Key Factors in the Continuance of Self-Service Technology and Its Mobile App Adoption—A Case Study of Convenience Stores in Taiwan ^†

Abstract

Past literature has advocated the integration of channels through the offline-to-online and online-to-offline models. However, little research has investigated the interrelationships effects between the two channels. Drawing on the literature from self-service technology (SST) and expectation–confirmation theory, this study aims to investigate key attributes of SST and assess their impact on consumer evaluations across offline-to-online and online-to-offline channels. A questionnaire survey was administered at convenience stores and 360 user responses were collected through a physical self-service kiosk. Two-stage structural equation modeling with linear structural relations (LISREL version 8.54) software was used for data analysis. The empirical results verified the attributes of physical SST (physical system’s service quality and perceived convenience) and online SST (virtual system’s service quality and perceived ubiquity) as critical antecedents of satisfaction and attitude and the subsequent behavioral intentions toward each channel. However, some transitional effects from offline (physical kiosk use) to online (kiosk app adoption) intention were not as significant as hypothesized. The offline attributes of perceived convenience and satisfaction had no significant impact on online SST significantly (kiosk app), except for the physical system’s service quality. Discussions and implications are provided, including strategies for concise functional design and essential SST services.

1. Introduction

Taiwan has the second-highest density of convenience stores in the world [1], with an average of three convenience stores per 10 km², which drives fierce competition in this industry. With increasingly diverse types of convenience store services and a limited workforce, self-service technology (SST) has become a convenient alternative to staff-provided services. In convenience store management, researchers have suggested that firms can retain customers by providing system-quality services and novel shopping experiences using SST, thus strengthening the customer–business relationship [2,3]. In-store SSTs (i.e., electronic kiosks) in convenience stores assist consumers in completing their consumption tasks independently in a time-efficient manner. As customers become more comfortable with SSTs, developing a well-designed SST strategy can strengthen customer relationships and generate profits [4,5].

Organizations today are increasingly adopting integrated approaches that bridge the digital and physical worlds, going beyond the implementation of self-service technology (SST). One notable strategy in this area is bi-directional online–offline integration models, commonly referred to as O2O (online-to-offline or offline-to-online) paradigms. This approach is exemplified by the “click and mortar” business model. Click and mortar is a term used to describe the strategic expansion of traditional brick-and-mortar operations by incorporating digital channels, which represents a significant shift in business strategy [6]. This synergistic approach enables companies to harness the advantages of tangible, in-person customer experiences while capitalizing on the expansive reach, convenience, and data-driven capabilities inherent in digital interfaces [7].

By adopting such integrated models, companies aim to create a seamless ecosystem that caters to diverse consumer preferences and behaviors, potentially leading to increased market share, improved customer satisfaction, and enhanced operational efficiency. This strategic evolution reflects the emergence of digital transformation in the business landscape as companies adapt to consumer engagement dynamics and market competition in an increasingly connected world [8]. Some researchers have concluded that channel integration promotes combined benefits. For example, one channel’s attractiveness may increase customers’ willingness to engage with another channel [9].

Understanding the interrelationship between offline and online channels and their impact on the customer-business relationship is critical in multichannel integration. However, this raises some important research questions that deserve further exploration. As in-store self-service technologies (SSTs) expand online, it is critical to examine whether consumer usage habits are evolving. It is also necessary to examine what attributes attract users across channels and how attitudes and behavioral intentions toward SSTs vary by channel. Past research has emphasized the importance of channel integration through O2O models. However, there remains a notable gap in understanding the interrelationships and effects between these two channels. This study addresses this gap by examining key attributes of SST and their influence on consumer evaluations across both channel types.

In the present study, based on earlier SST and O2O research [10], as well as the expectation–confirmation model (ECM) [11], we propose an integrated model to further investigate the interrelationships effects between physical kiosk use (offline channel) and kiosk app adoption (online channel). This proposed model is expected to answer the aforementioned questions, identify the key attributes of SSTs, and assess their impact on consumer evaluations across online and offline channels. Accordingly, this study aims to explore the key attributes of offline self-service technologies (SSTs) and their influence on consumer satisfaction and post-adoption behavior. Specifically, we examine the attributes of online SSTs and assess their impact on consumer attitudes and behavioral intentions. We also examine whether customers move from offline to online in the context of self-service technology and its mobile application (hereafter mobile app) adoption, providing insights to understand how consumers adapt to such service platforms.

2. Literature Review

2.1. Self-Service Technology (SST)

Meuter et al. [12] defined the SST as a digital interface that allows consumers to access services autonomously without direct interaction with company personnel. These platforms empower customers to conduct transactions and obtain information independently, streamlining the service process and reducing reliance on human intervention. The adoption of SST represents a significant shift in service delivery models, reflecting the growing trend toward automation and digital transformation across industries. The benefits of using SST are apparent (e.g., convenience, timesaving, and multiple locations); thus, SST is widely applied in various contexts, such as automatic teller machines in the banking industry [13], electronic kiosks in convenience stores [14], and airlines [15].

The launch of SST in various service contexts provides an alternative means for organizations to coproduce services and deliver additional value to customers. Consumer behavior and service management studies have shown a positive correlation between using self-service technologies and improved customer outcomes. Specifically, research suggests that when consumers engage with SST, they will experience higher satisfaction and loyalty to the service provider. This improved customer experience and loyalty can be attributed to factors such as the perceived control, convenience, and efficiency offered by these self-service platforms [13]. However, the use of self-service technologies (SSTs) is not always seamless. Research has shown that frustrating user interfaces and unreliable hardware, such as system failures, can negatively impact the customer experience. In addition, user unfamiliarity with the interface can lead to errors, further contributing to negative perceptions of SSTs [16]. When consumers have the opportunity to voluntarily use these technologies, they can act as co-producers of services and experience the benefits of time and cost savings [17], location convenience [18], self-efficacy [19], usefulness, and enjoyment of use [14].

Some SSTs located in physical stores for transaction purposes are called “in-store kiosks”. A leading convenience store chain in Taiwan has deployed multifunction self-service kiosks throughout its stores to provide 24-h access to various services. These in-store kiosks offer customers multiple options, including purchasing and picking up transportation tickets, booking accommodations and reservations for travelers, redeeming credit card rewards, and processing utility bill payments. This comprehensive suite of always-on services exemplifies the integration of digital conveniences into traditional retail environments. By offering these diverse capabilities, the company aims to enhance the customer experience and increase the store’s value beyond traditional merchandise sales. Although convenience stores have become a part of people’s daily lives, few studies have been undertaken of the multifunctional, high-usage kiosks in convenience stores. Based on the ECM model, Wang [14] proposed two key influencing factors (i.e., perceived control and convenience) of satisfaction with using kiosks in convenience stores. Research has investigated the impact of customers’ perceived value on their behavioral responses toward using convenience store kiosks [20], but this study focused on digital (virtual) kiosks. When digital technology evolves across channel boundaries, the trend toward brick-and-mortar integration is not limited to companies but extends to SST. SST facilitates the integration of online and offline services in various forms, such as self-checkout kiosks and mobile apps, enabling consumers to seamlessly switch between different shopping channels. For example, customers can upload documents to the cloud via an online platform and conveniently print them using SST, enhancing the overall convenience. Therefore, the present study is one of the few to examine how offline SST (i.e., kiosks) moves across channels and influences virtual SST (i.e., a kiosk app) in retail.

2.2. Integrative Model of Offline-Online Channels

2.2.1. From Online to Offline Services

With the rapid development of the Internet, most traditional brick-and-mortar companies encouraged customers to use their online services to expand their business service channels [21]. Research by Kwon and Lennon [22] suggests that offline services can attract online customers and that customers’ positive perception of a company’s offline brand image can increase their trust in its online platforms [23]. This approach can help companies to expand their services from brick-and-mortar stores to online platforms. For example, digital transaction activities in banking industries can be conducted timely and efficiently. Mobile payment can thus enhance the consumer experience [24]. Researchers reported that the advantages of mobile banking services (e.g., accounts inquiry and management, money transfer, and bill payment) have a positive impact on users’ willingness to switch to online services [25].

Conversely, when e-retail channels are extended offline, consumers can enjoy the shopping experience in a physical environment. For example, Amazon and Alibaba have launched trials of physical stores, enabling consumers to purchase using their mobile phones in the center of a showroom [26]. One of the benefits of online-to-offline retail channels is the spillover effect in terms of increased purchases [27]. For example, consumers who engage with online ads, reviews, and exclusive in-app discounts develop purchase intent. This shortens in-store decision time, increases conversion rates, and strengthens the spillover effect. Herhausen et al. [28] suggested that online and offline channel integration promotes the channel’s competitive advantage (e.g., willingness to pay) by improving perceived service quality and reducing perceived risk in online channels. Researchers have demonstrated that in-store technologies (e.g., QR codes and RFID-supported smart cards) can increase consumers’ purchase intention [29].

Accordingly, in-store SST services can play a role in encouraging consumers to engage with online kiosks. In the retail industry, when workforce service capability reaches saturation, SST provides an opportunity to serve customers. Research in the fashion industry by Flavian et al. [30] suggests that online services, such as web showrooms, have a positive impact on consumers’ purchase intentions due to their timesaving benefits. However, there is limited academic literature on the spillover effects of offline-to-online services in the retail industry, particularly in the context of convenience stores. Additionally, examining whether the synergy between online and offline services can enhance customer satisfaction and encourage continued usage or encourage customers to seek alternative channels, such as an online kiosk app, is essential.

2.2.2. Effects of Offline Satisfaction on Online Behavioral Processes

Oliver’s four-stage loyalty framework [31], which includes cognitive, affective, intentional, and active loyalty, can be applied to understand the offline-to-online behavioral transition processes across digital and physical channels. For example, a customer’s in-store experience (offline behavior) can influence their online interactions, while digital engagement (online behavior) can shape their expectations and actions in the physical store. This interplay between channels highlights the complex nature of customer engagement in the modern retail environment. Integrating these two behavior streams in a research setting is essential to understand these processes comprehensively.

Satisfaction is generally understood as an emotional response to a product or service that can shape a consumer’s overall attitude toward it [11]. This conceptualization suggests that user satisfaction and user attitude are closely related constructs. Consumer satisfaction with one channel does not always guarantee satisfaction with the other, as online and offline features may interact with or hinder each other, as highlighted by Cheema and Papatla [32]. Piercy and Archer-Brown [33] have also cautioned that failures of online services, such as higher prices, could potentially drive customers away from offline channels. Furthermore, Lu et al. [25] have argued that certain offline habits of bank customers may negatively impact their intention to transfer online. Recently, researchers have re-examined the impact of offline habits on users’ online behaviors [2,3,34]. Existing research on these questions lacks consistent conclusions. This gap motivates the need for the exploration in this study to determine whether consumers’ frequent use of self-service kiosks in convenience stores has a positive or negative impact on their behavior of utilizing online kiosks.

In the field of information technology (IT), Bhattacherjee [11] introduced the Expectation-Confirmation Model of IT Continuance (ECM-IT). This model proposes that users’ intention to continue using an IT system is primarily determined by their satisfaction with its use. Going beyond initial acceptance theories, ECM-IT emphasizes the importance of post-adoption satisfaction in predicting long-term usage behavior. This approach highlights the critical role of user experience and subsequent satisfaction in driving sustained engagement with technology. Organizations should focus on driving initial adoption and ensuring positive user experiences to drive long-term usage and loyalty. In retailing context, Wang [14] verified that consumers’ satisfaction with SST enhances their continuance use intention toward SST. The effect of satisfaction on continuance use intention has also been demonstrated in the context of mobile applications [34]. When customer satisfaction is enhanced, repeated consumption behavior becomes more frequent [11].

2.3. The Attributes of Technological Factors

2.3.1. System Service Quality

Service quality refers to customers’ perception of expected levels of service excellence and comprises factors such as interaction, the physical environment, and outcome [35]. In click-and-mortar retail environments, customers engage in both online and offline behavioral processes. In modern marketing, the in-store experience is a critical retail-related experience [36]. In-store SST can provide customer value and be an optimal strategy for providing an alternative means of accessing goods and services. When consumers operate SST (e.g., an ibon kiosk) in a convenience store, they receive services from an organization through technical facilities, not employees [37]. Therefore, the service quality of in-store SST depends on the system functions that SST provides, which we call system service quality. In the context of virtual SSTs, service quality refers to the quality of the virtual SST on the smartphone (e.g., ibon app). The virtual SST aims to present an online service with reliability, security, and value to users and customers [38].

2.3.2. Perceived Convenience

Research has emphasized the critical role of convenience in perceived service quality and in driving the evaluation of SST [39]. Armstrong et al. [40] asserted that the key attribute to retail success is retail location. A convenient location enhances consumers’ ease of access to services. In this study, the perceived convenience of SST reflects its flexibility attribute for consumers. Compared with traditional staff-assisted services in physical stores, SST offers advantages such as time flexibility, self-paced operation, and accessibility [14,18].

The convenience and accessibility of service providers play an essential role in shaping consumer perceptions and behavior. Research suggests that when customers can easily access services, they tend to view the outcomes more favorably [41]. This phenomenon extends to various service contexts, including SST. Studies have shown that geographic proximity is a critical factor in channel selection for traditional purchases, with distance negatively affecting usage [42]. In the context of SST, location becomes even more crucial. When SST is in congested areas, users must consider additional environmental factors, potentially diverting their attention and energy from using the technology. Convenience has emerged as a key driver of customer satisfaction in service encounters. When consumers perceive SST as providing convenience benefits, they are more likely to report higher satisfaction levels [39]. These benefits often include increased transaction efficiency and effectiveness, reduced waiting time, and reduced physical and mental effort. The positive impact of convenience on service evaluations goes beyond satisfaction. It can influence customers’ overall perceptions of service quality, their likelihood to use the service again, and their willingness to recommend the service to others. Therefore, service providers should prioritize convenience when designing and implementing SST to improve customer experience and drive positive outcomes.

2.3.3. Perceived Ubiquity

Perceived ubiquity incorporates the concepts of spatial, temporal, and mobility [43]. It can be defined as ready access to products or services at any time and in any place [44]. With more people worldwide owning mobile phones and the rapid development of mobile and wireless technology, perceived ubiquity is becoming increasingly popular. Studies have reported that one significant advantage of an online channel is its facilitation of product purchases from any location [45]. Consumers can benefit from using mobile apps, through which they can access a service whenever and from wherever they wish. Thus, the Internet has become a vital alternative to completing tasks with reduced time, money, and effort [46]. In this study, perceived ubiquity is considered a key factor of virtual SST because consumers can operate its functions through smartphones or websites without time and space constraints.

2.4. Research Framework and Hypothesis Development

Research has reported that in the multichannel context, positive evaluations of the primary (i.e., offline) channel are likely to serve as a reference for assessing the online channel [47]. Furthermore, integrating online and offline channels has generated competitive advantages, as evidenced by consumers’ increased willingness to purchase across both platforms [28]. We argued that if consumers have positive evaluations of an offline channel (i.e., physical SST) and exhibit continued usage, they tend to adopt the alternative channel (i.e., virtual SST). Thus, in this study, we proposed the following hypothesis:

H1: 

The continuance intention of the kiosk has a positive impact on the adoption intention of the kiosk app.

Consumer satisfaction and attitude have been studied in two parallel IT studies [38]. Based on Oliver’s four-stage loyalty framework [31] and extending attitude–behavior theory, satisfaction and attitude are regarded as powerful predictors of continuance use intention. The relationship between consumer satisfaction and post-purchase behavior has been a significant focus of service marketing research, particularly within expectation-confirmation theory (ECT) [31,48,49]. This association has received consistent support concerning both online and offline environments. Therefore, in this study, the following hypotheses were proposed:

H2: 

Satisfaction with the kiosk has a positive impact on (H2a) continuance intention of the kiosk and (H2b) adoption intention of the kiosk app.

H3: 

Attitude toward the kiosk app has a positive impact on the intention to adopt it.

In this study, system service quality refers to the extent to which the systemic functions of physical or virtual SSTs can help consumers obtain products or services effectively. Regarding the integration of offline SST (physical kiosk) and online SST (kiosk app) in retail, the system service quality of the functional kiosk interface can more fully and accurately explain users’ satisfaction with SST use [50,51]. If the system service quality of SST exceeds expectations, customer satisfaction increases [50], which in turn promotes the positive impact of the online channel. In the Internet context, the positive effect of service quality on satisfaction was verified by previous research [52]. Studies have indicated that offline service quality positively impacts online satisfaction [51]. Therefore, in this study, the following hypotheses were proposed:

H4: 

Perceived system quality of the kiosk positively impacts (H4a) satisfaction with the kiosk and (H4b) attitude toward the kiosk app.

H5: 

Perceived system quality positively impacts the attitude of the kiosk app.

Moreover, according to consumer efficiency theory [53], consumers favor channels that minimize various costs, such as monetary costs. In a multichannel environment (i.e., online and offline channels), offline purchase decisions sometimes motivate consumers to consider the costs associated with reaching physical stores [54]. Lu et al. [25] indicated that certain offline habits negatively affect consumers’ transfer usage from an offline to an online channel. Studies have also reported that the farther away the store is from the consumer, the more likely the consumer is to make purchases online. For example, Chintagunta et al. [55] discovered that consumers purchase groceries online to avoid traveling distances to reach stores. Conversely, when consumers perceive shopping at nearby stores as convenient, they are less likely to use an online channel. Shaw et al. [56] observed the influence of ubiquity on the intention to use mobile apps, such as digital wallets. Following these research findings, we posited that consumers’ experience of a degree of ubiquity is likely to increase their positive attitude toward using an online SST. Accordingly, SST’s critical offline and online attributes, specifically perceived location convenience and ubiquity, can also predict user satisfaction and attitude [18]. Thus, we proposed the following hypothesis:

H6: 

Perceived convenience of the kiosk positively impacts (H6a) satisfaction with the kiosk and (H6b) attitude toward the kiosk app.

H7: 

Perceived ubiquity positively impacts attitude toward the kiosk app.

Based on the above studies, our proposed theoretical framework for understanding the adoption of SST in retail settings emphasizes the interaction between offline and online platforms. It suggests that consumers’ intention to use SST is primarily driven by two key factors: satisfaction and attitude. These, in turn, are influenced by three critical perceptions: system service quality, convenience, and ubiquity. Considering offline and online SSTs, the framework covers the increasingly blurred boundaries between physical and digital retail experiences. The seven hypotheses (H1–H7) related to the conceptual framework are illustrated in Figure 1.

3. Methodology

3.1. Data Collection

To test the above hypotheses, this study takes the example of a prominent convenience store chain in Taiwan, 7-Eleven (headquartered in Xinyi District, Taipei City, Taiwan). This industry leader has pioneered a multifaceted service delivery approach encompassing physical, digital, and mobile platforms. They have implemented a multidimensional service system, starting with in-store kiosks (e.g., ibon) for convenient shopping, followed by an online marketplace. Recognizing the growing importance of mobile technology, the company launched its mobile app for smartphone use in 2014, as shown in Figure 2. The app was designed to enhance customer convenience and reduce transaction times, in line with the consumer demand for efficient on-the-go services. By combining online and offline services, retail companies can encourage offline customers to become online users. Users can confirm various services (such as ticket purchasing, payment, printing, and scanning services) using their mobile devices and then visit a nearby store to complete the transaction. Additionally, online membership registration facilitates the monitoring of customers’ transaction profiles and the timely communication of promotional offers to consumers, ultimately saving operational costs [57]. This strategy demonstrates how companies can leverage multiple channels to create a comprehensive service ecosystem, potentially improving customer engagement and operational efficiency [11].

The researchers of this paper administered the survey during scheduled class sessions, having obtained prior authorization from both students and faculty members. Research has shown that young people make up the majority of heavy users of mobile applications and other technologies [34], so this study considers student use to be appropriate. Prior to participation, each respondent provided informed consent. The questionnaire commenced with an overview of the research objectives and explicitly emphasized participant anonymity, stating that personal identifiers such as names and institutional affiliations were not required. The initial survey item presented the following query: Are you willing to provide informed consent and participate in this survey? Once participants agreed to take the survey, they could move on to the next questions. The paper-based survey took approximately 20 min to complete, and participants were thanked upon completion.

This study used a paper-based questionnaire to collect empirical data from actual ibon kiosk users in Taiwan. The choice of Taiwan as a research setting is particularly relevant due to its exceptionally high density of convenience stores, which provides a valuable context for studying the use of self-service technology. The questionnaire was developed by adapting and modifying previously validated scales [11,14,55,56,58,59,60,61,62,63] to ensure reliability and validity. To identify participants who were familiar with the ibon kiosks but not the mobile app, the survey asked if they had used an ibon kiosk at 7-Eleven and if they had downloaded the app. The design of this question was to identify participants who had experience with the physical kiosks but were new to its mobile app. The two screening questions helped identify participants who had experience with the physical ibon kiosks but had not yet engaged with the mobile app version. The study collected 380 questionnaires, of which 360 were fully completed for analysis. This resulted in a high response rate of 94.7%, indicating strong participant engagement.

Regarding the sample distribution in this study, male and female participants were 55.8% and 44.2%, respectively. In terms of age distribution, 80.3% of the respondents were in the 20–25 age group. This pattern is similar to a survey of university students who borrow books from convenience stores in northern Taiwan [2]. Such characteristics should be considered when interpreting the results of the study and considering their generalizability to broader populations. The high completion rate and focused demographics may provide valuable insights into the self-service technology preferences and behaviors of young adult consumers in the Taiwanese convenience store environment. Regarding consumer behaviors, more than half of the respondents visited a convenience store every day, and approximately 70% used an ibon kiosk at least once a week.

Table 1. Sample characteristics (N = 360).

Characteristics	Number	Percentage (%)
Gender
Male	201	55.8
Female	159	44.2
Age
Under 19	54	15.0
20–25	289	80.3
26–30	11	3.1
Over 31	6	1.7
How often to 7–11?
1–2 days	199	55.3
3–4 days	92	25.6
5–6 days	37	10.3
Over 7 days	32	8.9
Number of times to use ibon (per week)
Less than once	109	30.3
1	185	51.4
2	30	8.3
More than 3 times	36	10.0

summarizes the demographic characteristics of the respondents.

3.2. Research Instrument

All constructs were measured using established scales from previous research adapted to the current context using a 7-point Likert scale, with responses ranging from “strongly disagree” (1) to “strongly agree” (7). Measures were obtained as follows:

• Service quality: five items derived from Jiang et al. [58] and Negash et al. [59].
• Perceived convenience: four items based on Collie et al. [60].
• Perceived ubiquity: four items based on Mallat et al. [61].
• Satisfaction: three items adapted from Wang [14].
• Attitude: three items adapted from Davis et al. [62].
• Offline continuance intention: three items adapted from Bhattacherjee [11].
• Online adoption intention: three items adapted from Venkatesh et al. [63].

The instrument was translated from English to Chinese by an independent translator and then back to English by another translator. Any discrepancies between versions were thoroughly reviewed by additional translators to ensure consistency and accuracy of meaning across languages. This instrument development and translation approach aimed to increase the validity and reliability of the measures while ensuring their relevance to the Taiwanese context [2,3].

4. Results

This study used a survey method for data collection, followed by a two-stage structural equation modeling (SEM) approach for data analysis, as recommended by Anderson and Gerbing [64]. The analysis was conducted using LISREL software and consisted of two distinct phases. First, a measurement model was used to assess the reliability and validity of the constructs.

Table 2. Standardized loadings and reliability.

Construct	Loading
Physical System Service Quality (PSQ)
Adapted from Jiang et al. [55]; Negash et al. [56]; α = 0.95; CR = 0.95; AVE = 0.79.
PSQ1. The ibon kiosk interface enables me to quickly operate various functions.	0.83
PSQ2. The ibon kiosk and its various services fully meet my needs.	0.87
PSQ3. The ibon kiosk provides me with a variety of alternatives for solving my problems.	0.95
PSQ4. The ibon kiosk has a user-friendly interface.	0.91
PSQ5. The ibon kiosk provides prompt service to users.	0.89
Perceived Convenience (PC)
Adapted from Collie et al. [60]; α = 0.93; CR = 0.93; AVE = 0.76.
PC1. Having an ibon kiosk in a convenience store that enables me to easily initiate a transaction is important to me.	0.91
PC2. A convenient location makes me feel more comfortable using ibon kiosk services.	0.93
PC3. The location of the convenience store affects my decision to use its ibon kiosk.	0.75
PC4. Using an ibon kiosk makes my service transaction less time-consuming.	0.89
Virtual System Service Quality (VSQ)
Adapted from Jiang et al. [58]; Negash et al. [59]; α = 0.95; CR = 0.94; AVE = 0.75.
VSQ1. The ibon kiosk interface enables me to quickly operate various functions.	0.88
VSQ2. The ibon kiosk and its various services fully meet my needs.	0.92
VSQ3. The ibon kiosk provides me with a variety of alternatives for solving my problems.	0.92
VSQ4. The ibon kiosk has a user-friendly interface.	0.90
VSQ5. The ibon kiosk provides prompt service to users.	0.68
Physical System Service Quality (PSQ)
Adapted from Jiang et al. [58]; Negash et al. [59]; α = 0.95; CR = 0.95; AVE = 0.79.
PSQ1. The ibon kiosk interface enables me to quickly operate various functions.	0.83
PSQ2. The ibon kiosk and its various services fully meet my needs.	0.87
PSQ3. The ibon kiosk provides me with a variety of alternatives for solving my problems.	0.95
PSQ4. The ibon kiosk has a user-friendly interface.	0.91
PSQ5. The ibon kiosk provides prompt service to users.	0.89
Perceived Convenience (PC)
Adapted from Collie et al. [60]; α = 0.93; CR = 0.93; AVE = 0.76.
PC1. Having an ibon kiosk in a convenience store that enables me to easily initiate a transaction is important to me.	0.91
PC2. A convenient location makes me feel more comfortable using ibon kiosk services.	0.93
PC3. The location of the convenience store affects my decision to use its ibon kiosk.	0.75
PC4. Using an ibon kiosk makes my service transaction less time-consuming.	0.89
Virtual System Service Quality (VSQ)
Adapted from Jiang et al. [58]; Negash et al. [59]; α = 0.95; CR = 0.94; AVE = 0.75.
VSQ1. The ibon kiosk interface enables me to quickly operate various functions.	0.88
VSQ2. The ibon kiosk and its various services fully meet my needs.	0.92
VSQ3. The ibon kiosk provides me with a variety of alternatives for solving my problems.	0.92
VSQ4. The ibon kiosk has a user-friendly interface.	0.90
VSQ5. The ibon kiosk provides prompt service to users.	0.68
Perceived Ubiquity (PU)
Adapted from Mallat et al. [61]; α = 0.93; CR = 0.93; AVE = 0.78.
PU1. Using ibon app services on my mobile phone reduces queuing time.	0.85
PU2. Using ibon app services on my mobile phone is independent of time.	0.91
PU3. Using ibon app services on my mobile phone is independent of place.	0.91
PU4. Using ibon app services on my mobile phone is convenient because I usually have my phone with me.	0.85
Satisfaction (SA)
Adapted from Wang [14]; α = 0.95; CR = 0.94; AVE = 0.75.
SA1. I am satisfied with the ibon kiosk installed at my convenience store.	0.87
SA2. The ibon kiosk installed by the firm exceeded my expectations.	0.70
SA3. My experience of using ibon kiosk has been satisfactory.	0.83
Attitude (AT)
Adapted from Davis et al. [62]; α = 0.93; CR = 0.94; AVE = 0.85.
AT1. I think that using the ibon app is a good idea.	0.92
AT2. I think using the ibon app is beneficial to me.	0.94
AT3. I have a positive perception of using the ibon app.	0.90
Continuance Intention (CI)
Adapted from Bhattacherjee [11]; α = 0.95; CR = 0.95; AVE = 0.86.
CI1. I intend to continue using ibon kiosk for service transactions in the future.	0.93
CI2. I will regularly use ibon kiosk for service transactions in the future.	0.94
CI3. I intend to use ibon kiosk in my daily life.	0.91
Adoption Intention (AI)
Adapted from Venkatesh et al. [63]; α = 0.95; CR = 0.95; AVE = 0.87.
AI1. I intend to use the ibon app in the future.	0.92
AI2. I intend to use the ibon app in my daily life.	0.94
AI3. I plan to use the ibon app frequently.	0.94

Note: α = Cronbach’s alpha; CR = Composite reliability; AVE = Average variance extracted.

presents the criteria used to assess the reliability and validity of the instrument. Second, a structural model was used to examine the relationships among the eight research constructs. This approach allowed for a thorough evaluation of the measurement quality and the theoretical relationships proposed in the model.

4.1. Measurement Assessment

The study used confirmatory factor analysis (CFA) to assess the reliability and validity of the instrument (Table 2). Results indicated a good fit between the measurement model and the data, with acceptable goodness of fit indices: χ²/d.f. = 2.57 (970.58/377), NFI, NNFI, and CFI all at 0.98, IFI at 0.99, GFI at 0.85, RMR at 0.063, and RMSEA at 0.061. Composite reliability (CR) was used to assess internal consistency among the research constructs and measurement items, while convergent and discriminant validity were examined to ensure construct validity. CR values ranged from 0.84 to 0.95, exceeding the 0.7 threshold and demonstrating robust reliability. In addition, Cronbach’s alpha coefficients (0.85 to 0.95) exceeded the recommended level of 0.7, confirming the internal consistency of the research items [65].

Then, we employed convergent and discriminant validity analyses to assess the validity of our research constructs. Convergent validity was established through two methods. First, the average variance extracted (AVE) for each construct ranged from 0.65 to 0.87 (on-diagonals,

Table 3. Descriptive statistics, variance explained, and correlations.

	Means	S.D.	PSQ	PC	VSQ	PU	SA	AT	CI	AI
PSQ	5.23	1.29	0.79	0.53	0.38	0.32	0.50	0.35	0.55	0.25
PC	5.38	1.36	0.73	0.76	0.24	0.26	0.62	0.24	0.67	0.12
VSQ	4.83	1.26	0.62	0.49	0.75	0.53	0.24	0.58	0.26	0.34
PU	4.80	1.33	0.57	0.51	0.73	0.78	0.23	0.56	0.25	0.29
SA	5.09	1.21	0.71	0.79	0.49	0.48	0.75	0.22	0.67	0.15
AT	4.98	1.30	0.59	0.49	0.76	0.75	0.47	0.85	0.22	0.45
CI	5.35	1.42	0.74	0.82	0.51	0.50	0.82	0.47	0.86	0.13
AI	4.70	1.47	0.50	0.35	0.58	0.54	0.39	0.67	0.36	0.87

Note: S.D. = standard deviation. On-diagonals are the average variance extracted (AVE) (as indicated in boldface). The lower triangle is the correlation coefficient, and the upper triangle is the square of the correlation coefficient. PSQ = physical system service quality; PC = perceived convenience; VSQ = virtual system service quality; PU = perceived ubiquity; SA = satisfaction; AT = attitude; CI = continuance intention; AI = adoption intention.

), exceeding the critical value of 0.5 [66]. This indicates that the items within each construct effectively explained a substantial portion of the variance. Second, all factor loadings within each construct were statistically significant at the 0.01 level, with t-values ranging from 14.89 to 23.57. This further supports the convergent validity of our measurement model. Discriminant validity was assessed using a correlation matrix. The AVE for each construct (diagonal values) was more significant than the square of the correlation coefficient between constructs (upper triangle values) [67,68]. Additionally, the values did not exceed 0.85 [69]. These findings demonstrate that each construct measures a distinct aspect of the construct domain, minimizing overlap and ensuring that other constructs do not unduly influence them in the model. The result shows that our measurement model exhibits sufficient reliability and validity.

All measures were self-reported by the same respondents, thereby generating potential common-method bias. Two tests were conducted to examine potential common-method bias following the recommendation of Podsakoff et al. [70]. First, the Harmon single-factor test was employed to verify that the data and constructs were unaffected by common-method bias. We performed factor analysis and determined that the most significant extracted factors contributed to 45% of the total variance extracted, which is less than the threshold value of 50%. This result verified that the research data did not exhibit common-method bias. Moreover, concerns about common-method variance were raised because the same respondents self-reported the items. According to Kock and Lynn [71], a full collinearity test was used to assess a potential common-method bias. The variance inflation factors for all constructs (between 1.90 and 3.52) were less than 5 [72], indicating that all variables were distinct and that multicollinearity was absent. In addition, the test employing the common latent factor in CFA was performed [70], with the results revealing the fit of the common latent factor model (χ² = 12,212.46 on 405 d.f. and RMSEA = 0.285), with the results demonstrating the fit of the common latent factor model (χ² = 970.58 on 377 d.f. and RMSEA = 0.061). The difference in standardized regression weights between these two models was less than 0.05 for all items. Based on these results, we concluded that common-method bias was not a concern in this study.

4.2. Results of Structural Model

After verifying the measurement model, we examined the structural model, which showed an acceptable fit to the data (χ²/d.f. = 2.54 (989.71/389, p < 0.001). Goodness-of-fit indices further supported this conclusion: NFI = 0.96, NNFI = 0.98, CFI = 0.99, IFI = 0.99, GFI = 0.84, RMR = 0.033, and RMSEA = 0.066. These results indicate that the model accurately represents the relationships among the variables in the study. The result of standardized path coefficients also supported seven of the ten hypotheses. Consumers’ satisfaction with the ibon SST significantly influenced their intention to continue using the technology (β = 0.94, p < 0.001), supporting H2a. However, these factors did not predict intention to adopt the ibon app, rejecting H1 (β = −0.13, p > 0.05) and H2b (β = 0.18, p > 0.05). A significant relationship between user attitude and intention to adopt the ibon app (β = 0.66, p < 0.001) supported H3. The perceived system quality of the ibon SST significantly influenced user satisfaction (β = 0.24, p < 0.001) and attitude (β = 0.14, p < 0.05) toward the ibon app, supporting both H4a and H4b. However, the relatively modest effect size for H4b suggests that system quality may play a less substantial role in shaping attitudes compared to other factors in the proposed model. While system quality contributes to attitude formation, other variables likely exert stronger influences on users’ attitudes toward the mobile app. Future research should investigate whether factors such as perceived usefulness might have a more direct impact on attitude formation. This is because users may prioritize functional benefits over technical system performance. Additionally, perceived ease of use could potentially be a stronger determinant of attitudes toward mobile app, as intuitive interfaces and operational simplicity might be more immediately apparent to users than backend system quality. Perceived convenience significantly affected consumer satisfaction (β = 0.74, p < 0.001), supporting H6a, but did not affect attitude toward the ibon app (β = −0.12, p > 0.05), rejecting H6b.

Finally, regarding the virtual attributes of the ibon SST, consumers’ perceived system quality and perceived ubiquity significantly affected their attitude toward the ibon app, supporting both H5 (β = 0.44, p < 0.001) and H7 (β = 0.38, p < 0.001). In addition, the research framework determined 80%, 74%, 85%, and 50% of the observed variance regarding users’ satisfaction, attitude, continuance intention, and adoption intention, respectively. The results revealed that the applied research model can satisfactorily explain endogenous variables. The results of the structural model are summarized in

Table 4. Hypothesis-testing result.

Hypothesis	Path	β	t-Value	Result
H1	CI → AI	−0.13	−0.73	not significant
H2a	SA → CI	0.94 ***	22.34	supported
H2b	SA → AI	0.18	1.00	not significant
H3	AT → AI	0.66 ***	12.29	supported
H4a	PSQ → SA	0.24 ***	5.13	supported
H4b	PSQ → AT	0.14 *	2.33	supported
H5	VSQ → AT	0.44 ***	7.47	supported
H6a	PC → SA	0.74 ***	12.93	supported
H6b	PC → AT	−0.02	−1.13	not significant
H7	PU → AT	0.38 ***	6.62	supported

Note: * p < 0.05; ** p < 0.01; *** p < 0.001.

and Figure 3.

5. Discussion

5.1. Key Findings and Some Unsupported Hypotheses

This study aimed to deepen the understanding of the attributes of physical and virtual SSTs that influence both online and offline behavioral processes simultaneously and the transitional effects from offline to online regarding the usage of SST. The applied model exhibited satisfactory model fit indices and a high percentage of explanatory power, and seven of the ten hypotheses were supported. The key findings and some unsupported hypotheses are discussed.

First, we identified two critical attributes of physical SST, namely system service quality and perceived convenience, which have also been analyzed in related studies [14,18]. Consistent with previous studies, these two factors considerably influence consumers’ satisfaction [39,51,55], which in turn affects their subsequent usage. Moreover, all the hypothetical paths in the online environment were also supported. Two key attributes of virtual SST, namely system service quality and perceived ubiquity, were also identified. This finding is consistent with previous research [18,51,56]. Based on the results, we confirmed that system service quality and perceived ubiquity positively affect consumers’ attitudes toward virtual SST, affecting their intention to download and use the ibon app.

Second, in terms of the transition effect from offline to online, only one hypothesis was supported (i.e., the impact of perceived physical system service quality on attitude toward virtual SST), which echoes the research of Montoya-Weiss et al. [47]. This result indicates that consumers’ positive evaluation of the service quality of SST enhances their favorable attitude toward virtual SST. Specifically, it was observed that consumers’ positive assessment of the service quality of SST enhances their favorable attitude toward virtual SST. Based on the literature [51], this outcome implies that companies that prioritize enhancing service quality and functionality of in-store (offline) SST are more likely to enhance consumer satisfaction and foster loyalty among their customers, providing contributions to the SST literature.

Third, further discussion of the weak effect (H4b) and three hypotheses that were not supported from offline to online (H6b, H2b, and H1) are provided. Regarding H4b, the results support that the perceived system quality of the ibon SST influenced the attitude towards the ibon app. However, the relatively weak effect (t = 2.33) suggests that system quality may play a less substantial role in attitude formation compared to other factors in the proposed model. While system quality contributes to attitude formation, other variables are likely to have stronger influences on users’ attitudes toward the ibon app. Future research should investigate whether factors such as perceived usefulness have a more direct impact on attitude formation, as users may prioritize functional benefits over technical system performance. In addition, intuitive interfaces and ease of use may be more immediately apparent to users than back-end system quality in influencing user attitudes toward the ibon app.

Next, we found that consumers’ perceived convenience of physical SST locations failed to enhance their attitude toward online SST (H6b). Some possible explanations are provided, including user familiarity, digital resistance, and different interaction expectations between physical and digital SSTs. User familiarity with physical SSTs may create a comfort zone. Consumers who are used to interacting with ibon kiosks may evaluate the ibon app independently. Digital resistance also plays a role from another perspective, as consumers who prefer tangible interactions may maintain barriers to mobile alternatives regardless of the convenience of virtual SSTs. For example, some users may still prefer the immediate feedback and tangible interface of physical devices to digital alternatives. In addition, different interaction expectations between physical and virtual SSTs are likely to influence this relationship. While location convenience drives the use of physical SSTs, mobile applications are evaluated based on interface design and functionality rather than geographic proximity. This may reflect that consumers’ evaluation criteria for physical and mobile SSTs are in separate cognitive frameworks. This also explains why convenience in one domain does not necessarily translate into positive attitudes in the other.

Consumers’ satisfaction with physical SST (H2b) and continued use of physical SST (H1) had no significant effect on consumers’ willingness to download and use virtual SST. Based on the argument from the habit theory in the technology adoption literature [25,35], we inferred that one possible reason is the usage habits of consumers. This implies that offline habits may negatively affect online transfer intentions. In this sense, lack of habit formation may be the primary reason for this disconnect. Despite satisfaction with physical kiosks, users may not have developed sufficient habitual engagement to motivate the adoption of complementary digital channels. For example, occasional SST users may be satisfied with their infrequent interactions but lack the behavioral motivation to extend their usage to a mobile app. Next, a perceived lack of need could explain this result; consumers may view the mobile app as redundant if their service needs are already adequately met by physical SSTs. Finally, resistance to channel switching could also be a reason. Even satisfied physical SST users may be reluctant to switch to virtual alternatives due to switching costs, learning curves, or a general aversion to technological change. This resistance manifests itself as inertia. Despite being satisfied, consumers continue to use familiar physical interfaces because the perceived effort of adopting a new technology outweighs the potential benefits of the mobile app. Perspective from consumer efficiency theory [53] provides similar explanations as consumers favor channels that can reduce their costs, such as Internet search costs. Taiwan has a high density of 24 h convenience stores. Consumers are already accustomed to using convenience stores for various services and purchases, including physical SST. When consumers are unsure about how to use SST, some may prefer to seek assistance from staff immediately. In such cases, the convenience and satisfaction derived from using offline SST may hinder them from adopting online SST.

5.2. Theoretical and Practical Implications

This study offers theoretical contributions by providing insights into the SST and O2O integration model from three aspects. First, this study extends the application of the ECM-IT model to the field of SST and O2O research in a retailing context. Critical factors, such as service quality and perceived location convenience of SST, were identified as having substantial impacts on satisfaction with continuing use of physical services. Second, this study adopted O2O integration models to explore the characteristic variables of online and offline SST and examine their respective contributions to consumers’ behavioral intentions based on consumers’ satisfaction and attitude. This study indicated that online service quality and perceived ubiquity are reliable predictors of online attitude, which in turn affects their intention to adopt virtual SST. Third, this study further examines the transitional effect of SST from offline to online usage. This result suggests that retailers can convert occasional consumers into online members and have access to their customers anytime. Moreover, companies can use the service quality of the SST system as a means to simultaneously affect the evaluation, emotion, and attitude of consumers in online and offline environments.

Based on the results, several practical implications for retail managers are provided. First, perceived location convenience is a critical factor influencing consumer satisfaction toward SST. This is evidenced by the high density of convenience stores in major cities worldwide, many of which are strategically situated on main thoroughfares to enhance patronage and foster customer loyalty. Beyond the functional aspects of SST, retailers recognize that positioning these technologies in high-traffic, convenient areas is an effective strategy for increasing usage, expanding the store’s service area, and ultimately improving customer satisfaction.

Second, both offline and online system service quality constructs have been identified as significant factors in satisfaction and attitude. In this sense, providing a more consistent functional design can reduce user confusion and learning barriers across channels for extending services from offline to online. Consumers’ offline usage habits (e.g., being accustomed to completing transactions using SST in conveniently located brick-and-mortar establishments) can reduce the perceived necessity of downloading and using an app. However, emphasizing the ubiquity of mobile services, allowing consumers to enjoy services anytime and anywhere, increases consumers’ positive attitude toward and willingness to use digital services. Furthermore, this study further verified that consumer satisfaction increases consumers’ post-purchase behavior and loyalty. However, it suggests that satisfaction with services experienced in physical stores may not necessarily lead to a positive attitude and willingness to use virtual services. Therefore, in addition to improving offline consumer satisfaction, companies must provide more benefits to using online services, such as shopping discounts, shipping discounts, and lottery draws.

Regarding the transition effect, our results show that it is not significant for younger consumers in the context of SST use. This finding highlights the need for further investigation into the factors influencing consumers’ adoption of online SST channels and the potential challenges in transitioning from offline to online services in convenience stores. First, the finding challenges previous perceptions: younger consumers are generally more tech-savvy and will naturally adopt online SSTs after using offline SSTs. However, other factors such as habit formation, perceived ease of use, or situational preferences may play a more important role in determining whether consumers switch from offline to online SSTs. Moreover, when in-store staff are asked questions about the operation of SSTs, they should patiently explain the operation of SSTs to unfamiliar and elderly customers to encourage consumers to become more familiar with the interface operation of SSTs and switch to using online services.

Regarding the transition effect, our results show that it is not significant for younger consumers in the context of SST use. This finding highlights the need for further investigation into the factors influencing consumers’ adoption of online SST channels and the potential challenges in transitioning from offline to online services in convenience stores. The finding challenges previous perceptions: younger consumers are generally more tech-savvy and will naturally adopt online SSTs after using offline SSTs. Offline-to-online transitions are not as straightforward as commonly assumed in the service design literature. Even when targeting younger demographics, companies cannot rely solely on physical SST exposure to drive mobile app adoption. Several barriers may explain this disconnect. For example, digital literacy barriers may exist even among younger users, who may understand basic technology but struggle with specific applications or features of the virtual SST. Motivational issues may arise if the mobile alternative does not demonstrate clear advantages over physical kiosks, and service design issues may create friction in the transition process. For example, a convenience store customer may be familiar with in-store kiosks for printing services but may find the mobile app interface confusing or see no value in using it remotely. To address these challenges, companies should consider implementing cross-channel incentives that specifically reward switching behavior and develop clearer value propositions that articulate the unique benefits of mobile channels. In addition, in-store staff should be trained to assist older customers and highlight specific mobile benefits to tech-savvy users who may not recognize the full benefits of virtual SST despite their general comfort with technology.

Finally, we argue that this study holds significant value by addressing an important research problem through the lens of human behavior and decision-making. Specifically, our findings can inform practitioners and policy-makers about consumer preferences and barriers to adoption, which are critical for guiding the development and effective implementation of SST solutions. Hence, the scientific value of this work lies in its contribution to consumer behavior and social science research, forming a basis for further interdisciplinary collaborations. We respectfully suggest that this perspective aligns with the broader scope of knowledge advancement in this area. Some implications for offering SST services in both on- and offline channels are suggested. For example, SST services such as mobile printing, bill payment, and delivery services are essential for customers. This implied that retail managers should recognize the significance of these services and ensure that their SST systems in both on- and offline channels meet the necessary standards to provide customers with a positive experience. In addition, developing functions to support people in their daily lives and simplifying the interface operation are also critical, especially for offline SST use in the physical store. As such, strengthening on-the-job training for employees will enable them to assist customers unfamiliar with SST’s operating procedures and functions. In addition to convenience stores, other retail operators can refer to the successful case of the ibon kiosk when establishing SST, thereby attracting more customers to use SST and expanding the scope of service. It also suggests that companies should consider seamless integration between the on- and offline channels to improve the overall consumer experience. These insights can help companies optimize resource allocation and effectively address consumer needs across both platforms. These considerations can help retailers optimize resource allocation and effectively address consumer needs across both platforms.

5.3. Limitations and Recommendations for Future Research

Some limitations and research directions for future studies are noted as follows. First, this study focuses on SST in convenience stores, and therefore the results cannot be fully generalized to other types of SST in the retail industry. This is because most SST functions are relatively simple, such as beverage vending machines. However, SSTs in convenience stores provide a variety of functions. For some specific functions (e.g., mobile data purchases), they are relatively more complex to operate than a self-service vending machine. To further test the generalizability of this study, future research can apply this research model to examine the use of SSTs in other retail industries. Second, the main idea of SST in convenience stores is to serve customers of all ages. In this study, we focused on university students’ perceived values of using SST in the context of convenience stores. More specifically, the empirical data of this research were collected from universities in the North Taiwan metropolitan area. Taiwan has the second highest density of convenience stores in the world, especially in urban areas; they can be found on almost every street, especially in large cities in northern Taiwan. For future research, data collection could be expanded to include students in post-initial programs and students in other regions (e.g., suburbs) or counties to obtain a more stratified sample. Future studies could be conducted to validate the current findings. It is also suggested that consumers of other age groups be included to determine their online behaviors. Finally, the transitional effect of offline to online use of SST was not reflected in the perceived convenience, satisfaction, and continued use of in-store physical SST. About 30% of respondents used ibon less than once a week, which may have also affected their willingness to use the app. Therefore, follow-up research could separately examine the transitional effect of high-frequency and low-frequency users from offline to online to verify and compare the conclusions of this study.

6. Conclusions

This study integrated the ECM-IT and the models of SST and O2O and O2O to provide a theoretical framework for SST research in both online and offline contexts. In this study, we identified critical attributes of physical and virtual SSTs. Based on the results, we confirmed the effects on attitude/satisfaction and subsequent behavioral intention. We also discussed the link between offline and online SST use. Although most relevant studies have advocated the implementation of O2O integration or multichannel strategies, our study is one of the few to raise questions about the transition process between offline and online. That is, the salient characteristics of offline services may not necessarily lead to positive consumer attitudes toward online services. In addition, satisfaction with offline services may not be directly transferred to online services. It should be noted that some hypothetical transition paths were not confirmed in this study. Future studies are suggested to collect more samples to test the proposed model in different contexts of SST use. For some new digital services, further studies are also suggested to investigate customers’ intentions toward using SST’s new functions, such as Internet cards (SIM/e-sim) service. Such considerations are crucial and can enable companies to capitalize on satisfaction incentives and avoid satisfaction inhibitors when driving consumers to multichannel services. The findings of this study bridge the gaps in the SST and offline-to-online literature and offer a new perspective for studying this topic. This study also advanced our understanding of how and why an offline channel differs from an online channel, providing some managerial and theoretical implications.