCD encounters in retail stores, limiting its findings to offline store-based retailing services. As NCD is also likely to appear in online retailing, future research should address NCD effects in online shopping environments. Moreover, as bystander-customers encountering NCD in a specific store may switch to another store and not necessarily to online shopping, future research could target switching between physical stores. Lastly, this research employs variance-based SEM to test the effects. Future research should test the phenomenon using an experimental design, allowing for causational assumptions.

We thank Prof. Sabine Benoit for reviewing this article beforehand and for providing suggestions to enhance this research.

Table A1 

Table A2