Maintaining continuers vs. converting discontinuers - Semantic Scholar

108

Int. J. Mobile Communications, Vol. 6, No. 1, 2008

Maintaining continuers vs. converting discontinuers: relative importance of post-adoption factors for mobile data services Hoyoung Kim Chungjeong-ro 3-Ga, Seodaemun-gu, Seoul 120-709, Korea Fax: (+822)-2014-5157 E-mail: [email protected]

Inseong Lee* HCI Lab, School of Business, Yonsei University, Seodaemun-gu, Seoul 120-749, Korea Fax: (+822)-364-7828 E-mail: [email protected] *Corresponding author

Jinwoo Kim HCI Lab, School of Business, Yonsei University, Shinchon-dong 134, Seodaemun-gu, Seoul 120-749, Korea Fax: (+822)-364-7828 E-mail: [email protected] Abstract: Although Mobile Data Services (MDS) are expected to revolutionise internet use, they have yet to do so, in part, because many users are not continuing with MDS after initial use. This study examines which factors are most important in converting discontinuers into continuers, and whether they differ from those that are effective in maintaining continuers. We conducted an online survey to compare continuers and discontinuers empirically in terms of the relative importance of seven post-adoption factors to the behavioural intention to use MDS. The results show that usefulness and social influence were more important for discontinuers, ubiquitous connectivity for continuers. Keywords: continuer; discontinuer; mobile data services; post-adoption. Copyright © 2008 Inderscience Enterprises Ltd.

Maintaining continuers vs. converting discontinuers

109

Reference to this paper should be made as follows: Kim, H., Lee, I. and Kim, J. (2008) ‘Maintaining continuers vs. converting discontinuers: relative importance of post-adoption factors for mobile data services’, Int. J. Mobile Communications, Vol. 6, No. 1, pp.108–132. Biographical notes: Hoyoung Kim has worked at the Widerthan, merged with the RealNetwork in 2006, since he graduated the Yonsei University in 2002. He had been interested in the usage of mobile internet during the graduate school. He had published in Information System Frontiers and Electronic Markets. Inseong Lee is a Researcher at the HCI Lab at Yonsei University in Korea. His current research interest is the study of cultural factors that affect mobile user experience. He has published in Int. J. Human-Computer Interaction, Int. J. Electronic Commerce and CyberPsychology & Behavior. Jinwoo Kim is a Professor of HCI at the School of Business, Yonsei University. He is also working as the director of HCI Lab at Yonsei University. His research interests include cultural design of information appliances including mobile phones, social network analysis for blog users, user generated contents and media substitution.

1

Introduction

With the rapid advance of telecommunication technologies, Mobile Data Services (MDS), defined here as wireless access to the internet through a mobile communication network (e.g. GSM or CDMA) by means of hand-held devices such as mobile phones (Federal Trade Commission, 2002), are poised to reshape the internet industry (Scornavacca and Barnes, 2004; Chen, Zhang and Zhou, 2005). In Asian countries, for example, a MDS-enabled phone has come to be considered a necessity of daily life (Minges, 2005); the number of people in Japan adopting MDS already exceeded the number using the traditional stationary internet (Kim et al., 2004). Recent research predicted that the number of MDS users in the world will reach 934 million by 2007 (Mitchell and Higgins, 2002). Many forecasters suggested that in the near future, most internet access will involve small, wireless devices providing more powerful internet services than those presently available on mobile devices (Lee, Kim and Kim, 2005). Recent environments can offer the users most feasible conditions to enjoy MDS than ever. To begin with, in terms of network speed, most major operators in the world have rolled out the 3G network supporting more than 2M BPS (Bits Per Second) when users download the data (Chiu, Lee and Chao, 2006). In addition, the operators have increasingly expanded the network coverage to encourage the users to access anywhere and anytime (The UTMS Forum, 2005). This improvement can provide the user with more seamless and more brilliant MDS than ever. Device evolution to support various convergence services is also outstanding. Latest devices, for instance, are basically embedded in mp3 function with the large storage to store more than 100 songs, and in camera capable of taking fine pictures like the digital camera (The Ericsson, 2005). In terms of price, the operators are expected to lower the MDS price as the bandwidth expands more. In Korea, the SK Telecom dropped about 30% of MDS price down this

110

H. Kim, I. Lee and J. Kim

year (The Etnews, 2007). Lower price and multifunctional devices have sufficient condition to boost up the MDS. Despite this optimism and positive trends, however, MDS face serious problems in terms of low profits and shallow customer bases (Urbaczewski et al., 2003), in part because many people adopt MDS and then cease to use them at a later date (Lee, Choi and Kim, 2005). We call users who have had the capability to use MDS, have used them at least once and have subsequently stopped using them, ‘discontinuers’, and those who have kept using MDS after initial adoption ‘continuers’. Even though more than 30 million Korean people (63% of the total population) have subscribed to MDS at least once, only 32.6% of them kept using MDS after their initial adoption (Sir et al., 2004). In the USA, according to a survey, 70% of users who had used the mobile internet at least once said they did not want to use it again (Schultz, 2001). The low percentage of continuers is worth noting, since it is directly related to the low profits seen by MDS providers. It should also be noted that discontinuers may influence continuers through word-ofmouth (Oliver, 1997) – an effect likely to be enhanced by the fact that the mobile phone is designed precisely for communication. In order to increase the number of MDS users, and hence profits, we need effective strategies for converting discontinuers into continuers, as well as for maintaining continuers (Bruner, Cybernautics and Usweb Corporation, 1998). It has been found that discontinuers behave differently from continuers in terms of accepting new technologies (Parthasarathy and Bhattacherjee, 1998), suggesting that we need to implement different strategies for discontinuers than we do for continuers. To do so, however, we need first to understand the relative importance of diverse factors that may affect user behaviour after they initially adopt. In other words, which post-adoption factors are most crucial for continuers, and which ones for discontinuers? Despite the key role post-adoption factors play, research on MDS has thus far shown two limitations. First, the unique characteristics of MDS have not been considered appropriately, in spite of their importance. Although MDS can be used at anywhere and anytime, users have to deal with constrained resources, such as awkward input and output systems and low processing power (Lee and Benbasat, 2003). These benefits and constraints have a strong influence on the general usage of MDS (Chae and Kim, 2003), and should be considered in any effort to identify the critical factors that affect postadoption use decisions. Second, prior studies on post-adoption behaviour have compared continuers with discontinuers in terms of several important factors (e.g. usefulness of the service), but have not explained how these factors bear upon the intention to continue or to cease using the services (Karahanna, Straub and Chevany, 1999). Thus the studies could explain how continuers were different from discontinuers, but could not provide concrete guidance on how to convert discontinuers into continuers, or on how to maintain continuers. Our objective in this study was to compare continuers and discontinuers in terms of the critical factors that affect the post-adoption intention to use MDS. Our research tasks, then, were to identify which factors were effective in increasing the intention to use MDS among discontinuers, and to determine how these factors differed from those that influenced continuers. In order to address these questions, we drew seven post-adoption factors from prior studies in related fields, and conducted a large-scale survey to compare continuers and discontinuers in terms of how each factor affected the intention to use MDS. This comparison of the relative importance of each factor would do more than

Maintaining continuers vs. converting discontinuers

111

explain how the two groups were different from each other – it would, when those differences were identified, suggest different strategies for increasing each group’s intention to use MDS. In Section 2, we review prior research on technology innovation and post-adoption behaviour, and derive from the pertinent studies seven key post-adoption factors. The section following presents our theoretical framework and hypotheses. Subsequently, we describe our research methods and the data collected, and present the results of our analysis. The paper ends with an examination of the study’s theoretical and practical implications and a discussion of its limitations.

2

Theoretical background

Prior studies targeting post-adoption factors specific to MDS are scarce. For this reason, we collected potentially important post-adoption factors from prior studies of technology adoption in general, reviewing the literature in three related areas: marketing (e.g. Tolman, 1932; Zeithaml, 1988; Ajzen, 1991; Dodds and Monroe, 1985, 1991; Cronin, Brady and Hult, 2000), technology innovation (e.g. Rogers, 1995) and information systems (e.g. Davis, 1989; Davis, Bagozzi and Warshaw, 1989; DeLone and McLean, 1992; Parthasarathy and Bhattacherjee, 1998; Karahanna, Straub and Chevany, 1999). Marketing was selected because MDS have been used primarily by individual consumers for personal purpose; technology innovation because MDS were introduced only a few years ago and are distinctly different from prior services; and information systems because MDS are a type of information system, and many adoption studies (e.g. Technology Acceptance Model (TAM)) have been conducted in the area. We selected six key factors that have been frequently used in prior adoption studies in these three areas: usefulness, usability, system quality, social influence, compatibility and perceived cost.1 We also added a seventh factor, ubiquitous connectivity, one of the most representative features of MDS. Definitions of each factor, as well as accounts of why they are pertinent to this study, follow.

2.1 Usefulness Usefulness is how helpful a user feels a new product is to his or her work (Rogers, 1995). When the usefulness of a new product is high, the product is adopted rapidly in the market. Parthasarathy and Bhattrcherjee (1998) suggested that people are more willing to pay for useful online services than for less useful ones. Karahanna, Straub and Chevany (1999) have established that good quality and good function, as perceived by users, allow them to adopt a new system easily in an organisational environment. Usefulness appears to be a key factor in the adoption of MDS as well (Koivumaki, Ristola and Kesti, 2006; Mahatanankoon, Wen and Lim, 2006) – especially when their usefulness differentiates them from traditional internet services (Pedersen, Methlie and Thorbjonsen, 2002). For example, subjects in one study reported using MDS only when the usefulness of their mobility really mattered (Jarvenpaa et al., 2003).

112

H. Kim, I. Lee and J. Kim

2.2 Usability Usability, another subjective measure, describes how easy (Preece et al., 1994) and comfortable (Davis, 1989; Segars and Grover, 1993) people find a system to learn and use. Rogers (1995) suggested that a system is adopted quickly when a user can easily learn how to use it. Nielsen (1993) proposed that for better usability a system must be efficient to use. Ajzen (1991) found a high correlation between difficulty of use and choice of services. Usability will also be an important factor for the adoption of MDS (AlShaali and Varshney, 2005; Mahatanankoon, Wen and Lim, 2006). Providers are coming more and more to believe that usability is the key to retaining customers (Nachum, 2003); poor usability is thought to be the greatest barrier to the adoption of MDS (Venkatesh, Ramesh and Massey, 2003). However, compared with the resources of other systems, the resources available to make MDS more usable are severely constrained (Kristoffersen and Ljungberg, 1999).

2.3 System quality System quality refers to the perceived stability and efficiency of a system (DeLone and McLean, 1992). As a technology advances, the system quality becomes an important factor for users. For example, in an online shopping mall, the speed and stability of the system network affected the attitude of online shoppers significantly (Liao and Cheung, 2001). In addition, Chin, Diehl and Norman (1988) found that the speed and stability of a system was an important determinant of user satisfaction. Speed and stability have been found to be important in determining the overall satisfaction of MDS users (Kim and Steinfield, 2004). For example, Jarvenpaa et al. (2003) found that poor system quality (e.g. service with frequent ‘handoffs’) was perceived by users as the greatest limitation of mobile commerce services. A similar argument has been made about the adoption of mobile hand-held devices (Sarker and Well, 2003). Thus, the system quality of MDS may significantly affect both adoption and ongoing usage.

2.4 Social influence Social influence refers to the effect of interaction among people in their social context (Rice et al., 1990). Social influence helps determine whether technologies were adopted and whether products are purchased (Venkatesh, Ramesh and Massey, 2003). Fisher and Price (1992) found that when people purchased new products, others’ opinions influenced their purchasing decisions. Social influence may also affect the use of MDS. In fact, since MDS are a part of the telecommunications industry – an industry specifically designed to facilitate social interactions – social influence may be an even more important factor in service choice than it would otherwise be (Downes and Mui, 1998). In one study, an individual’s adoption of iMode, a famous MDS in Japan, was found to be determined primarily by whether it was adopted by acquaintances, such as friends or colleagues (Stuart and Sid, 2003).

Maintaining continuers vs. converting discontinuers

113

2.5 Compatibility Compatibility refers to the conformity of an individual user’s background – including their lifestyle and personal knowledge of frequently used products – with the services in question (Rogers, 1995). Rogers (1995) suggested that a user adopts a new system quickly when the system is compatible with his or her everyday experience. Conversely, a system or product that requires radically new behaviour may be difficult for the user to adopt. In addition, Parthasarathy and Bhattrcherjee (1998) suggested that an individual’s demographic profile, lifestyle and knowledge of other devices all affect how he or she uses online services. Another element of a user’s background was his or her culture (Stuart and Sid, 2003); Sarker and Wells (2003) identified culture and prior experience as two major factors influencing the adoption of mobile devices. With their awkward input systems and small screens, MDS devices are quite different from their stationary internet counterparts (Chae and Kim, 2003; Siau, Sheng and Nah, 2004). Thus the compatibility of MDS with the prior knowledge and experience of users of the stationery internet (or of other information systems) is an open question. This compatibility (or lack thereof) may be an important determinant of the value users find in MDS.

2.6 Perceived cost Perceived cost refers to the monetary costs and time that users say they encounter when using a given product or service (Zeithaml, 1988). Consumers’ sense of the financial and temporal costs of a product or service needs to be considered when evaluating them (Zeithaml, 1988; Dodds and Monroe, 1991). Garbarino and Edell (1997) suggested that when a consumer overestimates the cost of a product or service, his or her choice can be influenced adversely. Interestingly, Cronin, Brady and Hult (2000) found that excessive effort affected the user’s perceived cost. MDS involve a higher perceived cost than most other information systems. They involve several types of monetary cost, such connection fees and information usage fees, which make them more expensive than most stationary internet services (Wu and Wang, 2004). Moreover, small screens and difficulties of manipulation required MDS users to expend additional time, further raising the costs of usage (Kristoffersen and Ljungberg, 1999). Because of the higher perceived cost of MDS, most providers find they must reduce the initial cost of the service until customers are established as ongoing users (Funk, 2000).

2.7 Ubiquitous connectivity Ubiquitous connectivity is the distinctive capacity of MDS to be used anywhere and at any time (Ahmed and Hurst, 2000; Dey, 2001; Kini and Thanarithiporn, 2004; Lee, Kim and Kim, 2005). In other words, MDS users can use the services regardless of time or place (Dey, 2001; Jarvenpaa et al., 2003). For example, when a user is in transit or for some other reason cannot access the stationary internet, he or she can tap into internet services through a mobile device like a cellular phone. This ubiquitous connectivity, elsewhere referred to as ‘freedom of choice’ (Jarvenpaa et al., 2003), induces a relatively high level of observability and makes MDS unique (Stuart and Sid, 2003). A recent qualitative study found that ubiquitous connectivity in general or at work was considered

114

H. Kim, I. Lee and J. Kim

as one of the key benefits of MDS (Jarvenpaa et al., 2003; Gressgård and Stensaker, 2006).

2.8 Perceived value and behavioural intention Prior studies indicated that these seven factors – usefulness, usability, system quality, social influence, compatibility, perceived cost and ubiquitous connectivity – affect perceived value, and that perceived value in turn affects the behavioural intention to use technologies (Wakefield and Barnes, 1996; Woodruff, 1997; Cronin, Brady and Hult, 2000; Zeithaml and Bitner, 2000). Perceived value is a broad and abstract concept, involving all the benefits users find in or ascribe to the purchase and use of a product or service (Zeithaml and Bitner, 2000), and it is therefore difficult to define. One useful approach has been to measure it in terms of monetary value (Dodds and Monroe, 1991). Price is one of the important factors affecting the use of MDS, and people usually pay MDS usage fees out of their personal funds (Kleijnen, Wetzels and de Ruyter, 2004). Thus, perceived value of MDS might be measured in terms of the differences between benefits and costs in purchasing or using MDS (Woodruff, 1997). Perceived value affects behavioural intention, which is the degree of reported intention to use products or services in the future (Wakefield and Barnes, 1996; Cronin, Brady and Hult, 2000). For example, Oh (1999) found that when a high value was attributed to a specific service, the behavioural intention to use that service in the future was greater. In the case of MDS, too, high perceived value may affect behavioural intention favourably. Based on reviewing major literatures that were cited earlier, the research model with the seven post-adoption factors, as well as perceived value and behavioural intention, has been developed in Figure 1. Figure 1

Post-adoption model for mobile data services

Maintaining continuers vs. converting discontinuers

115

The main purpose of this study is not to validate the model in Figure 1. Many factors in Figure 1 have been examined extensively in prior studies (e.g. Davis, 1989; Davis, Bagozzi and Warshaw, 1989; Parthasarathy and Bhattacherjee, 1998; Venkatesh and Brown, 2001), and there is little need to add further empirical verification, even in a new area such as MDS. Nor does this study focus on comparing continuers and discontinuers in terms of the selected adoption factors (the circled elements in Figure 1, e.g. usefulness); adoption factors per se have already been compared in prior research (though in the context of online services rather than that of MDS) (Parthasarathy and Bhattacherjee, 1998). This research did not, however, provide direct indications as to how to approach continuers and discontinuers differently, which is the main research question of the present study. Our main task, then, is to compare continuers and discontinuers in terms of the relationships among the adoption factors, perceived value and behavioural intention (the lines in Figure 1, e.g. that from usefulness to perceived value). To compare the two groups in this way, we used multi-group analysis based on structural equation modelling. This approach allows us to suggest how continuers and discontinuers should be approached differently – a topic hardly investigated, as we have said, in prior studies of technology adoption or post-adoption behaviour. The following section presents our research framework and hypotheses.

3

Theory and research hypotheses

In general, critical factors that affect the adoption of products or services differ according to user type (Karahanna, Straub and Chevany, 1999; Venkatesh, 2000). In other words, different types of user of a product or service – an information system, for instance, like MDS – will consider different factors to be important. Our study classifies users of MDS by type into two groups, and our main hypothesis is that the type of user – continuer or discontinuer – moderates all the relationships among the chosen post-adoption factors. In particular, we hypothesise that all seven factors will affect perceived value more heavily for discontinuers than for continuers, and further that perceived value will affect behavioural intention more heavily for discontinuers than for continuers. In other words, discontinuers will be more sensitive to a given change in any of the seven factors, modifying their perceptions of value and also their behavioural intentions more readily. Our main hypothesis is based on a synthesis of two closely related studies, one in the field of technology innovation (Rogers, 1995) and the other in the area of post-adoption behaviour (Parthasarathy and Bhattacherjee, 1998). In the former study, Rogers (1995) proposed four different kinds of adopter: innovators, early adopters, early majority adopters and late majority adopters. The innovators are the individuals in a social system who first adopt an innovation readily – in the research by Rogers (1995) research, 2.5% of the population. The early adopters are next 13.5% of individuals in a social system, who adopt the innovation after the innovators. The other two groups, early majority and late majority adopters, are slow to adopt new technologies and constitute more than 70% of the total population. One common characteristic of the innovators and early adopters is their ability to cope with a high degree of uncertainty – uncertainty being a typical feature of innovative technologies like MDS. As they are adventurous, they are eager to try out new devices regardless of whether the technology offers, at that point in time, any clear practical value. In other words, they are insensitive to the specific benefits and costs a

116

H. Kim, I. Lee and J. Kim

technology entails at that time, and driven instead by personal curiosity and predilections (Rogers, 1995). In contrast, majority adopters approach innovations with caution and scepticism. They adopt new ideas only after most of the uncertainty about the innovation has dissipated. They are unwilling to risk scarce resources on untested technologies, or to adopt a new technology out of economic necessity or peer pressure. They are, therefore, more sensitive than the first two groups to the benefits and costs of a new technology, and this assessment of current value is the driving force behind their adoption behaviour. In their study of post-adoption behaviour, Parthasarathy and Bhattacherhjee (1998) found that innovators and earlier adopters were less likely to discontinue a new service than early and late majority adopters. In other words, they found that continuers were more likely to be innovators and early adopters, while discontinuers were more likely to be majority adopters (Parthasarathy and Bhattacherjee, 1998). It may be that majorities have unrealistically high expectations of a new service, or that they lack the skills or resources to utilise the new service effectively. In any event, majorities are less satisfied by the new service, and thus more likely to cease using it. Synthesising these two studies, we hypothesise that continuers are less sensitive than discontinuers to various adoption factors, because they are mostly innovators and early adopters. Conversely, discontinuers, being for the most part majority adopters, are likely more sensitive to the same adoption factors. Thus we propose as our main hypothesis: Hypothesis 0. The seven MDS adoption factors influence perceived value and behavioural intention more strongly for discontinuers than for continuers. This hypothesis can be further specified in terms of the eight relationships among the variables in Figure 1. We explain each of the eight sub-hypotheses only briefly, as a full description for each would result in unnecessary repetition of the main hypothesis. First, we hypothesise that continuers keep finding value in MDS even though they are not as useful as they thought they would be, whereas the perceived value of MDS drops significantly for discontinuers if they find they are not as useful as expected. Indeed, recent research reported that discontinuers readily stop perceiving value in MDS because they offered poor output and unverified usefulness (Landor, 2003). Thus we propose as our first sub-hypothesis: Hypothesis 1. Usefulness influences perceived value more strongly for discontinuers than for continuers. Second, we hypothesise that discontinuers are more sensitive to a technology’s ease of use, and that usability will therefore have a greater influence on discontinuers than on continuers. Unskilled users have a stronger negative reaction to a bad usability experience, because it further weakens their already weak confidence (Sumner and Taylor, 1997). Karahanna, Straub and Chevany (1999) also found that discontinuers were more sensitive to usability than early adopters and continuers, who keep finding value in a technology even though they experience difficulties in using it (Moore and Benbasat, 1991; Nachum, 2003; Stuart and Sid, 2003). Thus we propose as our second subhypothesis: Hypothesis 2. Usability influences perceived value more strongly for discontinuers than for continuers. Third, we hypothesise that discontinuers are more sensitive to the overall quality of a technology, and that system quality will be more influential on discontinuers than on

Maintaining continuers vs. converting discontinuers

117

continuers. In other words, discontinuers may be quicker to stop finding value in MDS when they experience system malfunctions. In contrast, it is common to find continuers proudly explaining how they overcame the system malfunctions of a new technology (Rogers, 1995). Thus we propose as our third sub-hypothesis: Hypothesis 3. System quality influences perceived value more strongly for discontinuers than for continuers. Fourth, we hypothesise that social influence is more powerful for discontinuers than for continuers. This hypothesis is consistent with the finding of Rogers (1995) that early adopters, most of whom are continuers, are likely to use innovative products out of their own curiosity, while later adopters are more influenced by other people’s opinions and experiences. Thus we propose as our fourth sub-hypothesis: Hypothesis 4. Social influence affects perceived value more strongly for discontinuers than for continuers. Fifth, we hypothesise that compatibility is more influential on discontinuers than on continuers. Continuers, most of whom are innovators and early adopters, tend not to be afraid of trying new and esoteric services (Rogers, 1995). In contrast, discontinuers may be more conservative, requiring a new technology to fit into their existing lifestyle – in this case, requiring MDS to work the way wired internet services do (Anckar and D'Incau, 2002). Thus we propose as our fifth sub-hypothesis: Hypothesis 5. Compatibility affects perceived value more strongly for discontinuers than for continuers. Sixth, we hypothesise that ubiquitous connectivity affects perceived value more strongly for discontinuers than for continuers. Discontinuers are for the most part late adopters, who tend to rely on the proven benefits of a new technology, and ubiquitous connectivity is the main benefit to be found in MDS. Thus we propose as our sixth sub-hypothesis: Hypothesis 6. Ubiquitous connectivity affects perceived value more strongly for discontinuers than for continuers. Seventh, we hypothesise that discontinuers are more sensitive than continuers to the perceived costs of using MDS. Early adopters are ready to accept the costs of a new technology, while late adopters are typically reluctant to pay those (Moore and Benbasat, 1991). Discontinuers, who are mostly late adopters, may consider the high cost of MDS more important than continuers do. Thus we propose as our seventh sub-hypothesis: Hypothesis 7. Perceived cost affects perceived value more strongly for discontinuers than for continuers. Finally, we hypothesise that perceived value affects behavioural intention more strongly for discontinuers than for continuers. Continuers, being early adopters, are more likely to intend using a new technology regardless of its currently perceived value, whereas discontinuers need its value to be proven before they will use it. Thus we propose as our final sub-hypothesis: Hypothesis 8. Perceived value affects behavioural intention more strongly for discontinuers than for continuers.

118

4

H. Kim, I. Lee and J. Kim

Research methodology

4.1 Development of questionnaire To test our hypotheses, we created a set of questions for a nationwide online survey of Korean MDS users. Questions for all constructs were adapted from related studies (e.g. Davis, Bagozzi and Warshaw, 1989; Rogers, 1995; Sheth, Newman and Gross, 1991; Venkatesh and Brown, 2001) to increase their content validity. To increase the validity of the empirical data, we constructed a research consortium consisting of representatives of all the mobile telecommunication companies in Korea (SK Telecom, LG Telecom and Korea Telecom Freetel), as well as major internet portals (Yahoo Korea, NATE, Daum Communication, etc.), major MDS providers and MDS application developers. We conducted in-depth interviews with each member of the consortium, and modified our survey questions according to the interview results. The modified questionnaire was pre-tested again with 48 heavy users, defined as users who use MDS for more than 200 min month–1 on average. We also conducted pre-tests with 53 average users, defined as users who use MDS for less than 200 min month–1 on average. The mobile telecommunication companies in our consortium referred these users to us as pre-test subjects. We started with 54 questions in our questionnaire, but ended up with 26 after deleting the questions that pre-test subjects found hard to understand.2 The final set of questions used in the main survey is presented, with references, in Appendix 1.

4.2 Data collection Respondents to the online survey were recruited through banner advertisements at the major portals participating in our consortium. Survey data was collected as follows. Before asking the questions, we emphasised that only those who had used MDS at least once were eligible to take the survey. We also made sure that only those who allowed us to check their usage data against the records of their telecommunication companies participated. We then asked the respondents to provide their mobile phone numbers, which the telecommunication companies would use to classify them as continuers or discontinuers. We attempted to ensure accurate reporting of phone numbers by sending the lottery prizes for their participation to the billing addresses of the reported numbers. Respondents went on to answer each of the survey questions shown in Appendix 1 on a seven-point Likert scale (ranging from 1 for strongly disagree to 7 for strongly agree). They finished the survey by providing demographic information. Once they had finished the survey, their phone numbers and survey responses were sent to the three telecommunication companies for data verification. The fact that all three companies providing MDS in Korea participated in our consortium increased the validity of the data significantly. The telecommunication companies checked whether the phone numbers reported were legitimately registered, and whether the owners of the phone numbers had used MDS at least once in the past. They also classified respondents as continuers or discontinuers, using log data in legacy systems. Those who had used MDS at least once, but had not used it at all in the past month, were categorised as discontinuers, while those who had used MDS more than four times in the past month were classified as continuers.3 The criterion of more than four uses in a month was an industry standard widely used by telecommunication companies for customer

119

Maintaining continuers vs. converting discontinuers

management purposes (HCI Lab, 2002). After verification and classification, the companies returned the survey data to the authors, with private information such as phone numbers deleted to protect the privacy of survey respondents. Cooperating in this way with telecommunication companies, we were able to increase the validity of survey data without significantly undermining the privacy of survey respondents. In the end, 1,789 respondents were classified as continuers and 1,770 as discontinuers. Demographic information on members of the two groups is presented in Table 1. Table 1

Demographic information for effective respondents

Age (years) Usage months Gender Male Female Educational status Junior high school completed High school completed Undergraduate completed Postgraduate completed

Mean 22.3 8.8

Continuers SD 5.3 6.5

Discontinuers Mean SD 24.0 6.7 9.2 6.97

60.6% 39.4%

59.0% 41.0%

18.5% 48.0% 28.5% 5.0%

23.9% 51.2% 21.8% 3.1%

Respondents in the discontinuer group were slightly older on average than those in the continuer group. The average number of usage months since the first use was also slightly higher among discontinuers than among continuers. However, these differences did not reach statistically significant levels, and the two groups were treated as two homogenous groups.

4.3 Measurement validation To ensure construct validity, explorative factor analysis was performed; results are shown in Appendix 2. Among the seven factors originally proposed, only six were extracted as independent variables: Usefulness (UF), Usability (US), System Quality (SQ), Social Influence (SI), Ubiquitous Connectivity (UC) and Perceived Cost (PC). The seventh construct, Compatibility, was not retained, as respondents took the two questions for this construct to pertain to usefulness or usability. In hindsight, their reading is understandable: the statement “MDS fits my daily routine well” can be interpreted as a question concerning usefulness, in that people may or may not use MDS daily for useful benefits they provide; and the statement “MDS is easy for me to adjust to” can pertain to usability, since people may or may not learn to adjust to MDS easily. We dropped the Compatibility construct in our further analysis and folded the two questions into the Usefulness and Usability constructs, respectively. Our finding was somewhat consistent with prior research, in which the questions pertaining to compatibility could not be merged into a single factor (Karahanna, Straub and Chevany, 1999). Two factors were extracted as dependent variables: Perceived Value (PV) and Behavioural Intention (BI). The variances explained were 59.4% for independent variables and 82.7% for dependent variables, and Eigenvalues of extracted factors were all above 1.

120

H. Kim, I. Lee and J. Kim

We calculated Cronbach’s alphas for each construct to confirm the reliability of the survey instrument. As shown in the last row of Appendix 2, Cronbach’s alphas for all factors except Perceived Cost exceeded the cut-off point of 0.7. We also conducted a Confirmatory Factor Analysis (CFA) to verify the construct validity of questions for the adoption factors. The results of the CFA (shown in Appendix 3) indicate that all the questions converged well into their respective constructs except the two compatibility questions. The Goodness of Fit Index (GFI) was 0.94, and the Adjusted Goodness of Fit Index (AGFI) was 0.92, exceeding threshold levels (Bentler, 1990; Browne and Cudeck, 1993), while the Normed Fit Index (NFI) and the Non-Normed Fit Index (NNFI) approached the recommended levels of 0.90. Also, t-values of each construct (including Perceived Cost but excluding Compatibility) were significant at the level of 0.01, which indicates high convergent validity of the seven variables (Anderson and Gerbing, 1988). The correlation and Average Variance Extracted (AVE) from confirmatory factor analysis are presented in Appendix 4. The diagonal elements in this appendix are the square roots of AVE, and the off-diagonal elements are inter-construct correlations. The six diagonal elements were all higher than 0.5 except Perceived Cost, and also higher than their corresponding correlation coefficients, which indicated that the metrics had appropriate discriminant validity (Gefen, Straub and Boudreau, 2000). In sum, the convergent and discriminant validity, as well as the reliability, of most measures were good enough to proceed to further analysis. Although questions for Perceived Cost did not reach the threshold value in terms of reliability and discriminant validity, we decided to retain them, first because the variable was converged into one factor in our explorative and confirmatory factor analyses, and second because cost was one of the primary features distinguishing MDS from the traditional stationery internet (HCI Lab, 2002).

5

Results

Structural equation modelling (SEM) analysis with LISREL version 8.13 was used to investigate the causal relations among these factors. Figure 2 presents two different causal models, one for continuers and one for discontinuers. The GFI shown at the bottom of Figure 2 indicated that the two models were valid enough to justify investigating the path coefficients of the causal relations.4 Multi-group analysis using LISREL was conducted for each causal relation in order to compare the relative strengths of influences for continuers and discontinuers. Table 2 summarises the results of the multi-group analysis. The effects of Usefulness and Social Influence on Perceived Value were significantly greater for discontinuers than for continuers, a result consistent with sub-hypotheses, Hypotheses 1 and 4. However, the influence of Ubiquitous Connectivity on Perceived Value was greater for continuers than for discontinuers at the level of 0.01, and the impact of Perceived Value on Behavioural Intention was significantly greater for continuers than for discontinuers at the level of 0.001. These two results contradict subhypotheses, Hypotheses 6 and 8. The other relationships – of Usability, System Quality and Perceived Cost to Perceived Value – were not found to be significantly different for continuers and discontinuers (sub-hypotheses, Hypotheses 2, 3 and 7). Table 3 summarises the test results for each of the sub-hypotheses comparing the two groups.

121

Maintaining continuers vs. converting discontinuers Figure 2

Comparison of continuers and discontinuers

Table 2

Sensitivity comparison for continuers and discontinuers

Type

Chi-square

df

|Difference|

Unconstraint

4327.94

556

Usefulness

4339.11

557

11.17**

Usability

4328.28

557

0.34

–

System Quality

4327.48

557

0.46

Social Influence

4356.16

557

28.22**

Ubiquitous Connectivity

4331.54

557

3.6*

Perceived Cost

4327.72

557

0.22

Perceived Value

4356.41

557

28.47**

**p < 0.001, *p < 0.01

122

H. Kim, I. Lee and J. Kim

Table 3

Comparison of causal relations for continuers and discontinuers

Type

Hypothesis

Result

Hypothesis 1

Usefulness influences Perceived Value more strongly for discontinuers than for continuers.

Supported

Hypothesis 2

Usability influences Perceived Value more strongly for discontinuers than for continuers.

Not supported

Hypothesis 3

System Quality influences Perceived Value more heavily for discontinuers than for continuers.

Not supported

Hypothesis 4

Social Influence affects Perceived Value more strongly for discontinuers than for continuers.

Supported

Hypothesis 5

Compatibility affects Perceived Value more strongly for discontinuers than for continuers.

N/A

Hypothesis 6

Ubiquitous Connectivity affects Perceived Value more strongly for continuers than for discontinuers.

Rejected

Hypothesis 7

Perceived Cost affects Perceived Value more heavily for discontinuers than for continuers.

Not supported

Hypothesis 8

Perceived Value affects Behavioral Intention more strongly for discontinuers than for continuers.

Rejected

6

Conclusion, discussion and implications

This study attempted to identify adoption factors that increase the behavioural intention of discontinuers to use MDS, and to discern how they differ from factors that affect the behavioural intention of continuers. We derived seven adoption factors from literature in the fields of marketing, information systems and technology innovation. Among the seven factors, this study identified two that had more effect on discontinuers (usefulness and social influence) and one that had more effect on continuers (ubiquitous connectivity). The stronger effect of ubiquitous connectivity on continuers was the opposite of our expectation. This result may stem from the fact that early adopters, who constitute most of the continuer group, rely more heavily on external sources of information than on internal sources like friends and family (Parthasarathy and Bhattacherjee, 1998). Ubiquitous connectivity was the characteristic of MDS most frequently mentioned by external sources such as advertising and public announcements (HCI Lab, 2002; Jarvenpaa et al., 2003). It may be that continuers are better acquainted than discontinuers with ubiquitous connectivity, and therefore more sensitive to it. The impact of perceived value on behavioural intention was, counter to our hypothesis, stronger for continuers than for discontinuers. It may be that the behavioural intention of discontinuers is less sensitive to a rise in perceived value because they have already used the service and decided not to use it anymore. One implication is that if discontinuers are to consider using the abandoned service once more, the rise in perceived value must be substantial enough to shake them out of a decision already made and acted on. This implies in turn that the task of converting discontinuers will likely be more difficult than that of maintaining continuers. The study has several interesting theoretical and practical implications. Theoretically, it suggests a new analytical approach, namely, comparison of relationships among key

Maintaining continuers vs. converting discontinuers

123

adoption factors. Prior research has identified the different characteristics of continuers and discontinuers, but has not explained directly how to approach them differently. Comparing the strength of the relationship in each group between perceived value and behavioural intention, as well as the relationships between perceived value and chosen post-adoption factors, offers concrete guidance on increasing the intention of a specific group to use MDS in the future. Second, the hypotheses were tested with a large sample of respondents, and the study enhanced the reliability of the proposed comparison by using objective data from telecommunication companies. It had telecommunication companies to verify that respondents were actual users of MDS, and had them classify respondents according to the accurate objective data stored in legacy systems. This method may suggest a new means of data collection that can secure the integrity of usage data without undermining company security or user privacy. Third, while the proposed factors for MDS are based on prior studies in technology adoption, the study accounts for important characteristics of MDS, such as ubiquitous connectivity. Further, it takes the valuable step of viewing post-adoption behaviours from the perspective of the individual user, providing a post-adoption analysis of MDS that can be applied to individuals both in the office and at home. In practice, the study suggests taking direct methods to keep up the MDS users and to transform the discontinuers into the continuers. One implication is that strategies that increase the usefulness of MDS or create high social influence are likely to be more effective in converting discontinuers than in maintaining continuers. For instance, the ability to take pictures with mobile phones and send them through MDS may have the potential to convert discontinuers into continuers, because it directly increases the usefulness of MDS. Another policy to perceive the usefulness of MDS may expand the coverage of free promotion to older population like employees. Most free promotion of the operators tend to focus on teenagers, so most employees with MDS phones have rare opportunity to experience the usefulness of MDS. At the same time, more useful services for employees are developed. Current services such as the Ring Back Tone and the Ring Tone do not offer the value to the workers, because they are less interested in representing themselves differently than teenagers. Attractive services for employees, for instance, are to improve the efficiency such as the prompt check and reply of attached files documented by MS Word. To operate a monitoring group will also probably be more effective in converting discontinuers than in maintaining continuers, because such play may strengthen the social influence of friends and family. When a new service without user group’s testing and designing is released, it generally includes a lot of system errors, broken links and unattractive features. Those problems can be lifted by the monitoring group. It suggests a useful idea to build the new service’s feature and, after that, triggers the viral marketing among the peer groups. In contrast, strategies that increase the ubiquitous connectivity of MDS will be more effective in maintaining continuers than in converting discontinuers. For example, developing a location-based monitoring service such as a ‘buddy-finder’ might be effective in maintaining continuers, because it emphasises the ubiquitous connectivity of MDS. And another example is to provide the ubiquitous data access service like the Melon of the SKT. Such service enables the users to enjoy the music anytime, anywhere and any device.

124

H. Kim, I. Lee and J. Kim

Lastly, the company will invest more resource on the factors cited earlier than others. Recent main trend is that most companies have done utmost effort to deploy faster network and to launch it earlier than competitors even though most users felt the network speed insignificant. The study implied that high speed to offer the MDS did not convert the discontinuers into the continuers. In sum, implementing different strategies for different types of MDS user might alleviate the problem of low profitability that telecommunication companies are experiencing all over the world, in spite of the general optimism about MDS.

7

Limitations

This study has several limitations. First, it suffers from a methodological constraint in that it relies on an online survey. The online survey method involves a random sampling error and a self-selection bias. For example, our results may be biased towards those who are willing to use the internet to answer our questions. Even though the online survey has economic and administrative merits (Kehoe and Pitkow, 1996), alternative data collection methods, such as stratified random sampling with an e-mail survey, should be used in future research to increase the reliability of the results. The second limitation involves the measurement of adoption factors. The questions for compatibility failed to converge into a single factor, and the perceived cost construct could not meet the reliability and discriminant validity criteria. Even though we started with at least four questions for each construct, only two questions were left for some constructs (e.g. ubiquitous connectivity and perceived value) after pre-tests with heavy and normal users. The relative dearth of questions for these constructs may explain why the results support only two of our hypotheses. To enhance the validity of measurement, appropriate measures for these constructs should be developed and validated in a future study. Third, the criterion that distinguished continuers from discontinuers was the number of uses of MDS in the month prior to the survey period. In order to obtain objectively verified usage data from the system log files of telecommunication companies, we had to accept the criterion standard to the industry. This is a common practice in academic research when no robust, verified criterion is available. For example, Aarnio, Heikkila and Hirvola (2002) studied the adoption of mobile services using an industry standard suggested by a telecommunication company (Nokia Telecommunications, 1999). However, in the case of our study, the criterion is arbitrary, with no theoretical or empirical basis. Unfortunately, we could not collect more detailed measures, such as total usage time or usage fee, because of security and privacy concerns on the part of the telecommunication companies. Nor did we collect self-reported data on the time respondents started to use MDS or the number of times they had used MDS, because people usually remember their past usage poorly. For instance, a follow-up study by the authors found that more than one-third of all respondents failed to report their past usage accurately. Thus we decided to choose an objectively verified but somewhat arbitrary classification, rather than a more meaningful but largely unverifiable one. We are currently conducting a follow-up study in which we try to identify alternative measures for continuance, which are theoretically sound and empirically verifiable, maintain a high level of data integrity and protect privacy and security.

Maintaining continuers vs. converting discontinuers

125

Fourth, our research strategy was, first, to collect important factors from prior research in technology adoption in general, and then to compare their relative importance to perceived value and behavioural intention. The assumption of this strategy – that factors important in general technology adoption will be also important to a comparison of continuers and discontinuers of MDS – has not been verified. Furthermore, there may be other factors that are not important for technology adoption in general, but that are important for a comparison of these two groups. For example, the brand name of a MDS may be an important factor (Liao and Cheung, 2001), one that could not be drawn from prior research on technology adoption in general. A future study should include additional factors that may differentiate the two groups of MDS users. Finally, because the online survey was conducted in Korea, with a largely Korean pool of MDS users answering questions written in Korean, the results can be applied to other countries and cultures only with caution. Just as some of the questions used in prior studies could not be asked of Korean users because of language differences, some of the conclusions drawn about Korean users may be culturally or linguistically specific. That said, MDS is a worldwide phenomenon, and factors that influence post-adoption behaviour in one culture might tend to exert influence in others in spite of a degree of cultural specificity (Urbaczewski et al., 2002). To increase the external validity of the study results, we are currently conducting a study in several other countries.

Acknowledgement This study was funded by the Korea Research Foundation (KRF-2003-042-B0045).

References Aarnio, A.E.A., Heikkila, J. and Hirvola, S. (2002) ‘Adoption and use of mobile services: Empirical evidence from a Finnish survey’, Paper presented in the Proceedings of the 35th Hawaii International Conference on System Sciences, Big Island, HI. Ahmed, Z. and Hurst, M. (2000) ‘The wireless customer experience’, Available at: http://www.creativegood.com/wireless/cg-wireless0900.pdf Ajzen, I. (1991) ‘The theory of planned behavior’, Organizational Behavior and Human Decision Processes, Vol. 50, pp.179–211. AlShaali, S. and Varshney, U. (2005) ‘On the usability of mobile commerce’, Int. J. Mobile Communications, Vol. 3, pp.29–37. Anckar, B. and D’Incau, D. (2002) ‘Value-added services in mobile commerce: an analytical framework and empirical findings from a national consumer survey’, Paper presented in the Proceedings of the 35th Hawaii International Conference on System Sciences, Big Island, HI. Anderson, J.C. and Gerbing, D.W. (1988) ‘Structural equation modeling in practice: a review and recommended two-step approach’, Psychological Bulletin, Vol. 103, pp.411–423. Bentler, P.M. (1990) ‘Comparative fit indexes in structural models’, Psychological Bulletin, Vol. 107, pp.238–246. Browne, M.W. and Cudeck, R. (1993) ‘Alternative ways of assessing model fit’, in K.A. Bollen and J.S. Long (Eds), Testing Structural Equation Models (pp.136–162). Newbury Park, CA: Sage Publications. Bruner, R.E., Cybernautics and Usweb Corporation (1998) Net Results: Web Marketing that Works. Indianapolis, IN: Hayden Books.

126

H. Kim, I. Lee and J. Kim

Chae, M. and Kim, J. (2003) ‘What's so different about the mobile internet?’, Communications of the ACM, Vol. 46, pp.240–247. Chen, M., Zhang, D. and Zhou, L. (2005) ‘Providing web services to mobile users: the architecture design of an M-service portal’, Int. J. Mobile Communications, Vol. 3, pp.1–18. Chin, J.P., Diehl, V.A. and Norman, K.L. (1988) ‘Development of an instrument measuring user satisfaction of the human-computer interface evaluations’, Paper presented in the Proceedings of the 1998 Conference on Human Factors in Computing Systems, Washington, DC. Chiu, H-C., Lee, W-R. and Chao, Y-P. (2006) ‘The development of 3G in Japan and Taiwan’, Int. J. Mobile Communications, Vol. 4, pp.47–62. Cronin, J.J. Jr, Brady, M.K. and Hult, G.T.M. (2000) ‘Assessing the effects of quality, value, and customer satisfaction on consumer behavioral intentions in service environments’, Journal of Retailing, Vol. 76, pp.193–218. Davis, F.D. (1989) ‘Perceived usefulness, perceived ease of use, and user acceptance of information technology’, MIS Quarterly, Vol. 13, pp.319–340. Davis, F.D., Bagozzi, R.P. and Warshaw, P.R. (1989) ‘User acceptance of computer technology: a comparison of two theoretical models’, Management Science, Vol. 35, pp.982–1003. DeLone, W.H. and McLean, E.R. (1992) ‘Information systems success: the quest for the dependent variable’, Information Systems Research, Vol. 3, pp.60–95. Dey, A.K. (2001) ‘Understanding and using context’, Personal and Ubiquitous Computing, Vol. 5, pp.4–7. Dodds, W.B. and Monroe, K.B. (1985) ‘The effect of brand and price information on subjective product evaluations’, Advances in Consumer Research, Vol. 12, pp.85–90. Dodds, W.B. and Monroe, K.B. (1991) ‘Effects of price, brand and price information on buyers’ product evaluation’, Journal of Marketing Research, Vol. 28, pp.307–319. Downes, L. and Mui, C. (1998) Unleashing the Killer App: Digital Strategies for Market Dominance. Boston, MA: Harvard Business School Press. Federal Trade Commission (2002) ‘The mobile wireless web, data services and beyond: Emerging technologies and consumer issues’, Available at: http://www.ftc.gov/bcp/reports/ wirelesssummary.pdf Fisher, R.J. and Price, L.L. (1992) ‘An investigation into the social context of early adoption behavior’, Journal of Consumer Research, Vol. 19, pp.477–486. Funk, J.L. (2000) ‘MDS market: lessons from Japan’s I-Mode System’, Available at: http:// e-conomy.berkeley.edu/conferences/9-2000/EC-conference2000_papers/Funk.pdf Garbarino, E.C. and Edell, J.A. (1997) ‘Cognitive effort, affect, and choice’, Journal of Consumer Research, Vol. 24, pp.147–158. Gefen, D., Straub, D.W. and Boudreau, M.C. (2000) ‘Structural equation modeling and regression: Guidelines for research practice’, Communications of the Association for Information Systems, Vol. 4, pp.1–77. Gressgård, L.J. and Stensaker, I. (2006) ‘The mobile service industry: Strategic challenges and future business models’, Int. J. Mobile Communications, Vol. 4, pp.509–531. HCI Lab (2002) ‘The 3rd MDS user survey’, Yonsei University, Seoul, Korea. Jarvenpaa, S.L., Lang, K.R., Takeda, Y. and Tuunainen, V.K. (2003) ‘Mobile commerce at crossroads’, Communications of the ACM, Vol. 46, pp.41–44. Karahanna, E., Straub, D.W. and Chevany, N.L. (1999) ‘Information technology adoption across time: a cross-sectional comparison of pre-adoption and post-adoption beliefs’, MIS Quarterly, Vol. 23, pp.183–213. Kehoe, C.M. and Pitkow, J.E. (1996) ‘Surveying the territory: GVU’s five WWW user surveys’, The World Wide Web Journal, Vol. 1, pp.77–84. Kim, D.J. and Steinfield, C. (2004) ‘Consumers mobile internet service satisfaction and their continuance intentions’, Paper presented in the Proceedings of the 10th Americas Conference on Information Systems, New York, NY.

Maintaining continuers vs. converting discontinuers

127

Kim, J., Lee, I., Lee, Y., Choi, B., Hong, S.J., Tam, K.Y., Naruse, K. and Maeda, Y. (2004) ‘Exploring e-business implications of the mobile internet: a cross-national survey in Hong Kong, Japan and Korea’, Int. J. Mobile Communications, Vol. 2, pp.1–21. Kini, R.B. and Thanarithiporn, S. (2004) ‘Mobile commerce and electronic commerce in Thailand: a value space analysis’, Int. J. Mobile Communications, Vol. 2, pp.22–37. Kleijnen, M., Wetzels, M. and de Ruyter, K. (2004) ‘Acceptance of wireless finance’, Journal of Financial Service Marketing, Vol. 8, pp.206–217. Koivumaki, T., Ristola, A. and Kesti, M. (2006) ‘Predicting consumer acceptance in mobile services: empirical evidence from an experimental end user environment’, Int. J. Mobile Communications, Vol. 4, pp.418–435. Kristoffersen, S. and Ljungberg, F. (1999) ‘Mobile informatics: Innovation of IT use in mobile settings’, IRIS’21 workshop report, SIGCHI Bulletin, Vol. 31, pp.29–34. Landor, P. (2003) ‘Understanding the foundation of mobile content quality: a presentation of a new research field’, Paper presented in the Proceedings of the 36th Hawaii International Conference on System Sciences, Big Island, HI. Lee, I., Choi, B. and Kim, J. (2005) ‘Cultural fit as user's perceptual lens: Focusing on mobile internet users' post-adoption behavior’, Paper presented in the Proceedings of the 11th International Conference on Human-Computer Interaction, Las Vegas, NV. Lee, I., Kim, J. and Kim, J. (2005) ‘Use contexts for mobile services: a longitudinal monitoring study on actual use of mobile internet services’, Int. J. Human-Computer Interaction, Vol. 18, pp.269–292. Lee, Y.E. and Benbasat, I. (2003) ‘Interface design for mobile commerce’, Communications of ACM, Vol. 46, pp.49–52. Liao, Z. and Cheung, M.T. (2001) ‘Internet-based e-shopping and consumer attitudes: an empirical study’, Information & Management, Vol. 38, pp.299–306. Mahatanankoon, P., Wen, H.J. and Lim, B.B.L. (2006) ‘Evaluating the technological characteristics and trust affecting mobile device usage’, Int. J. Mobile Communications, Vol. 4, pp.662–681. Minges, M. (2005) ‘Is the internet mobile? Measurements from the Asian-Pacific region’, Telecommunications Policy, Vol. 29, pp.113–125. Mitchell, C. and Higgins, M. (2002) ‘The mobile divide’, Available at: http://www.theage.com.au/ articles/2002/06/11/1022982835152.html Moore, G.C. and Benbasat, I. (1991) ‘Development of an instrument to measure the perceptions of adoption an information technology innovation’, Information Systems Research, Vol. 2, pp.191–222. Nachum, G. (2003) ‘Owning the mobile content customer’, Available at: http://www.totaltele.com/ interviews/display.asp?InterviewID=254 Nielsen, J. (1993) Usability Engineering. New York: Academic Press. Nokia Telecommunications (1999) ‘The demand for mobile value-added services’, Available at: http://nds2.ir.nokia.com/downloads/aboutnokia/press/pdf/study-vas.pdf Oh, H. (1999) ‘Service quality, customer satisfaction, and customer value: a holistic perspective’, Int. J. Hospitality Management, Vol. 18, pp.67–82. Oliver, R.L. (1997) Satisfaction: a Behavioral Perspective on the Consumer. Boston, MA: McGraw-Hill. Parthasarathy, M. and Bhattacherjee, A. (1998) ‘Understanding post-adoption behavior in the context of online services’, Information Systems Research, Vol. 9, pp.362–379. Pedersen, P.E., Methlie, L.B. and Thorbjonsen, H. (2002) ‘Understanding mobile commerce enduser adoption: a triangulation perspective and suggestions for an exploratory service evaluation framework’, Paper presented in the Proceedings of the 35th Hawaii International Conference on System Sciences, Big Island, HI. Preece, J., Rogers, Y., Sharp, H., Benyon, D., Holland, S. and Carey, T. (1994) Human-Computer Interaction. Boston, MA: Addison-Wesley.

128

H. Kim, I. Lee and J. Kim

Rice, R.E., Grand, A.E., Schmitz, J. and Torobin, J. (1990) ‘Individual and network influences on the adoption and perceived outcomes of electronic messaging’, Social Networks, Vol. 12, pp.27–55. Rogers, E.M. (1995) Diffusion of Innovations, 4th edn. New York: The Free Press. Sarker, S. and Well, J.D. (2003) ‘Understanding mobile handheld device use and adoption’, Communications of ACM, Vol. 46, pp.35–40. Schultz, B. (2001) ‘The m-commerce fallacy’, Network World, Vol. 18, pp.77–82. Scornavacca, E. Jr and Barnes, S.J. (2004) ‘M-banking services in Japan: a strategic perspective’, Int. J. Mobile Communications, Vol. 2, pp.51–66. Segars, A.H. and Grover, V. (1993) ‘Re-examining perceived ease of use and usefulness: a confirmatory factor analysis’, MIS Quarterly, Vol. 17, pp.517–525. Sheth, J.N., Newman, B.I. and Gross, B.L. (1991) ‘Why we buy what we buy: a theory of consumption values’, Journal of Business Research, Vol. 22, pp.159–170. Siau, K., Sheng, H. and Nah, F.F-H. (2004) ‘The value of mobile commerce to customers’, Paper presented in the Proceedings of the the 3rd Pre-ICIS Annual Workshop on HCI Research in MIS, Washington, DC. Sir, J-C., Cho, C-H., Yang, H-J., An, I-H. and Kim, J-R. (2004) ‘2004 Survey on the wireless internet use’, Available at: http://isis.nic.or.kr/report_DD_View/upload/mobile200412_ eng[1].pdf Stuart, J.B. and Sid, L.H. (2003) ‘Rising sun: iMode and the wireless internet’, Communications of the ACM, Vol. 46, pp.78–84. Sumner, T. and Taylor, J. (1997) ‘Coping with virtuality: steps toward a personal learning manager’, Paper presented in the Proceedings of the UNESCO Workshop on Virtual Learning Environments, Open University, UK. The Ericsson (2005) ‘Evolution towards converged services and networks’, Available at: http://www.ericsson.com/technology/whitepapers/convergence_b.pdf The Etnews (2007) ‘Telecommunication charges lowering’, Available at: http:// english.etnews.co.kr/news/detail.html?id=200702230007 The UTMS Forum (2005) ‘Towards mobile broadband and personal internet’, Available at: http://www.umts-forum.org/servlet/dycon/ztumts/umts/Live/en/umts/MultiMedia_PDFs_ Papers_Towards-Mobile-Broadband-Oct05.pdf Tolman, E.C. (1932) Purposive Behavior in Animals and Men. New York: Appleton-CenturyCrofts. Urbaczewski, A., Valacich, J.S. and Jessup, L.M. (2003) ‘Mobile commerce opportunities and challenges’, Communications of the ACM, Vol. 46, pp.31–32. Urbaczewski, A., Wells, J., Sarker, S. and Koivisto, M. (2002) ‘Exploring cultural difference as a means for understanding the global mobile internet: a theoretical basis and program of research’, Paper presented in the Proceedings of the 35th Hawaii International Conference on System Science, Big Island, HI. Venkatesh, V. (2000) ‘Determinants of perceived ease of use: integrating control, intrinsic motivation, and emotion into the technology acceptance model’, Information Systems Research, Vol. 11, pp.342–365. Venkatesh, V. and Brown, S.A. (2001) ‘A longitudinal investigation of personal computers in homes: adoption determinants and emerging challenges’, MIS Quarterly, Vol. 25, pp.71–102. Venkatesh, V., Ramesh, V. and Massey, A.P. (2003) ‘Understanding usability in mobile commerce’, Communications of the ACM, Vol. 46, pp.53–56. Wakefield, L. and Barnes, J. (1996) ‘Retailing hedonic consumption: a model of sales promotion of a leisure service’, Journal of Retailing, Vol. 72, pp.409–427. Woodruff, R.B. (1997) ‘Customer value: the next source for competitive advantage’, Journal of the Academy of Marketing Science, Vol. 25, pp.139–153.

Maintaining continuers vs. converting discontinuers

129

Wu, J-H. and Wang, S-C. (2004) ‘What drives mobile commerce? An empirical evaluation of the revised technology acceptance model’, Information & Management, Vol. 42, pp.719–729. Zeithaml, V.A. (1988) ‘Consumer perceptions of price, quality and value: a means-end model and synthesis of evidence’, Journal of Marketing, Vol. 52, pp.2–22. Zeithaml, V.A. and Bitner, M.J. (2000) Services Marketing: Integrating Customer Focus across the Firm, 2nd edn. Boston, MA: McGraw-Hill.

Notes 1

Our focus in this study was on user perspectives. Thus some terms that could, in another context, refer to objectively measurable factors, refer here to subjective measures. Usability is perceived usability; usefulness is perceived usefulness; and so on. 2 Some questions that had been used in prior studies were omitted because our pre-test users could not differentiate them from other questions. For example, the question ‘MDS fits my lifestyle very well’, which has been used to measure compatibility in prior studies, was deleted because our pretest subjects could not distinguish it from a question like ‘MDS fits my daily routine well’. We believe this is partly because the questions were translated into Korean, which has different lexical and semantic structures than English, and partly because MDS is a distinct technology with unique features. 3 Users who had used MDS between one and three times in the past month were not in either group, and dropped from the sample consequently. 4 It was found that the Ȥ2 GOFs of the two models were well above the threshold. However, the Ȥ2 GOF criterion is sensitive to sample size. As sample size increases (generally above 200), the Ȥ2 test has a tendency to indicate a significant probability level. Therefore, we believe that the high chi-square values are a result of our sample size, and do not indicate that the models are inappropriate. This belief is supported by the fact that all the other GOF indexes are within acceptable ranges.

130

H. Kim, I. Lee and J. Kim

Appendix 1

Questionnaire

Construct Usefulness

Usability

Questionnaire

Reference

UF1

MDS requires less time and effort than other ways of performing the same task.

UF2

MDS enables me to perform many tasks better than I could using other methods.

(Parthasarathy and Bhattacherjee, 1998)

UF3

MDS offers more value than other ways of performing the same task.

UF4

It is better to use MDS than to undertake the same activity another way.

US1

It is hard to learn how to use MDS.

US2

MDS is quite complicated to master.

US3

It is difficult to use MDS.

US4

MDS is complex and hard to use.

SQ1

MDS is quick.

SQ2

MDS responds promptly to my requests.

SQ3

MDS is stable.

SQ4

MDS is reliable.

SI1

I know many people around me who have been using MDS.

SI2

I know many people who have strong opinions about MDS.

SI3

Most of my friends and family have been using MDS.

CP1

MDS is easy for me to adjust to.

CP2

MDS fits my daily routine well.

Ubiquitous Connectivity

UC1

I can use MDS anytime.

UC2

I can use MDS anywhere.

Perceived Cost

PS1

Using MDS takes a lot of time.

PS2

Using MDS requires a lot of effort of me.

PS3

The price of MDS is very high.

PV1

I consider use of MDS an economically sound choice.

PV2

The price of MDS is appropriate considering its value.

BI1

The probability that I will use MDS in the future is very high.

BI2

If I had to choose between continuing to use MDS and stopping, I would decide to use it.

System Quality

Social Influence

Compatibility

Perceived Value

Behavioral Intention

(Parthasarathy and Bhattacherjee, 1998)

(Chin, Diehl and Norman, 1988)

(Sheth, et al., 1991; Venkatesh and Brown, 2001)

(Parthasarathy and Bhattacherjee, 1998) (Chae and Kim, 2003) (Cronin, Brady and Hult, 2000) (Dodds and Monroe, 1991)

(Cronin, Brady and Hult, 2000)

131

Maintaining continuers vs. converting discontinuers Appendix 2

Factor analysis results and reliability check Independent Variables

Type UF1 UF2 UF3 UF4 CP1 US1 US2 US3 US4 CP2 SQ1 SQ2 SQ3 SQ4 SI1 SI2 SI3 UC1 UC2 PC1 PC2 PC3 Cumulative Variance Explained Eigenvalue Reliability

Usefulness

Usability

System Quality

Social Influence

Ubiquitous Perceived Connectivity Cost

0.819 0.811 0.772 0.728 0.523 –0.044 0.024 0.017 0.037 –0.146 0.015 –0.038 –0.043 –0.062 –0.153 –0.170 0.104 0.136 0.135 –0.079 0.406 0.279 0.203

0.017 –0.004 –0.029 0.037 0.236 0.771 0.745 0.742 0.676 0.568 0.180 0.167 0.196 0.240 0.067 0.082 0.041 0.263 0.283 0.011 –0.004 –0.050 0.365

–0.030 0.036 –0.059 0.001 0.147 0.246 0.163 0.142 0.195 0.087 0.802 0.731 0.620 0.614 0.098 0.105 –0.017 0.233 0.336 0.066 –0.218 –0.419 0.442

–0.063 –0.076 0.005 –0.184 0.335 0.050 0.118 0.004 0.118 –0.086 0.039 –0.006 0.093 0.238 0.837 0.825 0.771 –0.201 –0.230 0.073 –0.063 0.159 0.503

0.027 0.061 0.031 0.130 –0.101 0.110 –0.050 0.115 0.109 0.174 0.085 0.028 0.128 0.081 –0.015 –0.006 0.201 0.612 0.559 –0.102 0.167 0.068 0.55

0.118 0.010 –0.035 0.120 0.128 0.015 –0.062 –0.043 –0.063 0.245 –0.040 –0.006 –0.066 0.008 0.047 0.040 –0.105 0.120 0.060 0.795 0.546 0.472 0.594

3.10 0.783

2.86 0.813

2.53 0.739

1.81 0.729

1.46 0.763

1.31 0.461

Dependent Variables Type PV1 PV2 BI1 BI2 Cumulative Variance Explained Eigenvalue Reliability

Perceived Value

Behavioral Intention

0.925 0.921 0.050 0.122 0.494 1.72 0.837

0.070 0.109 0.892 0.881 0.827 1.58 0.740

UF = Usefulness, US = Usability, SQ = System Quality, SI = Social Influence, UC = Ubiquitous Connectivity, PC = Perceived Cost, CP = Compatibility, PV = Perceived Value, BI = Behavioural Intention

132

H. Kim, I. Lee and J. Kim

Appendix 3

Confirmatory factor analysis (CFA) results Factor Loading

Type Usefulness

Usability

System Quality

Social Influence

Construct

Factor 1

UF1

0.92*

UF2

0.97*

UF3

1.05*

UF4

0.96*

CP1

0.79*

Factor 2

UA1

1.12*

UA2

0.97*

UA3

1.12*

UA4

1.02*

CP2

0.74*

Factor 3

SQ1

1.03*

SQ2

0.92*

SQ3

0.81*

SQ4

0.78*

Factor 4

SI1

0.91*

SI2

0.98*

SI3

0.50*

Factor 5 Factor 6

Ubiquitous Connectivity

UC1

1.19*

UC2

1.17*

Perceived Cost

PC1

0.29*

PC2

1.02*

PC3

0.85*

UF = Usefulness, UA = Usability, SQ = System Quality, SI = Social Influence, UC = Ubiquitous Connectivity, PC = Perceived cost, CP = Compatibility *p < 0.01 Appendix 4

Correlation and AVE of adoption factors

Type

Usefulness

Usability

Usefulness

0.59

a

–0.02

0.66

0.35

–0.01

Usability

System Quality

System Quality

Social Influence

Ubiquitous Perceived Cost Connectivity

0.60

Social Influence

0.30

0.23

0.31

0.55

Ubiquitous Connectivity

0.23

–0.18

0.14

–0.04

0.79

–0.01

0.28

–0.22

0.04

0.01

Perceived Cost a

reverse question

0.38