Elucidating continued usage of mobile wallets application from the lens of extended expectation-confirmation model Open Access

In the evolving digital payment landscape, mobile wallet (m-wallet) are playing a pivotal role in transforming consumer behavior and payment practices (Hanafiah et al., 2024). The m-wallet is an application-based tool that securely stores payment information and passwords for facilitating various financial transactions and online services (Kumar et al., 2024). Globally, the m-wallet market is forecasted to grow to $16.2tn by 2031, with a CAGR of 22.2% (PR Newswire, 2022). In India, m-wallet payments, by value, are expected to reach $4.1tn in 2025 with a CAGR of 42.7% (Pai, 2022). Recent studies consistently highlight determinants such as speed and convenience (Sharma et al., 2018), transaction security (Shao et al., 2019), perceived risk, app usability, fulfilment reliability and multi-payment capabilities as key drivers of both adoption and continued usage (citations; del Pilar Pizzan-Tomanguillo et al., 2024). Although evidence suggests that m-wallet providers integrate app features, transaction processes and digital payment options to influence consumer behavior, they continue to face challenges in driving app engagement, encouraging repeat usage, building customer loyalty, and, therefore, its continuance usage (de Luna et al., 2019; Gupta et al., 2020; Hilal and Varela-Neira, 2022).

Despite the growing body of literature on m-wallet usage, the field remains theoretically fragmented and methodologically inconsistent. A critical review of recent studies, as shown in Table 1, reveals a persistent tendency to treat adoption and continuance as separate phenomena, often relying on singular theoretical lenses that fail to capture the full user journey (Yousaf et al., 2025). For instance, studies such as those by Salah and Ayyash (2025) and Chand and Kumar (2024) extend the technology acceptance model (TAM) and unified theory of acceptance and use of technology (UTAUT), respectively, to explain adoption, yet stop short of addressing what drives users to continue using m-wallets postadoption. This narrow focus on initial behavioral intention, without tracing its evolution into sustained usage, limits the practical relevance of their findings. Conversely, Zaidan et al. (2025) and Nguyen-Viet and Hoang Nguyen (2025) attempt to explore continuance behavior but do so using frameworks like theory of planned behavior (TPB) and the mobile expectation-confirmation model (MECM), which either lack the conceptual depth to explain postadoption satisfaction (as in TPB) or are overly focused on context-specific features like gamification and grievance redressal, thereby sacrificing theoretical generalizability. Even when continuance is addressed, as in Nguyen-Viet and Hoang Nguyen (2025), their work fails to link preadoption beliefs with postadoption evaluations, leaving a conceptual void between what motivates users to adopt and what sustains their usage.

Moreover, the over-reliance on behavioral intention as a proxy for actual or continued use is a recurring methodological flaw. Studies such as Herzallah et al. (2025) and Joshi and Chawla (2024) report high explanatory power for intention but do not empirically validate whether these intentions translate into real-world continuance usage or not. This gap is particularly problematic in the context of m-wallets, where initial enthusiasm often fails to convert into habitual use due to usability issues, trust concerns or unmet expectations. Prior studies have also not examined the effect of external cues, such as social influence (Adams et al., 2017) and perceived trust, that can influence users’ continuance usage intentions toward m-wallets (Patil et al., 2020).

While studies confirm that a few variables, demographics majorly, moderate users’ continuance usage intentions toward m-wallets, the role of perceived value as a moderator in m-wallets is still unexplored and needs investigation. Although prior studies have examined various aspects of m-wallets adoption and usage, there remains a limited understanding of how these platforms can enhance satisfaction among existing users to foster loyalty-driven outcomes (Yousaf et al., 2025). Consequently, recent research emphasizes the need for integrative frameworks to offer deeper insights into the factors shaping loyalty-oriented behaviors in app-based contexts (Yousaf et al., 2025; Ravichandran et al., 2024).

In light of these arguments, our study proposes a novel integration of TAM and the expectation-confirmation model (ECM) to bridge the adoption-continuance divide. By combining TAM’s strength in modelling preadoption beliefs with ECM’s focus on postadoption satisfaction and confirmation, we offer a more comprehensive and generalizable framework. In addition to existing variables of TAM and ECM, a few additional variables are added to the postadoption model, namely, perceived trust and social influence. The moderating role of perceived value is also explored. This integration allows us to trace the full trajectory of m-wallet usage, i.e. from initial intention to sustained engagement, thereby addressing both the conceptual fragmentation and methodological gaps that have constrained prior research. In doing so, we challenge the prevailing reliance on intention-based proxies and offer a more behaviorally grounded understanding of m-wallet continuance usage.

From the unified perspective between ECM and TAM and/or UTAUT, Scholars (Gupta et al., 2020; Jaiswal et al., 2022) studied continued usage of mobile-based payment services in the backdrop of emerging economies. This blended approach acknowledged the presence of some important components (perceived trust, social influence and perceived value) on the intended use of technology, such as a digital payment platform. Therefore, trust, value and societal impact are one of the critical factors in shaping of users continued usage behavior.

The conceptual underpinning of TAM is built upon the theory of reasoned action and the TPB (Fishbein and Ajzen, 1975; Jaiswal et al., 2021). The TAM comprises two fundamental beliefs for acceptance of new technology: perceived usefulness (“the degree to which a person believes that using a particular system would enhance his or her job performance”) and perceived ease of use (“the degree to which a person believes that using a particular system would be free of physical and mental efforts”). These frameworks are valuable due to their emphasis on ease of use and perceived usefulness, which are key factors in determining the adoption of m-wallets (Chand and Kumar, 2024; del Pilar Pizzan-Tomanguillo et al., 2024). The extant literature applied TAM in the varied contexts of digital technology usage, e-shopping (Tong, 2010), e-commerce (Pavlou, 2003), mobile instant messaging (Jiang and Deng, 2011) and m-commerce (Wei et al., 2009). Scholars have also widely explored the TAM to adapt intent and continuance intent for technology research, especially in smartphone-led technology (Zhou, 2014; Cao et al., 2018).

Customers’ adoption, consumption and diffusion of new technologies have attracted researchers for years (Yousaf et al., 2021). However, most of these researches are limited to understanding intentions (Sarmah et al., 2021) and initial adoption of m-wallet apps and focus less on understanding the postadoption consumer behavior (Gupta et al., 2020). Most studies on m-wallet apps are bourgeoning phase only to initial acceptance, without considering postacceptance behavior. Recently, Al-Saedi et al. (2020) examined factors that led to the initial adoption of m-wallets usage among smartphone users. Across such studies, the preadoption intentions of m-wallet apps are focused primarily on minimal insights into postadoption user behavior. Considering this phenomenon, it is well recognized that positive confirmation from the consumption experience of a product/service affects consumer’s overall satisfaction, thereby determining a favorable postconsumption satisfaction level and repurchase intentions (Yousaf et al., 2021). Using self-perception theory, the postadoption ECM was proposed by Bhattacherjee (2001) to study consumers’ beliefs and attitudes toward the technology usage and re-usage, and their continuance usage behavior. It was argued that consumers constantly match their expectations with their own and others’ behavior, and, in turn, adjust this newer fact with beforehand accrued information (Bhattacherjee, 2001).

As a result, users keep modifying their expectations of an innovation or new technology, either through personal experiences or interaction with others, as they learn more about it (Thong et al., 2006; Yousaf et al., 2021). Perceived usefulness is one’s cognitive understanding about actual benefits, user satisfaction represents the meaningful ex post experience of the user with an IS product/service and continual usage intentions is the his/her usage intent to continue the technology in the forthcoming (Gupta et al., 2020). With the confirmation of expectations and perceived usefulness as its key determinants, the ECM expounds the satisfaction and continuance of innovative product and service usage (Rahi et al., 2021). With emerging technologies like m-wallet apps, understanding users’ continuance usage intentions is more important as they take time to adapt to the technology (Bhattacherjee, 2001; Yousaf et al., 2021). On the lines of Venkatesh et al. (2012), this study extends the ECM by integrating the primary antecedents of TAM and adding postadoption contextual factors, namely, perceived trust, perceived value and social influence. Perceived trust is like a service provider will take action that will benefit the consumers and will not engage in activities that will jeopardize their interest (Chellappa and Pavlou, 2002). Perceived value, for instance, a m-wallet service provider can provide user support programs that assist users in resolving issues, responding to inquiries and offering feedback on the service (Salah and Ayyash, 2025). Social Influence is consumers’ perception of critical people’s recommendations and support that will impact their decision to use the m-wallet (Yang et al., 2021).

In the present research, we link the ECM and TAM frameworks to theorize and test the determinants of continued intention toward m-wallet apps. This study extends the ECM and TAM by adding dimensions of ECM and TAM and additional contextual determining factors, namely, perceived trust, perceived value and social influence. The Proposed framework and hypotheses are presented in Figure 1.

Confirmation of expectation is the users’ evaluation of expectations vs actual use. The direct positive connection between confirmation with usefulness and users’ satisfaction is well established in the extant ECM literature (Bhattacherjee, 2001; Jaiswal et al., 2022). It is argued that users’ have expectations from m-wallet apps, and if their expectations are confirmed after the actual use of m-wallet apps, they feel the application is useful and feel satisfied (Thong et al., 2006; Yousaf et al., 2021). Ambalov (2018) in his work confirms the association between confirmation, usefulness and user satisfaction. In their work, Gupta et al. (2020) argued that users’ specific functionalities of a m-wallet app are captured by usefulness and are built over the fundamental expected functions. If these primary functionalities of the m-wallet app are much above users’ rudimentary expectations, the customers will be delighted, enhancing perceived usefulness and satisfaction with the m-wallet application and vice versa (Jaiswal et al., 2022; Lu et al., 2017). Other researchers in similar technological contexts have also established the same relationship, i.e. mobile shopping apps and mobile banking (Shang and Wu, 2017), smartwatches, intelligent wearables and fitness apps such as digital health wearables (Karjaluoto et al., 2019) and OTT platforms (Yousaf et al., 2021). Thus, this leads to the following hypothesis:

The extant technology adoption and ECM literature suggest perceived usefulness as the main predictor that has the highest influence on user satisfaction with the IS product/service and its continuance usage (Jaiswal et al., 2022; Venkatesh et al., 2003). Perceived usefulness reflects users’ belief that using a particular IS product/service will augment their effectiveness (Davis, 1989). In m-wallet systems, users are more likely to adopt the technology when they perceive substantial benefits, such as increased efficiency and cost effectiveness, which businesses facilitate by offering discounts and streamlined processes (Amoroso and Magnier-Watanabe, 2012). The ECM also theorizes the positive relationship of perceived usefulness with users’ satisfaction and their continuance usage intentions of IS products/services (Hong et al., 2017). These connections were also statistically tested by Gupta et al. (2020). Thus, the following hypotheses were formulated:

Perceived ease of use reflects users’ belief as the easiness/complexity associated with the use/consumption of an IS product/service. Ease of use is defined as “the degree to which the prospective user expects the target system to be free of effort” (Davis, 1989). In the context of m-wallet, perceived ease of use plays a critical role in shaping user adoption (Alhassan et al., 2020). The perceived ease of use is one of the most significant factors in user attitude and behavioral intention to adopt and use technology (Chawla and Joshi, 2023). According to TAM, perceived ease of use directly influences perceived usefulness, as users tend to perceive systems that are easier to use as more beneficial (Mehra et al., 2021). In the extant technology adoption and postadoption consumption literature, ample evidence is available that shows that perceived ease of use toward mobile technologies such as mobile devices and m-commerce positively impacts perceived usefulness and users’ satisfaction with the IS product/service (Wei et al., 2009; Gupta et al., 2020). Davis et al. (1989) argued that even if users view an IS product/service as useful; they may not be motivated to use it if it is perceived as complex/challenging. While using/adapting new technology, users of IS products/services see if its use requires consideration or less effort; if the technology is easy to use, it is considered more useful (Davis et al., 1989; Barry and Jan, 2018). Recently, Gupta et al. (2020) established perceived ease of use as a primary driver for perceived usefulness of m-wallet applications and empirically validated the same. Similar observations were noted by scholars in the milieu of e-wallet services (Yousaf et al., 2021). Therefore, we posit the following:

Perceived trust is the users’ belief that the m-wallet applications promise/process is reliable, trusted and can enhance user dependency on the IS product/service (Chawla and Joshi, 2019). Perceived trust, which includes a user perception that the app is trustful in its commitments and promises, as well as is acting in his/her best interest in terms of saving his/her personal financial information (credit/debit cards or UPI information), enhances users’ satisfaction and continuance usage intentions of the m-wallet application. They must trust that the transaction will be completed according to expectations and that any data shared will not be shared with inappropriate parties (Chellappa and Pavlou, 2002). According to Matemba and Li (2018), a trust-based model in the context of m-wallets emphasizes the importance of user trust in the service provider, influencing their willingness to use the wallet for transactions. This model recognizes that trust is a critical factor in shaping consumer attitudes and intentions to adopt and use m-wallets. The user must experienced the platform as trustworthy and committed to his/her explicit motives to hold the application and use it continuously (Hasan and Gupta, 2020). Users need to believe that the m-wallet provider is reliable, trustworthy and has their best interests at heart. A market offering with trust indicates that investments decision for skills, resources and care to the customer relationship structure have been taken care of by that company. In an online marketplace, transactions take place when there is a belief in the reliability and honesty of the parties, some of whom may be unknown to the consumer (Ganguly et al., 2010). Users tend to favor businesses with a good trust in the IS domain as they conceive lower risk and ambiguity and are attentive where to seek benefits from the community if something really goes wrong (Alalwan et al., 2017; Chao, 2019). Lack of trust in product/service will be a barrier to continuance usage since users have to share their personal and financial information with these apps (Gupta et al., 2020). If the user trusts the goodwill of the m-wallet application developer/merchant, they will continue using the application and will be satisfied (Chao, 2019). This aligns with the expectation that a positive reputation of an online business enhances customer trust, leading to increased mobile payment utilization. The relationship of users trust with their satisfaction and continuance usage intentions is well established in m-wallet literature (Alalwan et al., 2017; Zhou, 2014). Hence, this led to the following hypotheses:

In the postadoption/consumption behavior literature, satisfaction with an IS product/service has been cited as the most important predictor of continuance usage intentions (Bhattacherjee, 2001; Gupta et al., 2020). User satisfaction with a m-wallet application shapes users’ positive attitude toward the technology and motivates him/her to continue using the application (Jaiswal et al., 2022). With hundreds of m-wallet services accessible for download and installation on mobile devices, the choice to continue using one specific m-wallet platform is mainly predicted by individual satisfaction with the application (Gupta et al., 2020), which, in turn, is driven by confirmation, usefulness and user trust (Yousaf et al., 2021). The above arguments lead to a hypothesis:

Social influence is the degree to which a user decides to adopt an IS product/service is influenced by societal stakeholders or individuals who are important to a user, such as family, friends, superiors and colleagues (Fishbein and Ajzen, 1975; Venkatesh et al., 2012; Jaiswal et al., 2022) and/or other factors such as magazines, internet, television/radio/newspaper ads. This important phenomenon is conceptualized as the influence of close societal settings (peer groups, family members, relatives and friends) on users’ willingness to use new technology such as digital payment services (Zhou et al., 2010). Matemba and Li’s (2018) study reveals that trust is a pivotal factor influencing the acceptance of social influence. Social influence is an important determinant of user’s continuance usage as it reflects the perceived opinions of others, which are important to the user (Alalwan et al., 2017; Wei et al., 2009). The positive relationship between social influence and users’ continuance usage intentions has been well established and validated in the mobile payment application studies (Alalwan et al., 2017; Jaiswal et al., 2022). Based on the above grounds, it is hypothesized that:

Scholars defined perceived value as an individual’s subjective trade-off between two different facets for exchange of value – what benefits receive against what costs give (Kant et al., 2019; Zeithaml, 1988). This indicates that if the monetary and nonmonetary costs are high, there is a curtailed net perceived value from the customers’ context (Biswas et al., 2021; Shaw and Sergueeva, 2019). Under the considerations of the cost-benefit paradigm, this cognitive phenomenon implies users’ rational decision-making paradigm for innovative technology adoption (TAM, UTAUT) seeking to maximize value or benefits (quality, convenience and performance) in the evolving smartphone-based shopping and payment services (Wang et al., 2020).

In addition, this study adds perceived value into the framework as the moderator variable to widely explain user satisfaction and intention to continuous use of e-wallet (Liu et al., 2020). It is considered fundamental for all marketing activities to maintain competitive advantages (Karjaluoto et al., 2019). Perceived value will give some emotional consumer’s expression, such as fun, enjoyment and expressiveness, that lead to the positive effect on satisfaction (Liu et al., 2020). Thus, it is expected that consumer perceived value will strengthen the relationship between user satisfaction and intention to continuously use the e-wallet services. Extant literature indicates that the higher perceived value leads to higher satisfaction and behavioral intentions of online payment technology and further its continued usage, such as digital wallet transactions for utility and e-shopping services (Wang et al., 2020). Tamilmani et al. (2021) noted that the connection between perceived value and behavioral intentions had varied outcomes.

Consumers assess the value provided by different products or services and opt for those that deliver the highest perceived value. Moreover, perceived value leads to continuance usage of technology directly and indirectly via trust and satisfaction in the context of online services (Pavlou, 2003; Xiong et al., 2022). However, limited studies noted the moderation of perceived value, indicating that this subjective measure strengthens the casual connection between satisfaction and behavioral intention in online banking (Kumar et al., 2020). Furthermore, no such studies tested the moderating effect of perceived value in the milieu of e-wallet payment services, particularly for the direct relationship between perceived trust and continued behavioral intentions. In a similar observation, the authors argued that perceived value could also moderate the association between perceived trust and continued intention. Thus, a low perceived trust with users’ high perceived value could lead to a greater number of behavioral intentions than having high perceived trust with low perceived benefits in the backdrop of e-wallet technology’s continued usage. Hence, to fill the crucial literature gap, this timely study postulates the following hypotheses:

To analyze the integrated framework, a research instrument was developed to obtain responses from existing smartphone wallet users. All measurement items used in this study were adapted from well-established scales widely used in prior technology-adoption research. Each construct was selected based on its demonstrated theoretical relevance to the chosen context. Although the original instruments were developed in different cultural settings, these scales have been repeatedly applied across diverse national contexts and have shown strong reliability and validity, indicating their robustness across cultural boundaries. To ensure suitability for the present study, we followed a systematic adaptation process: items were first reviewed by three domain experts to assess conceptual equivalence, after which a bilingual academic panel evaluated semantic clarity and cultural appropriateness. Minor linguistic adjustments were made to enhance comprehensibility without altering the underlying constructs. A pilot test with a small sample of 40 target users further confirmed that the adapted items were easily understood and retained their intended meaning. This multistep process supports the cultural validity and methodological soundness of the measurement instrument used in this research.

The survey instrument comprised a thanking note followed by two succeeding parts which ensured information given by participants would be used for only for research purposes and kept confidential. The first section was linked to users’ sociodemographic attributes (gender, age, education, income and residing area) by asking general questions seeking information about the frequency, purpose and types of using m-wallet payment platform. Before this, the dichotomous questions were asked to ensure the target population of interest was the existing users of m-wallet platform or related apps. The next part comprised measurement items with their sources (mentioned in  Appendix) to capture the research phenomena using a seven-point Likert scale with anchored 1 (“Strongly disagree”) to 7 (“Strongly agree”).

A cross-sectional online questionnaire approach was used in this study. The study focuses on active users of smartphone wallets (m-wallets) in the Indian digital payment market (Kapoor et al., 2024). Additionally, to capture cutting-edge technology adoption behavior shaped by fintech innovation, the researchers need to focus on m-wallet users (Jaiswal et al., 2023; PwC, 2023). Therefore, this contextual setting is crucial for the advancement of technology adoption theory (Kapoor et al., 2024). This includes people who regularly use m-wallets for digital payments and fintech-related services (such as banking, recharging, utility bill payment, shopping and financial payment). An online survey was administered through e-mails and WhatsApp links to reach out to the target population of interest, who have smartphones and are currently using m-wallet apps (Shetu et al., 2022). A nonprobabilistic purposive sampling technique was used to contact the network of peers in academia, business and students to collect responses (Jaiswal et al., 2023). In light of this, an online survey facilitates timely and economic data collection and can reach smartphone users, particularly those who use m-wallet applications across a range of geographic locations, including rural, semiurban and urban areas of India (Jaiswal et al., 2022; Rana et al., 2023). Thus, online survey methods are acknowledged to offer the benefit of having a broad geographic reach (Saunders et al., 2009).

Over 900 respondents were approached, starting with dichotomous questions related to ensuring they were current users of m-wallet apps or related payment platforms; only 574 data were valid for the analysis after eliminating mono or alike type answered throughout the survey, followed by outliers. The sample size of 574 was considered adequate for applying multivariate techniques or SEM, which exceeded the minimum verge (ten sample sizes per parameter) for normal distributions (Kant et al., 2024). Table 2 presents the details of participants’ sample profile, indicating that the respondents are increasingly young adults (millennials and Gen X) and educated between the age group (18–44 years). Table 3 indicates that participants frequently use Google Pay, Paytm and PhonePe in that order of preference, primarily for mobile recharges, shopping and utility bill payments, respectively.

This empirical work used structural equation modeling (SEM) to test the linkage framework based on the guidelines of Anderson and Gerbing (1988) and Kline (2023). First, construct reliability, validity and dimensionality were analyzed using a measurement model through conducting confirmatory factor analysis (CFA) with AMOS 24.0. Next, the research hypotheses were tested, and the model’s predictor power and effect size were verified. Prior to running SEM, the data were cleaned using boxplot, cross-tabulation, skewness and kurtosis to scrutinize missing responses, outliers, biasness and normality-related issues. The results were adequately fit to perform multivariate analytical tools like SEM based on the methods suggested by Hair et al. (2015). Using Harman’s one-factor test, common method variance (CMV) was examined and found to be less than the 50% criterion (Podsakoff et al., 2003), indicating that CMV was not an issue. Additionally, the data were treated using the rigorous unmeasured latent method construct (ULMC) approach from the single-source, self-report design (Podsakoff et al., 2003) and assessing the possibility of CMV. The method adequately absorbed systematic variance when the original and ULMC-controlled CFA models were compared; the CMIN/DF ratio improved significantly from 2.505 to 1.771, below the threshold of 2.0 (Hair et al., 2015). Notably, the comparison indicates that the structural path coefficients’ signs and significance levels remained unchanged. This consistency implies the reported theoretical relationships are not artifactual and offers evidence that CMV did not significantly affect the study’s conclusions (Tehseen et al., 2017; Williams et al., 2010).

The measurement model was assessed using goodness-of-fit indicators, and reliability and validity measures via CFA. The goodness-of-fit indicators was adequate, namely, χ2/df = 2.50, SRMR = 0.0429, GFI = 0.910, AGFI = 0.886, IFI = 0.967, TLI = 0.961, NFI = 0.946, CFI = 0.967, RMSEA = 0.051, which satisfying the condition for unidimensional of the measurement model (Hair et al., 2015; Kline, 2023), as shown below in Table 4. Even though the AGFI value of.886 is marginally below the cut-off of.90, it is still within an acceptable range (Gefen et al., 2000), and could be caused by the model complexity and the ratio of estimated parameters to sample size (Kline, 2023). According to Hu and Bentler (1999), an acceptable fit may be indicated by an AGFI value of 0.80 or higher.

The reliability measure of Cronbach’s alpha of all latent constructs exceeded (0.886–0.923) the threshold of 0.70, which satisfies the internal consistency of construct reliability. Table 4 presents the results of composite reliability (CR), which were also larger than the recommended verge value of 0.70 (Gefen et al., 2000; Kline, 2023), thus confirming both condition of convergence consistency and internal consistency of scale reliability. Item loadings ranged from 0.748 to 0.927, which exceeded the cut-off mark of 0.50, thus satisfying the condition for construct dimensionality and scale structure (Hair et al., 2015; Jaiswal et al., 2020). The validity measure of average variance extracted (AVE) of each construct was greater than the recommended cut-off of 0.50, which accepts the satisfactory condition for convergent validity, as shown in Table 4. Similarly, Table 5 shows that the square root of AVE exceeded its shared variance, which was tested based on the recommended method by Fornell and Larcker (1981) and Kline (2023) and thus, confirms the adequate results of discriminant validity.

The structural model was estimated using predictive power (R²) and effect size (f²), followed by the goodness-of-fit measures along with hypothesized direct and indirect effects (mediation and moderation) using path analysis. The explanatory power and effect size of the research model were good enough (R² = 0.790, f² = 3.830), thus confirming that “Extended ECM” had a robust model and the strongest coefficient of determination. This indicates that 79% of the total variance of the endogenous construct of CI with extended expectation-confirmation model (EECM) is explained by its direct and indirect predictors, namely, CF, PU, EU, ST, TR, PV and SI, reported in Table 6. Hence, the results established the robust predicator power and effect size of the EECM in the present paper.

As detailed in Table 6, the goodness-of-fit indicators confirmed that the EECM achieved an acceptable level of fit. The EECM exhibited robust fit statistics (χ2/df = 3.16, SRMR = 0.0879, GFI = 0.913, AGFI = 0.884, IFI = 0.964, NFI = 0.948, CFI = 0.964, RMSEA = 0.06). The model’s overall fit is considered robust and adequate (Hair et al., 2015). While the SRMR indicator was marginally above the suggested cut-off ≤ 0.08 (Hu and Bentler, 1999), a value below 0.10 is reasonably deemed fit for SEM (Kline, 2023). AGFI value equal to or exceeding. 80 is recognized by Hu and Bentler (1999) as a criterion for demonstrating satisfactory model fit. Finally, the assessment of the structural model (EECM) verified that multicollinearity did not exist; all variance inflation factors were found to be below the crucial threshold of 5.0 (Hair et al., 2015).

The hypothesized model (EECM) was analyzed via hypotheses testing to test the direct, meditated and moderated effects using the path coefficient (β) and p-value. Table 6 shows the results of direct effects, which indicate that all the direct hypothesized relationships (H1–H10) were positively significant in the integrated model. The results revealed that the path of EU → PU (H5. β = 0.809, p  < 0.001) was extremely significant followed by the path between ST→ CI (H9. β = 0.567, p  < 0.001), EU→ ST (H6. β = 0.324, p  < 0.001) and PT→ ST (H7. β = 0.293, p  < 0.001). While the direct effect of CF on PU (H2. β = 0.071) and CF on ST (H5. β = 0.085) were the least significant in the model shown in Figure 2.

The mediating effects of “Satisfaction” (ST) and “Perceived Usefulness” (PU) were tested in the research model following the procedures of Baron and Kenny (1986), which are shown in Table 7. This is a fundamental technique (Schneider et al., 2005) and is widely used in the varied contexts of recent studies for determining mediating relationships by establishing the necessary conditions through a series of regression analyses (Biswas et al., 2021; Jaiswal et al., 2023; Kant et al., 2024). A satisfying initial condition for mediation analysis is that the predictor should be significantly correlated with both the mediator and the outcome variable. In the second stage, we examined whether the mediator and the outcome variable were significantly correlated. If the predictor and the outcome variable were not significantly correlated, this was deemed to be full mediation. The outcome variable is significantly correlated with both the predictor and the mediator, a phenomenon known as partial mediation (Schneider et al., 2005). Therefore, all of the underlying mediating paths’ results (EU-PU-ST, CF-PU-ST, PT-ST-CI and PU-ST-CI) showed and confirmed the acceptable condition for the study’s partial mediation (Table 7).

We used the interaction regression approach to assess the moderating role of perceived value (PV) to investigate the interaction effects of the proposed model (Hayes and Rockwood, 2017). As recommended by Hayes (2017) and Igartua and Hayes (2021), the analysis was carried out using the PROCESS macro with SPSS 29.0. The results for the interaction effects of PV with Satisfaction (ST) and Trust (TR) on Continuance Intentions (CI) are depicted in Table 8 and Figure 3(a)–(b).

The hypothesized moderation effects received some support from the findings (reported in Table 8). In particular, the interaction path PV × TR supported H11 by being statistically significant in predicting CI (β = −0.153, p  < 0.001). Nevertheless, this negative slope suggests an attenuating effect in the current model. The results indicate that when users perceive great value, the critical function that trust plays in creating continuance intentions is diminished. The conditional impact is statistically significant even with the small effect size (f² = 0.04) (Cohen, 2013). In contrast, H12 was rejected because the interaction impact of PV × ST on CI was not significant (β = −0.046, p  > 0.05). Additionally, the effect size (f² = 0.004) indicates that there is no significant or sustained interaction effect of PV on CI (Cohen, 2013).

As recommended by Soper (2021), we perform a multi-group analysis with SPSS 27.0 to evaluate differences in sociodemographic characteristics (gender, age, education, income, area of residence and weekly use of m-wallets). As per the suggested procedures, these variables were divided into two groups, such as “gender” into male (n = 332) and female (n = 242), “age” into <25 years (n = 310) and ≥25 years (n = 264), “education” into bachelor’s or associate’s degree (n = 215) and master’s degree and above (n = 359), “income” into ≤Rs 50,000 (n = 259) and >Rs 50,000 (n = 315), “area of residence” into urban (n = 363) and semiurban and rural (n = 211) and “weekly M-wallet usage” into up to three times (n = 136) and more than three times (n = 438) to estimate the variations in their beta (β) coefficients using t-values. As demonstrated in Table 9, the difference in t-statistic suggests that education for the ST → CI (2.037*) path and age for the PT → CI (3.090**) path significantly influenced their variations (Kant et al., 2024). However, the t-values for the other sociodemographic variables with underlying relationships have not shown significant variation in the model.

M-wallets are the perfect solution for improving financial inclusion (Herzallah et al., 2025). They reduce cash dependency and offer a safer way to conduct financial transactions online with minimal risk of fraud and theft. It can also boost financial inclusion by accommodating migrant workers who do not have bank accounts to access their salaries. For example, with m-wallet apps, workers can receive their salaries directly on the app and then use it to do transfers quickly, even without a bank account (Fosso-Wamba, 2024). The present study has adopted ECM and TAM along with other contextual variables, such as perceived trust and social influence, to get insights about users’ continuous usage intention of m-wallet. The users’ expectation confirmation was found to have a positive impact on the perceived usefulness and satisfaction of using m-wallet, which is in line with previous literature (Yousaf et al., 2021), which emphasize that meeting consumers’ expectations is crucial, more specifically in the context of new products such as m-wallet.

Perceived usefulness reported a positive association with user satisfaction and continuous usage intention of m-wallet supported the previous findings, such as Gupta et al. (2020) and Hong et al. (2017). This shows the importance of perceived usefulness if any new product has to be adopted among users on a large scale. Further, perceived ease of use and satisfaction while using m-wallets also reported a positive impact on perceived usefulness and satisfaction among users, which is in line with earlier findings in the context of new technology such as m-wallet adoption (Barry and Jan, 2018). Also, perceived trust reported a significant impact on m-wallet users’ satisfaction and continuous usage intention, which supports earlier findings (Hasan and Gupta, 2020). The regulatory framework in India, overseen by the RBI and NPCI, is continually adapting to address the evolving landscape of digital payments. These regulatory bodies implement guidelines and standards to ensure the security and integrity of digital payment systems (Chauhan et al., 2019).

Further, user satisfaction and social influence were also found to have a significant impact on the continuous usage intention of m-wallet, which is similar to previous studies (Chen and Li, 2017; Gupta et al., 2020). Finally, the relationship between perceived trust and continued usage intention is significantly moderated by perceived value. However, the moderating effect of perceived value had a negligible effect on user satisfaction with continuing usage of mobile-based payments. The results are consistent with research that has highlighted how customers’ decisions about new technologies, including m-wallets and other apps, are indirectly influenced by perceived value (Karjaluoto et al., 2019; Salah and Ayyash, 2025).

From a managerial perspective, this study offers significant insights for m-wallet apps, financial, information technologies and mobile commerce-related vendors to carefully examine the underlining factors. The study endows some valuable contributions theoretically to the existing literature on acceptance and postacceptance of using mobile-based digital payment platforms. First, our study broadens the understanding of users’ continuous usage intention of m-wallets by focusing on the broad spectrum of the construct, such as perceived trust, social influence and ECM and TAM. Second, the study has included social influence that the researchers have not focused on in the Indian context; the social influence will add a new spectrum to m-wallet continuous usage intention in collectivistic societies such as India. Third, the present study has added to the literature by examining the complex mechanism using mediation and moderation (Perceived value) to understand the users’ continuance usage intention of m-wallet, which will extend the existing literature on m-wallet adoption and usage intention.

Further, the present research offers several implications for policymakers and marketers that may help them to devise suitable strategies for better satisfaction and increased intention of using m-wallet among users. First, the marketers should ensure that while promoting the usage of m-wallet, they should never overpromise about its features, which may lead to a difference between the consumers’ expectations and confirmation after its usage, as confirmation is very crucial for users’ satisfaction and adoption of app-based services such as m-wallet.

Second, marketers should make sure that the app interface of m-wallet is easy to use, emphasizing the necessity of aligning the technology with customer knowledge, so that they can use it easily (Sarmah et al., 2021). Also, ease of use is crucial for the perceived usefulness of m-wallets and consumer satisfaction. Third, policymakers should focus on developing trust among consumers while using such services where a monetary transaction is involved, as trust is crucial for satisfaction and continuous usage intention. The companies should ensure that transaction through m-wallet is safe and secure and they follow data encryption, password-protected trade and a secure payment gateway (Gupta et al., 2020). Further, consumers should be asked about their normal transaction values, and beyond that limit, or in case of high transactional value, confirmation should be done by contacting customers. Also, consumers should be made aware through several platforms about how to use m-wallet and not click on any doubtful link, as such information will enhance consumers’ knowledge, leading to trust in m-wallet apps (Singh and Sinha, 2020). The marketers should have knowledge of users’ adoption preferences and postadoption behaviors that can support targeted promotional strategies, loyalty programs and communication campaigns that enhance engagement and retention.

Fourth, social influence was identified as crucial for users’ continuous usage intention; therefore, marketers should focus on developing a consumer community of m-wallets apps over social networking sites where users can share their experiences and concerns through blogs (Park, 2019). Such communities will create a social bond among users, and that may lead to the adoption and usage of m-wallet on a larger scale. For Gen Z, who are highly engaged in social networks and influenced by peer recommendations, SI plays a pivotal role in shaping their adoption behaviors toward digital payments (Hanafiah et al., 2024). Finally, marketers can enhance the perceived value of m-wallet by emphasizing the trust factor, reducing the risk associated with it and focusing on various aspects of customer satisfaction, as both constructs are reported to have a moderating impact on continuous usage intention of m-wallet.

The m-wallet payment landscape is categorized by rapid and ongoing alteration driven by improvements in AI- and fintech-enabled technologies, such as real-time fraud detection, algorithmic credit assessment and personalized payment interfaces. The present research has a few limitations, like any empirical research. First, the study’s cross-sectional methodology captures user adoption intentions at a single point in time, which restricts the ability to observe temporal variations in user behavior. Future research could use longitudinal designs to track the progression from behavioral intention to actual usage and sustained engagement. This approach would provide deeper insights into the dynamic nature of m-wallet adoption and how external factors (e.g. technological advancements and economic conditions) influence adoption patterns over time. Second, the study’s sample was collected via nonprobabilistic purposive sampling, which favors younger participants; therefore, the results might not sufficiently reflect the varied Indian setting. To strengthen external validity, future studies might concentrate on a balanced spread of responses from the young to the elderly. Third, the present research has focused only on the moderating influence of perceived value on various factors; future research may emphasize understanding the moderating impact of other crucial variables, such as domain innovativeness and technological readiness, among the user. Further, variations in digital infrastructure and literacy rates across regions, particularly in rural or developing areas, affect the adoption rate and effectiveness of m-wallet payments. Cultural attitudes toward technology, privacy and traditional banking differ significantly; resistance to change and trust issues with automated, nonhuman or AI-driven systems are often culturally contingent. Given that user interactions with these technologies are likely to advance over time, the results should be taken as reflecting consumer behavior at a specific stage of technological change. Future research could employ longitudinal designs or cross-country comparative studies to examine how evolving AI capabilities and cultural settings shape the m-wallet payment adoption over time. The above limitations also present opportunities for theoretical improvement by investigative how culturally embedded trust mechanisms interact with mobile payment architectures across markets.

While our quantitative approach aligns with most previous technology adoption studies in India, it may not fully capture the nuanced factors influencing m-wallet adoption. Future studies could benefit from qualitative or mixed-method approaches to explore deeper, contextual insights into user behavior. Finally, as the study employed a nonprobabilistic sampling approach, the sample may not perfectly reflect the characteristics of the broader population of m-wallet users. Consequently, the findings should be interpreted with caution, as the results may have limited generalizability beyond the specific respondent group included in the study.