2.1. Cloud-Based Bookstores
Cloud computing, a type of computing model aimed at providing end users with a reliable, customized, and dynamic computing environment [
12], is a current trend that reveals the architecture of next-generation applications [
13]. Cloud computing delivers computing resources as a service rather than as a product, so organizations can dynamically acquire and leverage required services via network connections [
14]. Thus, cloud computing is regarded as the most promising business opportunity for the information technology industry after Web 2.0 [
15]. Due to security and standardization issues remaining as big issues in the business environment coupled with a lack of consistently successful business models, many organizations have chosen not to adopt cloud computing at this time [
13].
Despite most organizations taking a wait-and-see attitude about embracing cloud-computing technology, several organizations have already utilized cloud-computing technology to further the development of an online e-book environment. Some well-known cloud-based bookstores include Kindle eBooks (Amazon), iBooks Store (Apple), NOOK Store (Barnes & Noble), and Google Books (Google), which provide e-book transaction services to their consumers. One of the better known cloud-based bookstores is Raz-Kids. It is a teaching-aid product that provides comprehensive learning resources for both teachers and students alike. This platform offers hundreds of e-books at 29 different levels of reading acuity. Students can easily read content at an appropriate level determined by teacher-student agreement. Beyond cost savings, e-book users may use the cloud-based bookstores to collect and exhibit e-books, which offers them the benefits of movement, flexibility, and value-added functionality whenever they need to search and manipulate digital information [
4]. A number of studies have found factors that encourage users to use e-books [
1,
16,
17,
18] or, they examine how users perceive e-books in general [
19]. Most of these studies adopted various theoretical perspectives, such as innovation diffusion theory [
18], task–technology fit [
11,
16,
17], technology acceptance theory [
1,
17,
20], or expectation confirmation theory [
10], to undertake their respective studies. In [
14], the authors deeply explored the diffusion of innovations theory that includes Rogers’ Diffusion of Innovations curve to create innovation categories suitable for understanding e-book usage. In [
11], the authors proposed that e-book usage remains dependent on how individuals perceive the fit of this technology tool to the tasks they undertake, determining what value-added functions are provided by the content information delivery technology used to enhance reader performance. In [
20], the authors applied the technology acceptance model (TAM) to address the causal psychological mechanisms posited by the TAM, which found that perceived usefulness is more significant than perceived ease of use to satisfaction with e-books, and the greater satisfaction with e-book usage prompted the willingness to continue using e-books. However, few studies [
7,
10] further focused on how to motivate users’ continuous usage of cloud-based bookstores or digital libraries. Furthermore, it may be supposed that there is a need to further the study of e-books from adoption/usage to how to better manage a large volume of e-books via cloud-based bookstores, which forms the basis of our study.
2.2. Continuous Use of Information Systems
In the field of marketing, Oliver [
21] provides Expectation Disconfirmation Theory (EDT), which is used mostly for studying consumers’ repeat-purchase behavior (see
Figure 1). Based on EDT, Bhattacherjee [
22] developed an Expectation Confirmation Model (ECM) to forecast individuals’ continuous behaviors to use information systems (see
Figure 2). EDT is based on the assumption that the expectation will change according to a consumer’s experience, albeit perceived performance is not easy to measure [
22]. To overcome these certain limitations, Bhattacherjee [
22] incorporated perceived usefulness into the model to explain why an individual’s perceived belief affects his/her repeated use of an IS (information system). Per ECM, users first record their original expectations (prior to use) before starting the use of an IS. After gaining usage experience, the users of an IS will be able to judge their own perceived usefulness of an IS. Meanwhile, users will evaluate their original expectations to establish their perceived level of satisfaction regarding use of an IS. Finally, the extent of users’ satisfaction (dissatisfaction) will drive them to continue (or, to discontinue) using such an IS.
Ever since Bhattacherjee [
22] proposed the ECM, it has been used extensively to analyze the continuous use of an IS. However, prior research points out that the ECM has some limitations in explaining users’ continuous use of an IS [
23,
24,
25,
26]. Evidence also finds that factors other than the confirmation of users’ expectations serve as a significant predictor of continuous IS use [
27]. The findings related to expectation confirmation may be insufficient for practitioners to undertake specific improvements of an IS. The findings of continuous studies that have utilized ECM may obtain further insights if some extensions are subsequently made to ECM, which is the other focus of our study.
2.4. Deconstructed Task–Technology Fit Theory
Task–technology fit is reflected by the interaction among task complexity requirements, user abilities and information technology/information system functions [
28]. For the decision-making tasks, [
30] summarizes relevant task classifications that readily belong to one of Hackman’s [
31] four conceptualizations or definitions for a task: (1) task qua task, (2) task serving as a behavioral requirement, (3) it acts as a behavioral description, and (4) it involves ability requirements. However, [
31] also declares that the task-qua-task approach emphasizes the actual materials to be used in a task being shown as part of the decision-making process; the behavior requirements of a task determine both what individuals have to complete to meet designated goals and how they should achieve those goals; and, the last two concepts are unlikely to advance research relative to decision-making tasks. [
32] thinks that task complexity can be integrated into the last two tasks conceptualized by [
31]: the task-qua-task approach and the task as behavioral requirements approach. Furthermore, this complexity lays stress on the task’s characteristics as presented to the decision makers.
Based on the structural contingency theory, the definitions of fit have been identified according to three distinct approaches, including: (1) fit means internal consistency, (2) fit functions as interaction, and (3) fit is treated as congruency [
33]. In [
34], there is further extension of these three approaches into six fit perspectives: gestalts, profile deviation, covariation, moderation, mediation, and matching. [
34] declared that the last three aspects are criterion-free; put another way, their applicability is not unique (i.e., universal), and it is not subject to any particular dependent variable (e.g., effectiveness). However, these three conceptualizations are not suitable to link task technology fit to an effective performance of decision making [
35]. Furthermore, out of the six concepts, the first two are finite in terms of the number of variables, which are taken into consideration; precisely, they are commonly employed in assessing an association between a single predictor variable, a single moderating or intervening variable, and a single dependent variable [
30], all of which are inappropriate for a consideration of task complexity. Following structural contingency theory, [
30] proposes that the optimal definition for technology/task fit is an ideal profile with an internally congruent set of task contingencies and technology elements which impacts the performance of decision. [
36] reported that higher level managers especially favor ample media for information processing and communication. This particular result may also imply that a task can be best supported when the right information technology is chosen.
With portable devices in abundance, the services of a cloud-based bookstore can be employed from any time and/or any place, by any user. Therefore, the use of cloud-based bookstore is removed from a non-mobile IS environment. [
37] thinks there is a substantial difference between mobile and non-mobile IS results from the context relevant to the IS used. With consideration of the technological use contexts or business processes perspectives, the use context has its own role centering in the socio-technical approach to IS [
38,
39] regard the use context as a critical indicator between mobile and non-mobile IS. The ideal concept is one in which a mutual understanding may serve to create a bridge between people and the situated context where users act, thereby leading to a continuous fit of the IS with users’ immediate needs, capabilities, and skills [
40]. Meanwhile, it is important to understand the complexity of the interwoven factors at large. These factors are composed of the user, technology, and the environment that surrounds what is focused on. Furthermore, the study of [
30] also demonstrated that information technology can provide users with the best support only when it is under the right use context.
In addition, a cloud-based system has been proven as a factor that provides end users with a reliable, customized, and dynamic computing environment, referring to the fact that cloud-based systems allow for individuals to access information resources anywhere. A suitable cloud-based system is frequently limited due to location, time criticality, functionality, etc. [
39]. To account for these limitations, it has been suggested to include the individual use context into the design of informational systems [
37].
In a socio-technical system, the communication application environment often plays a central role in IS utilization. Thus, in cloud services, it is necessary to comprehensively understand the complex and dynamic network of interrelated factors in which user behaviors and available technology surround global factors [
39]. For the overall technology evaluation that affects the technology usage and performance, [
41] claim that the performance evaluation of technology should include user–context characteristics in addition to task and technology characteristics. [
39] recommend that the task–technology fit can be further deconstructed to realize the effects of fit on the outcome of an IS. Thus, this study deconstructs the TTF into two segments: one is ideal task–technology fit, and the other is individual use context–technology fit, since “task–individual fit” does not consider the technology characteristics, which are consistent with the assertions of [
39].
In [
29], the authors argued that TTF is an important user evaluation concept in forecasting the utilization of a specific technology. Actually, TTF is a developing construct, and diverse forms of the TTF-based model currently subsist [
16]. In their study of information system continuance, [
27] found that perceived technological characteristics are insufficient to increase continuance intention. Thus, it is necessary to extend the TTF construct to information system continuance by integrating other concepts. Therefore, this study combines the TTF and ECM to propose a comprehensive explanation of cloud-based bookstore continuance.