Metaverse customer experience: conditional mediation of avatar self-congruity, user similarity and well-becoming Open Access

The metaverse is an immersive and massively scaled simulated space where users can communicate and interact simultaneously with one another via their digital avatars (Hadi et al., 2024). By transcending the boundaries between material and virtual worlds, the metaverse has revolutionised the virtual landscape, leading to profound changes in how users identify and interact (Koohang et al., 2023). This may significantly enhance customer experience (Rahman et al., 2025a), and enable brands to target consumers with novel forms of advertising, gamification and activation (Mehrotra et al., 2024).

By 2030, 40% of people globally are forecast to be using the metaverse for work, entertainment, socialising and retail purposes. The metaverse is forecast to grow annually by 37.73%, reaching US$507.8bn in volume and US$1,081bn in revenue by 2030 (Statista, 2024). This rapid growth is driven by advances in blockchain technologies, artificial intelligence and virtual and augmented realities (Huynh-The et al., 2023), enabling greater synchronicity and interconnection between users (Dieck et al., 2023) and the creation of avatars as digital self-representations (Kim et al., 2023).

Through their avatars, metaverse users can reimagine themselves by altering and integrating selected parts of the self (Etienne et al., 2023) or choosing an entirely new set of physical and personality attributes (Roy and Putatunda, 2023). An individual’s digital avatar can influence their real‐life attitudes and behaviours, a phenomenon known as the Proteus effect (Yee et al., 2009). The ability to construct a unique identity also facilitates customers’ self-expression, socialisation, immersion, enjoyment and sense of control, boosting their metaverse usage and purchase intentions (Domina et al., 2012; Park and Kim, 2024). However, altering one’s identity excessively can disrupt one’s self-concept, leading to identity disjunction and self-discrepancy (XinYing et al., 2024; Yang et al., 2024) and compromising a consumer’s sociological and psychological well-being (Ud Din and Almogren, 2023). This raises important questions about the role of user–avatar self-congruity – the degree to which a user’s avatar reflects their actual or ideal self – in shaping consumer well-being (CWB) within the metaverse. Given that self-congruity can enhance engagement and emotional connection, its implications for well-being warrant deeper exploration. In addition, the extent to which perceived similarity with other metaverse users influences this relationship remains an open question, as excessive similarity may either reinforce or undermine the benefits of self-congruity. Despite research on the varied effects of avatars, self-construction in the metaverse remains underexplored (Kim et al., 2024).

Various factors have a positive or negative influence on CWB. For example, Plangger and Campbell (2022) discovered that misleading elements in the metaverse – such as deceptive advertising, fabricated social interactions or unrealistic portrayals of identity and lifestyle – may adversely affect CWB by raising unrealistic expectations. Conversely, immersive and interactive features that promote self-expression and social connection, such as avatar customisation and virtual communities, can enhance well-being by strengthening a sense of belonging and reinforcing identity (Dwivedi et al., 2023).

Well-being is an umbrella term and has been used to denote: the global consumer’s static state of mind (CWB); the static state of the consumer as a result of having a product (well-having); the real-time state of a consumer involved in activities such as consumption (well-doing); and the dynamic state or imaginary status of the consumer (well-becoming) (Eshaghi et al., 2023). Unlike hedonic well-being, which focuses on a short-term static state of happiness (Xu and Zhang, 2021), consumer well-becoming (CWBec) represents a dynamic, ongoing process of achieving and recreating well-being through various life experiences (Eshaghi et al., 2023; Schomburgk and Hoffmann, 2023). This distinction is crucial in metaverse studies, as CWBec follows a temporal perspective and captures the continuous and evolving nature of user interactions in a virtual environment and their long-term impact on life domains such as family, social life, work and finances. Therefore, focusing on CWBec assists in understanding how the metaverse can augment or detract from users’ quality of life over time, making it a more relevant measure than overall well-being in this context.

As a digital community, the metaverse embodies a sense of co-presence (the “we mode”) by stimulating empathy and synchronicity (Riva et al., 2024). A positive experience leads to feelings of pleasure (Suh, 2024), belonging (Kim et al., 2024) and immersion (Dudley et al., 2023), and it can reinforce customer satisfaction, metaverse usage intentions and purchase intentions (Chen and Yang, 2022; Rahman et al., 2025a; Roy et al., 2017; Siqueira et al., 2019). However, researchers are yet to examine whether customer experience in the metaverse (CXMV) influences CWBec or whether user–avatar self-congruity has a direct or indirect impact on CWBec. Given that CXMV represents users’ engagement with and perception of the metaverse, understanding its mediating role is key to clarifying the broader well-being effects of avatar self-congruity and user similarity. Therefore, this study addresses the following contemporary research questions:

To address these questions, we draw on stimulus–organism–response (S–O–R) theory to explore the effect of environmental stimuli on the individual’s cognitive, emotional and behavioural state (Bagozzi, 1986; Mehrabian and Russell, 1974). We propose a new empirical framework that explains the influence of metaverse self-congruity directly on CXMV and indirectly on CWBec and metaverse usage intentions, and whether this is shaped by users’ perceptions of other avatars. Empirically, we apply advanced partial least square (PLS)-based conditional mediation analysis to precisely test the hypotheses (Cheah et al., 2021).

This study advances the literature with four main contributions. First, we extend the marketing literature by developing a model of the antecedents and outcomes of CXMV. Second, we examine the indirect effect of self-congruity on CWBec and metaverse usage intentions via CXMV, offering strategic insights for metaverse platform developers. Third, we demonstrate the moderating effect of users’ perceptions of other avatars, providing insights for managers seeking to enhance customer immersion and engagement. Fourth, we operationalise new measurement scales for key constructs in the metaverse context, including self-congruity, metaverse user similarity and CWBec.

The paper is structured as follows. First, we define the key constructs of our conceptual framework: self-congruity, CXMV, CWBec, metaverse usage intentions and metaverse user similarity. Next, we develop our hypotheses and introduce the methodology and the data collection processes we used. We then present the quantitative results and post hoc qualitative study. Finally, we discuss the insights from this study and explain the implications for theory and practice. We also offer future research directions.

Understanding the factors that shape customer experience and well-becoming in digital environments requires a strong theoretical foundation. This section reviews existing literature on CWBec, self-congruity and metaverse engagement, providing the basis for our conceptual framework.

Well-being has evolved as a multifaceted concept. Scholars have typically conceptualised well-being through two distinct perspectives: hedonic and eudaimonic well-being (Ryan and Deci, 2001). The hedonic perspective focuses on pleasure and pain avoidance, emphasising subjective happiness and life satisfaction (Disabato et al., 2016). In contrast, the eudaimonic perspective emphasises the pursuit of a meaningful, purposeful and fulfilling life (Ryan and Deci, 2001; Ryff and Singer, 2008). Building on these foundational perspectives, marketing research has expanded CWB to include various interpretations. At its core, CWB reflects consumers’ overall state of mind and satisfaction in life (Diener et al., 1999). It also captures the benefits derived from product ownership and consumption, including both material and symbolic value (Belk, 1988). Furthermore, CWB extends to active engagement in consumption experiences (Rahman et al., 2025b), where consumers co-create value through their interactions with products and services (Anderson et al., 2013). However, evolving virtual technologies have transformed our understanding of CWB from a static state to an ongoing process of identity construction and personal transformation (Batat, 2021).

The emerging perspective of CWBec represents a shift from a static to a dynamic conceptualisation of CWB (Xu and Zhang, 2021). While building upon eudaimonic well-being perspectives, emphasising the dynamic, forward-looking trajectory of personal growth (Ryff and Singer, 2008), CWBec extends this perspective by focusing on the dynamic process through which consumers actively construct and work towards their future selves, particularly through virtual or imaginary experiences (Eshaghi et al., 2023). This future-oriented perspective aligns with research on possible selves theory, which suggests that individuals are motivated by their visions of what they might become through self-transformation (Castilhos and Fonseca, 2016; Markus and Nurius, 1986). Studies have shown how immersive technologies enable consumers to engage in self-transformation, moving beyond mere self-expression (Yee et al., 2009). This ongoing process of technology-mediated transformation is beyond the traditional interpretation of eudaimonic well-being. CWBec is also distinct from hedonic well-being in terms of temporal orientation, as it focuses on the future state of users.

The theoretical foundation of CWBec lies in consumers’ psychological life space, where experiences are stored as “units of meaning” that influence overall life satisfaction (Lee et al., 2002; Xu and Zhang, 2021). The consumers’ self-concept is central to the formation of the psychological world as it reflects beliefs and values related to their actual and ideal selves. These are, in turn, instrumental for self-evaluations within all life domains and in shaping their future self (Sirgy, 2012; Markus and Nurius, 1986). Recent research shows that consumers use short-term well-being to project both back and forward in their real lives (Preece and Skandalis, 2024), suggesting that CWBec is an ongoing process of recreating well-being through life experiences, resulting from interactions between the consumer’s internal and external worlds (Xu and Zhang, 2021). Similarly, our post hoc study shows that metaverse users identified various ways in which metaverse experience facilitates growth, change and striving towards improved future states of life.

Consumers’ future states are becoming increasingly important for both consumers and researchers, as advancements in technology shape identity construction and enable individuals to reimagine and work towards their ideal selves in digital environments (Eshaghi et al., 2023). However, how the metaverse enables and shapes consumers’ future states is underexplored. This is an important research gap as the metaverse provides consumers with the potential to (re)define “identity” and enables consumers to (re)produce the “self” and their perceptions of well-being across life domains (Lee et al., 2002; Sirgy, 2012; Zhu et al., 2024). The metaverse makes this possible by providing an “external” simulated environment for consumers to engage in reciprocal activities in parallel virtual and real-life realms (Filimonau et al., 2024; Xu and Zhang, 2021). Engaging within a metaverse environment affects consumers’ daily personal, societal, financial and work-related lives and may engender both positive and negative impacts on family and social dynamics (Dwivedi et al., 2023; Sun and Yuan, 2023). Importantly, interacting within virtual environments supports learning and can help expand consumers’ problem-solving skills in their actual lives (Cowan and Ketron, 2019), helping them feel a higher quality of life and achieve their desired future state of being.

CWBec, in this research, is positioned as a longer-term, ongoing cognitive appraisal applied to a consumer’s state of being, resulting from an imaginary or simulated environment, such as the metaverse (Eshaghi et al., 2023). Unlike the traditional conceptualisation of subjective well-being that captures current pleasure-based states, CWBec measures the process of personal development through virtual experiences. This view aligns with the Aristotelian concept of eudaimonia while extending it to the technological mediation of human flourishing in digital environments. We, thus, consider CWBec to be an evaluative (subjective) variable representing a consumer’s life domains and related to themselves (CWB self-beneficiary) (see Table Web-appendix W2 for more CWBec definitions). In this research, CWBec captures four interrelated well-being aspects, namely, consumers’ family, social, work and financial life (Dwivedi et al., 2023; Eshaghi et al., 2022). This evolving conceptualisation of well-becoming highlights the need to explore how digital environments, particularly the metaverse, shape customer experiences and interactions. Given the transformative potential of immersive technologies, it is essential to examine their role in shaping consumer perceptions and behaviours.

Social media platforms and immersive technologies have revolutionised customer experience and brand interaction by enabling real-time interactions, which facilitate peer-to-peer recommendations and social connectivity (Constantinides and Fountain, 2008). These platforms have demonstrated how digital environments can enhance user engagement through features like real-time feedback, content sharing and interactive tools (Kaplan and Haenlein, 2010). Virtual reality (VR) applications have further demonstrated the power of immersive experiences in gaming and entertainment contexts (Steuer, 1992). Moreover, augmented reality (AR) has transformed customer experiences by overlaying digital content onto the physical world, enhancing product visualisation and trial experience (Javornik, 2016). Building upon these foundations, the metaverse represents a unique evolution by offering persistent, interconnected virtual spaces where users can express themselves through customisable avatars (Batat, 2022, 2024). The metaverse provides customers with multifaceted sensorial experiences and collaborative interactions with brands, bridging the divide between physical and virtual reality (Habil et al., 2024; Hallikainen et al., 2022) (see Web-appendix Table W1 for more metaverse definitions).

CXMV is a multifaceted concept that represents a customer’s evaluation of their experiences before, during and after their metaverse journey (Mansoor et al., 2024). With the development and advancement of virtual platforms, customer experience has increasingly become a focus point (Kumar et al., 2023). Unlike VR gaming or social media platforms that offer relatively isolated experiences, the metaverse provides a persistent and interconnected virtual environment that remains active even when users are offline (Dolata and Schwabe, 2023; Dwivedi et al., 2022). This characteristic, combined with its accessibility through various devices, from basic computers to advanced VR headsets, enables continuous engagement and identity expression through customisable avatars. The metaverse’s unique combination of immersion, real-time synchronicity and enhanced user control not only differentiates it from other digital platforms but also influences how users experience virtual interactions and perceive their overall quality of life (Mansoor et al., 2024). Given these unique features and their impact on user interactions, CXMV can be conceptualised through six key aspects: affective, cognitive, physical, relational, sensory and symbolic (Gahler et al., 2023).

The affective aspect refers to the moods and emotions generated by interactions with others in the metaverse (Gentile et al., 2007). For example, positive emotions can be triggered by immersive experiences through AR and VR (Talukdar and Yu, 2021), the ability to customise content (Habil et al., 2024) and the experience of flow (Balakrishnan et al., 2024). Suh (2023) argued that users’ emotional reactions to VR experiences are primarily determined by perceived opportunity and threat. The cognitive aspect refers to the thoughts generated during the customer experience (Gentile et al., 2007) and is partly determined by platform responsiveness and personalisation (Deng et al., 2010). The physical aspect reflects customers’ perceptions of their movement (Joy and Sherry, 2003). For example, AR can improve motor skills in people with disabilities (Kang and Chang, 2020), while the metaverse offers users immersive travel experiences (Buhalis et al., 2023). The relational aspect refers to customers’ relationships with other customers and brands (Gentile et al., 2007) and is enhanced by social connections, co-creation and social presence (Hadi et al., 2024; Koohang et al., 2023). The sensory component refers to customers’ visual, auditory and tactile experiences during virtual interactions, facilitating users’ immersion in virtual experiences (Batat, 2024; Brakus et al., 2009). For example, users may “feel” objects using haptic gloves (Masterson, 2022) or engage via auditory channels with other users (Minotti, 2021). Finally, the symbolic aspect refers to self-expression and self-affirmation during the interaction (Gahler et al., 2023). Building on these insights into customer experience in immersive digital spaces, it is crucial to examine how users construct and perceive their identities within the metaverse. In particular, self-congruity plays a key role in shaping users’ interactions and experiences in virtual environments.

In a period of increasingly digitised interaction, self-identification and construction now occur across multiple realms. Consumers create, configure and adapt their self-concepts across a range of digital platforms, resulting in a new exploration of the self. The rise of virtual environments and metaverse worlds allows consumers to trial different avatars that symbolise their potential selves, providing them with an opportunity to manipulate and design their possible selves in a low-risk environment (Taylor and Carlson, 2025). This enables consumers to consider and reflect upon multiple dimensions of their future-self and dynamically craft their self-conceptualisations. Consumers can further ensure their self-concept and its digital representation in the form of a metaverse avatar are aligned (Taylor and Carlson, 2025).

Consumer identity refers to the way consumers define themselves through consumption choices, habits and behaviours and is generally an expression or reflection of the consumer’s self-concept (Hadi et al., 2024). Self-concept theory provides a framework for understanding how individuals construct and express themselves in various contexts (Sirgy, 1986). The theory has been extensively applied in consumer research, where studies have shown how self-concept influences brand preferences and purchase decisions in retail settings (e.g. Japutra et al., 2019). While Sirgy’s work identifies multiple dimensions of self-concept, including actual, ideal, social and ideal social selves, research has demonstrated their varying importance across different consumption contexts. For instance, in luxury consumption, the ideal and the social selves often dominate purchase decisions (He and Mukherjee, 2007), while for everyday consumer goods, the actual self plays a more prominent role (Kressmann et al., 2006). These consumption patterns extend to the metaverse, where consumers, through their metaverse avatar identity, engage with metaverse platforms to access digitally created content or interact with others depending on their need for information, entertainment, self-expression or connection (Mehrotra et al., 2024). These avatars’ anthropomorphic characteristics have been found to support socialness perceptions and induce varying degrees of cognitive, affective and social responses during interactions with others in virtual worlds (Miao et al., 2022).

How users create their metaverse identity is therefore crucial to understanding the role of identity in virtual environments (Lee et al., 2023). Markus and Nurius’ (1986) theory of possible selves offers a parallel foundational framework for understanding the motivations for identity construction within the metaverse. Possible selves are considered to act as future-oriented self-representations encompassing the selves that an individual might become, what they would like to become and what they are afraid of becoming (Markus and Nurius, 1986). These possible selves serve as motivating factors that guide and shape behaviour and assist with self-regulation to increase the likelihood of individuals achieving their self-congruity goals (Markus and Nurius, 1986), thereby advancing our understanding of consumers’ motivations to “become”.

Recent research has applied Markus and Nurius’ (1986) possible selves within virtual worlds. The virtual self is defined as “the technology-mediated or mentally processed self, present or simulated on computers, computer networks, electronic games and any other virtual and digital media environments” (Jin, 2012). Virtual identities in virtual worlds are considered to be highly malleable since users can alter both their embodied and disembodied selves, dramatically altering their avatars’ physical, behavioural, social, emotional and moral attributes (Yoon et al., 2025). Jin (2012) notes that users of virtual worlds engage in a creative incarnation of the virtual self; thus, the concept of identity is essential to understanding self-conceptualisation in virtual worlds. Whilst identity construction in digital environments has been extensively studied in various settings including social media, where users can curate their ideal selves and compare themselves to other aspirational reference points (Taylor and Carlson, 2025), and websites, which offer users VR and AR tools to aid decision making, the metaverse gives users unlimited agency over their visual identities (Miao et al., 2022).

In addition, whilst prior virtual platforms, such as SecondLife, offered users the ability to create unique avatars, these platforms lacked interoperability, limited their integration of AR and VR technology and were primarily confined to gaming applications (Mystakidis, 2022). In contrast, the metaverse enables users to carry their “chosen self” across platforms, providing the opportunity for a seamless transition both between virtual spaces and with real-world applications, including but not limited to specific activities such as gaming (Mansoor et al., 2024). The metaverse, therefore, provides users with a more integrative and expansive realm for identity construction and enhancement by blending advanced avatar customisation with persistent immersive socialisation, gaming, entertainment, working, education and asset trading, including real estate and e-commerce economic integration within platforms, whilst also enabling real-time social presence simultaneously across these contexts (Purdy, 2023). The Harvard Business Review notes that the metaverse has reshaped virtual interactions by offering new immersive forms of collaboration and through the emergence of new digital, synchronous AI-enabled interactions, the acceleration of gamified technologies and the rise of a metaverse economy (Purdy, 2023). Metaverse e-commerce, for example, is already projected to reach US$42.1bn in 2025, with an annual growth rate of 37.96% until 2030, leading to a projected market value of US$210bn by 2030 (Statista, 2025).

Unlike prior virtual platforms, metaverse users can closely mirror, alter or replace their real-world identity, including their physical and emotional attributes, with a specifically crafted avatar (Hadi et al., 2024; D. Y. Kim et al., 2023). This self can have a profound effect on a user’s behaviour, both within and outside of virtual environments, through the concept of ethopoeia (impersonating a character) (Teng et al., 2023; van der Westhuizen, 2018). As prior research has shown, a user’s metaverse identity significantly informs their social engagement and enjoyment (Lee et al., 2023), trust (Teng et al., 2023) and propensity to engage in prosocial behaviours (Hadi et al., 2024). As such, our understanding of how consumers develop and project their sense of self relative to past, comparatively passive forms of virtual platforms is limited, and new research-based perspectives of self-construction are required to remain aligned with the more dynamic technological changes in the market associated with the metaverse (Taylor and Carlson, 2025; Yoon et al., 2025).

To examine metaverse identity creation, this study builds upon the work of Sirgy (1986) and Markus and Nurius (1986) by examining the construction of the self within the context of the metaverse (Kim and Sundar, 2012; Suh et al., 2011). According to Teng et al. (2023), in virtual environments, the self is constructed from both the actual and the ideal self. The actual self is the individual’s perception of themself in the present, while the ideal self is the individual’s perception of their future self (Balakrishnan et al., 2024). According to Balakrishnan et al. (2024), both forms of self are crucial in influencing behavioural actions. The Proteus effect suggests that users’ behaviours and self-perception can be influenced by the characteristics of their digital avatars, creating a dynamic interplay between virtual representation and identity (Yee and Bailenson, 2007). This phenomenon is particularly relevant in the metaverse, where users often combine their actual and ideal selves in their avatars (Belk, 2013; Malär et al., 2011), leading to a deeper integration (i.e. formation of the self-congruity construct) between their virtual and real selves (Kim et al., 2023, p. 2). Statistically, this structure treats actual and ideal self-congruity as distinct, reflectively measured first-order factors/dimensions that jointly contribute to the overall self-congruity construct. This formative specification is appropriate when each subdimension captures a unique but essential aspect of the broader concept (Becker et al., 2023), aligning with the identity co-construction processes observed in metaverse contexts. This psychological alignment between the user and their avatar determines the user’s preferences and immersion in the metaverse and leads to the co-creation of meaningful experiences (Yoo et al., 2023). Therefore, self-congruity is key to a positive CXMV and the success of metaverse platforms, which in turn may influence CWBec. However, empirical inquiries into the role of self-congruity in CXMV are lacking.

This section develops the main and moderating hypotheses, drawing upon the S–O–R paradigm and self-congruity theory.

Our conceptual framework is underpinned by S-O-R theory, which seeks to explain the effect of visual, auditory and tactile stimuli on an individual’s emotional and cognitive state and, thus, their behavioural responses (Bagozzi, 1986; Xu et al., 2024). S–O–R theory is commonly used in consumer psychology research to explain consumer behaviours such as brand loyalty, purchase intentions and brand equity (Kim et al., 2020; Yang et al., 2022). It also explains how the stimuli embedded in immersive technologies influence consumer decisions and behaviours (e.g. Kim et al., 2023; Suh and Prophet, 2018). For example, Xu et al. (2024) used an S–O–R model to examine the effects of AR features on consumers’ perceptions and behaviours. Chong et al. (2023) identified that perceived enjoyment and value co-creation were enhanced where high levels of interactivity, telepresence and social presence were observed, in turn promoting ongoing usage intentions. Moreover, interactivity and co-creation in sensory-rich virtual environments stimulate a sense of presence (Cowan and Ketron, 2019; Yim et al., 2017), enhancing CXMV. Therefore, we conceptualise metaverse self-congruity (a higher-order construct resulting from a user’s actual and ideal self-congruity) as a stimulus for CXMV, which in turn leads to CWBec. This suggests CXMV is the mediator in our conceptual model.

Self-congruity is a higher-order concept comprising the actual and ideal selves (Ahn et al., 2013; Hosany and Martin, 2012). An individual’s perceptions of their actual self and ideal self-combine to form their overall self-concept (Balakrishnan et al., 2024; Lazzari et al., 1978; Malär et al., 2011).

Technological advances have enabled individuals to modify and extend themselves through their digital avatars (Belk, 2016). Users can emancipate themselves from their physical bodies and be whoever they want to be, selecting, modifying and accessorising this representation of the self (Bryant and Akerman, 2009). Hadi et al. (2024) argued that the concept of identity takes on an entirely new meaning in the metaverse. However, according to self-verification theory (Yildiz et al., 2024), users tend to design their avatars as reflections of their actual selves. In this way, they can portray a consistent identity and avoid a discrepancy between their online and offline identities (Hadi et al., 2024; Swann, 1987). Therefore, a metaverse user may achieve self-congruity by developing an avatar that is similar to their actual and ideal selves (Donath, 2002; Yoo et al., 2023).

A consumer’s extended self in the form of an avatar incorporates both their actual and ideal selves in virtual form (Barrera and Shah, 2023; Miao et al., 2022). If a consumer designs an avatar that is similar to themself, their metaverse experience is more immersive and enjoyable Park and Lim, 2023). In the metaverse context, self-congruity can enhance a consumer’s affective, sensory, physical, behavioural and intellectual experiences (Balakrishnan et al., 2024; Kim et al., 2023; Moedeen et al., 2024; Soderlund et al., 2021). Therefore, we propose the following hypothesis:

The metaverse may enhance CWB by offering freedom of expression, supportive interactions, reduced loneliness and increased social self-efficacy (Mansoor et al., 2024; Oh et al., 2023b; Oh et al., 2023a). When individuals socially connect through shared-interest groups or communities, they mutually influence one another’s emotions, behaviours, perceptions and psychological outcomes in a process known as entitativity (Campbell, 1958; Rothaermel and Sugiyama, 2001). This supports a sense of belonging and connectedness, influencing a sense of well-being.

Despite the call for more research on the mediating role of CXMV in the relationship between self-congruity and well-being, studies are limited (Moedeen et al., 2024). Batat (2021) found that AR can improve the affective, behavioural, social, sensory and intellectual dimensions of well-being. VR and AR technologies in tourism can also promote emotional well-being in the real world (Yu et al., 2024), while physical movement in VR environments has been found to affect subjective well-being in reality (McLean et al., 2023). Metaverse experiences can also affect consumers’ emotional states (Jiang et al., 2023), in turn influencing their well-becoming (Fredrickson, 2001). When consumers perceive a similarity between their real-world and virtual selves, this may enhance their sense of authenticity and well-becoming (Pimentel and Kalyanaraman, 2020). Moreover, individuals who create friendly avatars have been found to be more social in real life (Ko and Park, 2021; Yee et al., 2009). Reducing the differences between the actual and ideal self can decrease psychological distress and enhance well-being (Ambika et al., 2023; Schafer et al., 1996). Therefore, we posit:

Previous research has demonstrated that a positive customer experience directly affects customers intentions to continue using a virtual platform (e.g. Deng et al., 2010; Habil et al., 2024; Hsu and Lu, 2004; Huang et al., 2021; Qin et al., 2021). For example, Kumari et al. (2024) found that metaverse usage intentions are enhanced through hedonic gratification (e.g. a visually appealing virtual environment), social gratification (e.g. shared events) and utilitarian gratification (e.g. virtual asset accumulation and ownership). Therefore, we posit the following:

To engage their consumers, brands convey images that evoke a typical metaverse user (Kolańska‐Stronka and Krasa, 2024). Brand/product image is an important determinant of customer experience, well-being and purchase intentions (Gorbaniuk et al., 2021; Kini et al., 2024). If a customer perceives a high similarity between themself and the brand image, their purchase intentions will increase (Sirgy, 1982). This similarity is even more critical in metaverse settings, where vast numbers of users communicate and interact (Oh et al., 2023b; Riva et al., 2024).

Virtual relationships and communities may contribute to users’ social, emotional and psychological needs, helping them to build their sense of self and enhance their social identity (Chih et al., 2017; Riva et al., 2024). Individuals who feel socially supported and included are happier and have higher self-esteem and well-being (Oh et al., 2023b; Oh et al., 2023a). Higgins (2021) terms these community-oriented behaviours the “we mode”, which facilitates social categorisation, identification and comparison. The greater the sense of “we”, the stronger the social cohesion through behavioural and cognitive synchrony, shared attention and intentional attunement (Riva et al., 2024).

However, the implications of perceived metaverse user similarity, which refers to the extent to which an individual perceives their own avatar as resembling the typical metaverse user in terms of appearance, behaviour and identity markers, need to be considered. In metaverse environments, individuals encounter varied representations of users, shaped by platform norms, social interactions and brand cues. As users navigate these spaces, they often form mental images or prototypes of what a typical metaverse user looks and acts like. This prototype may include common appearance features (e.g. avatars with fashionable or gamified styles), behavioural patterns (e.g. exploratory or collaborative behaviour) and identity signals (e.g. social roles, cultural markers). Drawing on Ahn et al. (2013), we conceptualise metaverse user similarity as the extent to which an individual perceives alignment between their own avatar and this imagined archetype across visual, behavioural and symbolic dimensions. This similarity influences users’ engagement, sense of belonging and, ultimately, their well-being in the metaverse (Riva et al., 2024; Sirgy, 1982).

Cognitive dissonance theory suggests that excessive perceived similarity between a user and other metaverse avatars can trigger psychological tension through two mechanisms (Cooper, 2012). First, when users perceive high similarity with others, it may threaten their need for uniqueness and distinctiveness, creating cognitive discomfort between their desire for belongingness and individuality (Brewer, 1991). Second, users may experience dissonance when their crafted avatar confronts numerous similar others, challenging their perception of authentic self-expression (Kim and Sundar, 2012). This cognitive dissonance leads users to modify their behaviours or even withdraw from the environment to reduce psychological tension (Kumar and Shankar, 2024).

The impact of cognitive dissonance on user behaviour in virtual environments may operate through several psychological processes. When users encounter high similarity with others, it may trigger social comparison processes that challenge their sense of uniqueness (Leonardelli et al., 2010). Psychological unease may also arise for consumers if they have dissonant cognitions related to their knowledge, beliefs, attitudes or actions concerning an environment or others around them. This discomfort may lead an individual to reduce the inconsistency by changing their cognition. Users also may respond by reducing their emotional investment and engagement in the metaverse environment (Swann et al., 2009). For example, if a consumer evaluates others in the group as being too similar to themself (negative element), this may prompt them to withdraw, disrupting the “we mode” (Higgins, 2021).

Our qualitative research supports perceived psychological unease resulting from user similarity. As one participant noted: “I think the similarity makes it easy for me to connect with other users and have a good time, but I like to meet people that are different than me too” (28-year-old male, a frequent metaverse user). This quote shows how moderate levels of similarity can facilitate connection, while suggesting that too much similarity might reduce the enriching diversity of social interactions. Another participant observed that “having the structure of the metaverse keeps people similar in ways we don’t have in real society” (54-year-old male, a frequent metaverse user), highlighting how the metaverse environment itself may amplify similarity effects.

Therefore, cognitive dissonance arising from perceived user similarity may trigger social comparisons and conflict feelings, leading them to cease interacting or emotionally connecting with others. Reduced we-mode cognition lowers the possibility of immersing-together, becoming well-together and intending-together and, as a result, may dampen the positive relationships proposed in this study. Therefore, we posit the following:

Figure 1 illustrates the proposed conceptual framework.

We used Qualtrics to develop our survey. In total, we used 22 items to measure our constructs. Specifically, we modified six items from Japutra et al. (2019) to measure our two first-order constructs: actual and ideal self-congruity. We used items from Mansoor et al. (2024) to measure CXMV and metaverse usage intentions. Four items from Eshaghi et al. (2022) were modified to measure CWBec, while three items from Ahn et al. (2013) were modified to measure metaverse user similarity. All items were based on a seven-point Likert scale (agree to disagree). The survey was written in English and checked for content validity by three marketing researchers with extensive publication experience. To check face validity, five respondents with metaverse experience completed the survey. We then modified any items flagged as difficult to understand by the participants and marketing researchers. Finally, we distributed 21 online surveys as a pre-test.

To ensure sufficient power and generalisability of our structural equation model (SEM) (Hair et al., 2019b), we used an online calculator (Soper, 2025) to assess the minimum sample size. A sample size of 123 was needed to generate a moderate effect size (d = 0.5). Next, we conducted a power analysis using G*Power, which showed minimum sample sizes of 144, 63 and 43 for small (0.2), moderate (0.5) and large (0.8) effect sizes, respectively, at 95% statistical power and a 10% significance level (Faul et al., 2007). Given the nature of the population, the available resources and the research context, a sample size of 155 was deemed sufficient for the PLS-SEM model (Hair et al., 2022; Hair et al., 2019a) to obtain minimum path coefficients of 0.11–0.2 at a 0.5 significance level. Our statistical results (i.e. confidence intervals, effect sizes, fit indices and parameter estimates) confirm the adequacy of the sample size (Hair et al., 2022; Jobst et al., 2023).

We collected survey data from metaverse users aged 18 years or older in Australia over a three-week period in early 2024. To ensure that they were regular rather than one-off metaverse users, respondents were asked about their favourite metaverse platforms (e.g. Sandbox, Roblox, Zepeto), frequency of usage and duration of engagement before being presented with the main items. To manage common method bias, survey items were randomised and respondents were given sufficient time to complete the survey. The survey finished with some demographic questions. In total, 169 surveys were returned. One was excluded because of possible straightlining concerns (Hair et al., 2022), leaving 168 surveys for analysis. The data set shows a balanced gender distribution, with most participants aged 25–34 years. The dominant family status is single family, and the majority have a bachelor’s degree. Full-time employment is the most common employment status. In terms of metaverse usage, most respondents reported using it for 7–12 months. (see Table Web-appendix W3 for detailed demographic information.)

The data set comprises multiple variables measured by indicators, with observed values falling within the defined 1–7 scale range, and no data is missing, ensuring completeness for analysis. The mean and median values across indicators are closely aligned, such as a mean of 3.21 and a median of 3.0 for ASLF1, showing consistent response patterns (Hair et al., 2022). The standard deviations range from 1.35 (ISLF1) to 1.62 (ASLF2), reflecting sufficient variability in responses. Assessments of excess kurtosis, such as −0.80 for ASLF1 and −0.18 for ISLF1, and skewness values, such as 0.20 for ASLF1 and 0.46 for ISLF1, show that responses are balanced without extreme biases (Hair et al., 2019). In addition, the sample size of 168 respondents exceeds the requirement of at least 10 times the number of indicators for the most complex construct, ensuring sufficient statistical power (Hair et al., 2011). (see Web-appendix Table W4 for detailed descriptive estimates.) These characteristics, alongside the absence of missing data and adequate dispersion, confirm the data set’s suitability for PLS-SEM.

PLS-SEM is a variance-based multivariate analysis method that is suitable for assessing complex moderated mediation models with both reflective and formative constructs and latent scores for subsequent analyses (Hair et al., 2019b; Hair et al., 2011). We used SmartPLS Version 4.1.0.0, which offers extended algorithms and covariance-based capabilities (Ringle et al., 2022).

We used confirmatory composite analysis (CCA) to assess the contribution of each item to its construct and check construct validity and reliability. CCA is both exploratory and confirmatory, and thus is suitable for both reflective and formative items and for operationalising measures in a novel context (e.g. the metaverse) (Hair et al., 2020). First, we assessed the validity of our reflective constructs using 10,000 complete bootstraps with random seed sampling. Standardised loadings were all acceptable (t > 1.96), and their squared values showed sufficient variance between items and their constructs. Cronbach’s alpha and composite reliability were > 0.7, confirming construct reliability (Hair et al., 2020). Moreover, the average variance extracted from all constructs was > 0.5, confirming convergent validity (Hair et al., 2019b) (see Table 1). Finally, the Heterotrait–Monotrait ratio of correlations was below 0.90 for all constructs, confirming the discriminant validity of the model (Henseler et al., 2015) (see Table 2). Thus, CCA confirmed the reliability and validity of the reflective measurement model.

Next, we used CCA to assess the dimensional structure and validate the metaverse self-congruity construct. CCA is particularly suitable for confirmatory analysis assessing a complex measurement model such as our Type II construct, ensuring the proper operationalisation of measures in a new context like the metaverse (Hair et al., 2020). Following a top-down conceptualisation approach (Becker et al., 2023), metaverse self-congruity was conceptualised as a reflective-formative second-order construct, with the actual and ideal self as its dimensions. This follows the theoretical formation of self-congruity proposed by Sirgy and Su (2000), and its employment in virtual settings (Kim and Sundar, 2012; Suh et al., 2011).

The reflective-formative conceptualisation represents distinct but interrelated aspects of self-congruity, collectively forming a broader aspect of metaverse self-congruity, suggesting that the actual and ideal selves contribute uniquely to the construct. This conceptualisation aligns with user insights, such as one metaverse participant who noted:

This shows how users perceive a dynamic interplay between their actual and ideal selves, supporting the appropriateness of the reflective-formative structure for metaverse self-congruity.

By theorising actual and ideal self-congruity as two dimensions of metaverse self-congruity, we reduced the number of hypothesised relationships and, thus, the complexity of the structural model (Becker et al., 2023), producing more reliable estimates (Sarstedt et al., 2016). Supported by existing theory, participant insights and the validity of the first-order reflective constructs (actual and ideal selves) via CCA, we validated metaverse self-congruity using a two-stage approach (Sarstedt et al., 2019). First, the path coefficient (0.88) between the actual and ideal self-congruity scores was much higher than the threshold (0.7), strongly supporting convergent validity of the metaverse self-congruity construct. Moreover, the variance inflation factors of both indicators (ideal self = 2.006, actual self = 1.000) were below the conservative threshold of 3, demonstrating that multicollinearity was not a concern (Sarstedt et al., 2019). Next, we examined the significance of each dimension. The outer weights of 0.5 for actual self-congruity and 0.6 for ideal self-congruity (p < 0.001) show that both indicators significantly contributed to metaverse self-congruity. Furthermore, the outer loading of ≥ 0.5 indicates that both actual and ideal self-congruity are necessary for forming metaverse self-congruity (Hair et al., 2020) and, thus, should never be excluded. Overall, the results confirmed the validity of metaverse self-congruity as a Type II construct.

We used various methods to reduce common method bias. For example, in the survey design, we avoided complex and ambiguous terms, created a separate section for each variable (Kock et al., 2021) and sought the most appropriate scale (Rahman et al., 2022). In our data collection, we used clear instructions (Kock et al., 2021) and did not collect personal identifiers (Olivier et al., 2023). Respondents’ annual income, age and family status, which were theoretically unrelated to the constructs, were used as marker variables for the common method bias test (e.g. Aldhamiri et al., 2024; Benhayoun et al., 2024). Moreover, a variance inflation factor of < 3 indicates that the study was not significantly affected by common method bias (Hair et al., 2019b).

After confirming the validity of the constructs, we used PLS-SEM to analyse the structural model (Ringle et al., 2022). A bootstrapping procedure (10,000 bootstrap subsamples with a fixed seed) was used to examine the path coefficients with a 95% bias-corrected confidence interval. Table 3 shows that metaverse self-congruity has a significant and positive effect on CXMV, supporting H1.

To test our mediation hypotheses, we used a PLS-SEM bootstrapping procedure with 10,000 subsamples. Table 4 shows that metaverse self-congruity has a significant and positive direct effect on CWBec but a nonsignificant effect on metaverse usage intentions. The indirect effects of metaverse self-congruity on CWBec (effect size = 0.298, t = 5.027) and metaverse usage intentions (effect size = 0.346, t = 3.828) were both significant. This confirms that CXMV partially mediates the relationship between metaverse self-congruity and CWBec and fully mediates the relationship between metaverse self-congruity and metaverse usage intentions, supporting H2 and H3.

To assess the robustness of the path estimates, necessary condition analysis (NCA) was conducted for both H2 and H3 using SmartPLS v. 4 (Ringle et al., 2022). NCA is used to identify the necessary conditions for hypothesised relationships (Richter et al., 2023). We used 10,000 permutations with parallel processing and a fixed seed at a significance level of 0.05 (Ringle et al., 2022). The ceiling line (CE-FDH) for H2 shows that both CXMV (effect size = 0.166, p = 0.002) and metaverse self-congruity (effect size = 0.136, p = 0.047) are necessary conditions for CWBec, further supporting H2. The ceiling line for H3 shows that both CXMV (effect size = 0.299, p = 0.000) and metaverse self-congruity (effect size = 0.252, p = 0.009) are necessary conditions for ongoing metaverse usage intentions, further supporting H3.

Next, we assessed the predictive ability of the model (Ringle et al., 2022). The PLSpredict with 10 folds and 10 repetitions results of both PLSpredict (CWBec: Q² = 0.326; metaverse usage intentions: Q² = 0.156) and the CVPAT test (CWBec: −0.330, p = 0.000; metaverse usage intentions: −0.155, p = 0.028) confirmed the predictive capability of the model (Hair et al., 2022; Hair et al., 2019a).

In conditional mediation models, PLS-SEM is frequently combined with regression analysis, which complicates the analysis of relationships between observed and latent variables and hinders the ability to account for inherent measurement errors in abstract concepts (Sarstedt et al., 2020). This may affect the reliability and generalisability of the findings (Sarstedt and Moisescu, 2024). To overcome the shortcomings of moderated mediation analysis (e.g. artificial dichotomisation and correlation attenuation), we used PLS-based conditional mediation (CoMe) analysis, which combines mediation and moderation analyses to test moderated mediation hypotheses. Unlike traditional mediation approaches that assume the constant indirect effect of an independent variable on a dependent variable across all observations, CoMe enables the examination of how mediation effects vary across different levels of the moderator (Hayes, 2017). Also, the traditional approach does not account for interaction effects between the variables in the model, potentially oversimplifying relationships (Cheah et al., 2021; Sarstedt et al., 2020). Furthermore, the traditional approach focuses on overall mediation effects, making it challenging to offer path-specific recommendations for practical applications (Cheah et al., 2023; Riggs et al., 2024). CoMe is particularly valuable in metaverse research, where user interactions and experiences are highly contextualised and influenced by varying degrees of perceived similarity.

H4 and H5 posit that metaverse image similarity weakens the indirect positive effect of metaverse self-congruity on CWBec and metaverse usage intentions. Following Cheah et al. (2021), we tested H4 and H5 using all conditional mediation modes (Modes A, B and C).

Table 5 shows the CoMe results for the first stage of the mediated relationship (MSCG →CXMV). Metaverse user similarity significantly and negatively moderated the effect of metaverse self-congruity on CXMV (ω = −0.076, t = 2.379, 95% CI [−0.132, −0.027]). The low, medium and high CoMe effects are also significant as they do not include zero. The drop in the CoMe effect from 0.279 (low) to 0.126 (high) suggests a significant decline in the strength of the mediated path similarity (see Figure 2), such that when metaverse user similarity is one standard deviation (1.003) above the mean, the CoMe effect is reduced by 0.076.

Table 6 shows the CoMe effects for the second stage of the mediated relationship (CXMV → CWBec). Metaverse user similarity significantly and negatively moderates the effect of CXMV on CWBec (ω = −0.078, t = 1.988, 95% CI [−0.147, −0.019]). The low, medium and high CoMe effects are also significant. This confirms that metaverse user similarity weakens the relationship between CXMV and CWBec. When metaverse user similarity is one standard deviation (1.003) above the mean, the CoMe effect is reduced by 0.078 (see Figure 3).

Given that metaverse user similarity moderates both the first- and the second-stage conditional relationships in H4, we conducted Mode C analysis. Table 7 shows that metaverse user similarity negatively moderates both halves of the mediated relationship, such that when metaverse user similarity is one standard deviation (1.003) above the mean, the CoMe effect is reduced by 0.154. Figure 4 shows that both low and high CoMe effects increase, with the former having a steeper incline.

The CoMe index of −0.154 for Mode C is considerably higher than that of Mode A (−0.076) or Mode B (−0.078), indicating that when metaverse user similarity is higher, the indirect relationship between metaverse self-congruence and CWBec through CXMV is weaker. The reduction in the CoMe effect from 0.338 to 0.079 indicates a significant decline in the overall strength of the mediated relationship (i.e. the indirect effect is still positive, but its strength reduces). These results suggest that the moderating effect of metaverse user similarity is not isolated to either the first or the second stages of the indirect relationship. It also implies that a change in one stage of the relationship may lead to a change in the other. Therefore, the role of metaverse user similarity should not be considered in isolation. These findings suggest that for metaverse users who perceive a higher similarity between themselves and the typical metaverse user, the effect of CXMV on CWBec may not be as strong.

Table 8 shows the CoMe results for the first stage of the mediated relationship (MSCG → CXMV). Metaverse user similarity significantly and negatively moderates the effect of metaverse self-congruity on CXMV (ω = −0.091; t = 2.475), suggesting that as the similarity between the metaverse user and typical users increases, the effect of metaverse self-congruity on CXMV decreases (Figure 5). The drop in the CoMe effect from low (0.334) to high (0.151) indicates that the relationship diminishes with an increase in metaverse user similarity.

Table 9 shows the CoMe results for the second stage of the mediated relationship (CXMV → USIN). Although the CoMe effect slightly decreases from 0.259 to 0.228, the CoMe index is nonsignificant. For completeness, we also conducted Mode C CoMe analysis, the results of which were nonsignificant. Therefore, CoMe analysis does not support a combined moderating effect, meaning that metaverse user similarity moderates the effect of metaverse self-congruity on CXMV but not metaverse usage intentions.

Following the recommendation of Chou et al. (2023), a post hoc qualitative study with metaverse platform users was conducted to validate our quantitative results and provide deeper insights. A qualitative study protocol was developed to address three research questions (see Web-appendix Table W5):

1. How do users conceptualise and experience the relationship between their actual and ideal selves in the metaverse?

2. What aspects of CXMV contribute to their well-being in real life?

3. How do users understand and define a typical metaverse user?

Participants for this study were recruited via an online platform, and the survey was designed using Qualtrics. To ensure data quality and confirm that participants were active metaverse users, a two-step data collection process was implemented (Sharpe Wessling et al., 2017). In the first stage, a participant pool was created by identifying individuals who described their current engagement with metaverse platforms. In the first stage, a participant pool was created with individuals identified as current. In the second stage, participants responded to seven open-ended questions (see Web-appendix Table W5). Compensation was provided for their time and effort.

A total of 33 metaverse users participated in the study. Given the complementary nature of this research and in alignment with qualitative methodologies that often rely on relatively small, purposive samples (Patton, 2014), participants were selected based on their ability to provide rich, detailed insights into the phenomenon under investigation. In addition, as noted by Chou et al. (2023), the research team determined that the sample size was appropriate for a focused study aimed at enriching the survey results.

Content analysis was conducted independently by three researchers who subsequently discussed their interpretations of the data. This collaborative process facilitated triangulation by incorporating diverse perspectives and interpretations (Patton, 2014). The researchers approached the data with reflexivity and analysed participants’ descriptions in the transcripts to ensure alignment with their experiences (Chou et al., 2023). The data was manually coded, using deductive thematic analysis. Predetermined concepts from prior empirical studies, for example, self-congruity and user similarity, guided the deductive analysis, providing deeper insights into these concepts (Fereday and Muir-Cochrane, 2006).

The detailed quotes and analysis of the post hoc study can be found in Web-appendix Table W6. In this section, we provide a summary of the main points. The analysis shows alignment between users’ avatars and their actual selves as well as their ideal selves. This supports the hypothesised link between the concepts under investigation, highlighting that such an alignment self-enhances immersive experiences and contributes to both CXMV and CWBec. Moreover, participants provided details of how the metaverse positively influences various life domains, validating the hypothesised connection between CXMV and CWBec. In addition, responses reflecting the long-term, evolving impacts of metaverse engagement align with the conceptualisation of CWBec in the research model, supporting claims about the dynamic benefits of metaverse use for personal growth in real-life settings. Furthermore, responses regarding user perceptions of others in the metaverse clarify the role of perceived similarity in shaping CXMV. The findings highlight the moderating influence of user similarity on the relationship between self-congruity, CXMV and CWBec, while also providing qualitative evidence of the challenges associated with homogeneity in a metaverse environment.

This study has examined three key questions about metaverse users’ behaviour: the influence of self-congruity on CXMV, the influence of CXMV on CWBec and usage intentions, and the moderated mediation effect of user similarity. The findings support our initial assumption and make four contributions to the literature. Firstly, it responds to calls for a new theoretical model of CXMV (Hadi et al., 2024) and CWBec (Eshaghi et al., 2023) by providing a theoretical basis for examining metaverse user behaviours. Secondly, drawing on S-O-R logic and self-congruity theory, it identifies a new determinant of CXMV, namely, metaverse self-congruity, extending the literature on self-identity in virtual worlds. Thirdly, it examines how CXMV influences CWBec and metaverse usage intentions. This study also operationalises CWBec in immersive technologies. Fourthly, we use CoMe analysis to test the moderated mediation relationships between self-congruity, CXMV, CWBec and metaverse usage intentions, demonstrating the dampening effect of metaverse user similarity.

This study enhances our understanding of metaverse user behaviours. While researchers have investigated CWBec, its role in the metaverse has not been fully examined yet (Eshaghi et al., 2023). Accordingly, we explore the operation of CXMV and its influence on CWBec (Gahler et al., 2023). Despite considerable research on customer experience and behaviour in virtual worlds, such as gaming, AR and VR (e.g. Ko and Park, 2021; Li and Lwin, 2016; Wu and Hsu, 2018), research on CXMV and CWBec is scarce. This study addresses this gap by operationalising six CXMV aspects (Mansoor et al., 2024) and demonstrating its mediating role between self-congruity, CWBec and usage intention. The uniqueness of the metaverse, including its immersiveness and synchronicity and the interconnectedness of avatars (Hennig-Thurau et al., 2023; Jiang et al., 2023), can improve CXMV and CWBec across various life domains, such as family, social, work and financial domains. The study also operationalises a scale for CWBec in the metaverse. Therefore, it contributes to both the well-being and metaverse experience literature, providing a useful guide for future researchers.

This study’s findings extend self-concept and possible selves theory by showing how virtual experiences can translate into users’ real-life well-being across multiple life domains. Previous studies on the role of the self in virtual settings have shown that both the actual and the ideal selves play an important role in shaping consumers’ holistic self-perceptions (Balakrishnan et al., 2024). In the context of the metaverse, self-congruity is influenced by users’ avatar designs (Park and Kim, 2024). As a unique representation of the self, avatars enable users’ sense of immersion (Kim et al., 2023), realism (Zhao et al., 2022), social connection (Oh et al., 2023b) and enjoyment and pleasure (Shin et al., 2023). The closer an avatar is to a user’s real-world self, the greater the sense of congruity and “fit” (Park and Kim, 2024). However, the effect of user–avatar congruity on CXMV has not previously been examined. Our findings show that self-congruity is necessary for a positive CXMV, contributing to our theoretical understanding of behaviour in the metaverse. Our findings also show the positive influence of metaverse self-congruity on CWBec and metaverse usage intentions, demonstrating the effects of virtual environmental characteristics on both immersive experiences and real-world responses.

This study also makes theoretical contributions by refining the conceptualisation of self-congruity within metaverse research. We draw on Sirgy’s (1982, 1986) foundational framework, which examines the self and positions self-congruity as a higher-order construct encompassing multiple self-image dimensions. Specifically, we focus on actual and ideal self-congruity and conceptualise metaverse self-congruity as a reflective–formative second-order construct formed by these two dimensions. This approach aligns with prior work in consumer research (e.g. Hosany and Martin, 2012, p. 688), where the interplay between actual and ideal selves jointly defines the self-congruity experience. Methodologically, this modelling allows us to account for multicollinearity between dimensions while capturing a parsimonious representation of avatar-based self-congruity in the metaverse context (Becker et al., 2023).

By integrating CXMV and user similarity into the S–O–R framework, this study demonstrates how stimuli in the metaverse can influence users’ responses. CXMV is significantly influenced by the design of the metaverse platforms. If the platform thus provides sufficient avatar design options, users can achieve higher self-congruity (Yee et al., 2009). Our findings demonstrate that self-congruity contributes to a positive CXMV, in turn enhancing CWBec and metaverse usage intentions. This extends the framework’s applicability to immersive environments and provides a foundation for further exploration of consumer behaviours in the metaverse.

By applying advanced PLS-SEM-based CoMe analysis, our study gains insights into the interaction between personal and social factors. Specifically, we observe how user similarity moderates the pathway from avatar self-congruity to CXMV and further to CWBec and usage intentions. This analysis shows that the indirect effect of avatar self-congruity is significantly stronger when user similarity is low, providing evidence of the importance of diversity in user representation. This finding challenges the common assumption that similarity enhances virtual experiences in a linear way, suggesting that users seek a balance between belonging and distinctiveness in metaverse environments.

CoMe also allows us to quantify these effects, showing how mediation strength changes across varying levels of user similarity. This study is the first to use this technique to explain the moderating role of metaverse user similarity in these relationships. If we had relied solely on traditional mediation analysis, we would have overlooked a critical understanding of how user similarity influences the relationships in our model. Traditional mediation assumes that the indirect effect of avatar self-congruity on CWBec and usage intentions through CXMV is uniform across all conditions. This approach would fail to explore how the strength of these effects actually varies depending on the level of user similarity. For example, we would not have identified that the positive impact of avatar self-congruity diminishes at higher levels of user similarity, which could result in oversimplified conclusions and missed opportunities for targeted platform interventions. Also, without considering the dampening effect of metaverse user similarity, lower-than-expected CWBec may be difficult to explain. These findings make novel contributions to the metaverse literature, suggesting that researchers should consider including metaverse user similarity as a contingent variable to avoid estimation errors.

This study offers critical insights for managers. Firstly, it shows that the greater the similarity between a metaverse user and the image of a typical metaverse user, the lower their CWBec and metaverse usage intentions. This may be because the metaverse is a medium through which customers seek a vastly different experience from their real lives (Han et al., 2022; Pal and Arpnikanondt, 2024). It follows that if other avatars are too similar to a customer’s real-life social peers, the platform may fail to offer the desired escape from real life that customers seek. Therefore, metaverse platform designers should identify the real-life situations from which users wish to escape and the myriad ways in which users seek to personalise their experiences. This knowledge would allow designers to offer avatar design tools that enable users to differentiate themselves from others and provide a highly personalised virtual escape. These tools could include advanced facial recognition, diverse cultural attributes and unique personalisation options. This may reduce the dampening effects of high user similarity and encourage users to continue using the metaverse platform.

In addition, our findings offer valuable guidance for platform managers seeking to enhance user experience and engagement. Given that users construct avatars blending their actual and ideal selves, platform designers should consider expanding avatar customisation options that enable users to express aspirational traits. Moreover, personalisation is a critical component of metaverse customer experience (Rahman et al., 2025a). Platform managers can enhance user well-being by delivering personalised onboarding experiences and curating diverse social environments that reflect varied user identities and interests. These efforts may reduce perceived user similarity, allowing users to feel more at ease expressing their individuality and engaging in meaningful interactions.

Secondly, this study demonstrates the positive effect of CXMV on CWBec and metaverse usage intentions. Metaverse platform managers can offer personalised avatar design and customer experiences and create an immersive metaverse environment. Metaverse users attempt to construct a new identity by creating a balance between their actual and ideal selves while simultaneously differentiating themselves from other avatars. Applying realistic human avatars is more effective for long-term relational exchanges (Miao et al., 2022) because they enrich the user experience and boost future usage intentions. Moreover, users wish to be immersed in an emotionally and cognitively intelligent environment. Rich and multifaceted metaverse experiences will contribute to consumers’ ongoing metaverse usage intentions and real-world financial, social, family and work well-being.

In other words, the metaverse is not merely a tool for fun; rather, it may contribute to a user’s overall life satisfaction. Therefore, metaverse platform designers should focus on enriching CXMV and consumers’ quality of life by extending metaverse functionality to life-enhancing domains targeting various life aspects such as entertainment, daily activities, education, health, shopping and banking. For instance, virtual therapy sessions could support users’ mental well-being, while immersive educational experiences can enhance learning outcomes. Moreover, the metaverse may enhance user independence and self-efficacy, particularly for vulnerable users who find it challenging to connect with others in real-world environments. Incorporating virtual fashion, customisable personal spaces and identity-aligned communities can strengthen user engagement and promote long-term usage intentions. Consequently, it can enrich social interactions and enable users to participate in collective activities and well-being, reducing the diluting effect of metaverse user similarity.

Thirdly, metaverse platform managers should periodically assess the success or failure of new initiatives. Overemphasis on avatar personalisation may lead to cognitive dissonance or dissatisfaction if users feel their avatars deviate too much from reality. Therefore, platforms should monitor and provide feedback mechanisms to help users align their virtual and real-world identities. This study offers a validated and easy-to-administer survey through which to gather feedback from customers on metaverse platforms, enhancing their perceptions of control (Mehrotra et al., 2024). This will improve the platform–customer relationship and induce in customers a sense of ownership of their experience and the metaverse platform.

We acknowledge that the relatively modest sample size may limit the broader applicability of our findings. This constraint stems from our data collection approach, which focused on active metaverse users within the Australian Qualtrics panel. While this panel is known for delivering high-quality data that closely reflects real consumer behaviour (Arndt et al., 2022; Douglas et al., 2023), the use of strict screening criteria significantly increased costs, thereby limiting the number of responses we could afford to collect. Moreover, financial limitations further reduced our capacity to expand the sample. Nonetheless, the final sample meets the minimum requirements for PLS-SEM analysis as outlined by Hair et al. (2022) and supports the validity of our proposed model.

However, a larger sample size could significantly enhance the scope and depth of future analyses. For instance, a larger sample would enable a multi-group analysis to examine potential differences in user behaviour across segments such as demographics, metaverse platform usage frequency or levels of technological literacy. This approach could offer further insights into how different groups experience and interact with metaverse platforms, informing more targeted managerial strategies. In addition, a larger data set would allow for the exploration of unobserved heterogeneity using PLS-based advanced techniques like finite mixture PLS (FIMIX-PLS) or PLS-POS (prediction-oriented segmentation) (Sarstedt et al., 2023). These methods could uncover latent segments within the data, identifying valuable consumer profiles that are not apparent in aggregate analyses. For example, segments might differ in their sensitivity to avatar self-congruity or their reactions to user similarity in the metaverse, offering deeper theoretical and practical insights.

Future studies could also leverage additional samples to explore cross-national differences, comparing metaverse experiences and behaviours in Australia with those in other countries. Future research could examine how cultural dimensions (e.g. individualism-collectivism) influence avatar creation and self-presentation in the metaverse, as these factors may significantly influence user behaviour and experience across different cultural contexts (Kim et al., 2016). This would enhance the external validity of the findings and provide a deeper understanding of metaverse user behaviour globally.

Moreover, while our study identifies user similarity as a significant moderator, further research is needed to explore the psychological mechanisms underlying this effect. Experimental studies could examine how cognitive dissonance mediates the relationship between user similarity and CWBec. Such research could show how varying levels of user similarity may create psychological tension between users’ needs for uniqueness and belongingness. In addition, the concept of self-congruity in this study is based on a combination of the actual and ideal selves. However, the social self and the ideal social self are also important (Sirgy, 1986). We acknowledge that other dimensions such as social and ideal social self are part of Sirgy’s broader framework but were not included in the current operationalisation due to scope and relevance considerations in our metaverse setting, which prioritises personalised avatar experiences over social approval mechanisms. Therefore, research is needed on other dimensions of the self to further understand the effect of self-congruity on CXMV and its outcomes.

Different metaverse platforms offer varying stimuli such as graphics styles, user interfaces and interaction possibilities, influencing the quality of avatar customisation and interactions (Javornik et al., 2021). Future researchers could examine how the technological features of platforms influence avatar creation, shedding light on the complexity of self-congruity. In addition, platform-specific features and affordances (e.g. Roblox vs Decentraland) may moderate the relationship between self-congruity and customer experience, showing how technological capabilities could shape user identity expression and social interaction patterns (Wang et al., 2024; Zhang and Chang, 2021). Moreover, while we used subjective measures of well-being, future studies could use objective measures such as eye tracking and brain scanning during immersive experiences (Eshaghi et al., 2023).

Finally, while PLS-SEM analysis offered valuable insights, future research could use alternative methods to uncover deeper patterns in metaverse user behaviour. For example, cluster analysis could help identify distinct user groups based on avatar creation and interaction style, while conjoint analysis could explore which metaverse features users value the most, helping platforms improve their design. Furthermore, longitudinal studies that track users over time could offer insights into how behaviour and CWBec change with ongoing metaverse use, showing the long-term effects of virtual engagement. In addition, future research could also examine the negative consequences of metaverse usage, including perceived social, emotional and behavioural risks to user safety such as harassment, bullying and customer ill-being (Suh, 2024; Suh and Prophet, 2018).

The supplementary material for this article can be found online.