پویایی اثرات محلی و تعاملی بر پذیرش نوآوری: مورد تجارت الکترونیک

Innovation adoption is determined not only by the opportunities and constraints resulting from organizations’ characteristics (local effects), but also by reaction to the adoption of their interdependent and referable others (interactive effects). This study examines the dynamics of innovation adoption by considering both local and interactive effects in early adopters relative to later adopters, and then investigates the electronic commerce adoption as an empirical example. Analysis results show that the crucial stimulating effects in the early market are focused on the nature of innovations, while those in the later market are concentrated on the practical implementation issues of innovations.

Adopting electronic commerce (EC) in business represents an innovative move for an organization, based on Schumpeter's (1934) argument that an innovation is something that reduces costs and increases quality and performance. Brynjolfsson and Saunders (2009) reviewed literature on productivity for the period 1995–2008 and confirmed that information technology played an important role in productivity increases and was an important driver for economic growth. Substantial evidence confirms that enterprises can benefit from EC. However, most of the numerous studies of EC adoption by organizations have been concerned with whether their characteristics affect the decision to adopt EC (Al-Qirim, 2007, Corrocher, 2006, Daniel and Grimshaw, 2002, Dinlersoz and Pereira, 2007, Grandon and Pearson, 2004, Hong and Zhu, 2006, Joo and Kim, 2004, Lai and Ong, 2010 and Mustaffa and Beaumont, 2004). Comparatively few studies have considered the external driving mechanisms which embed organizations within networks of interdependencies (Dos Santos and Peffers, 1998 and Shih, 2008). Many mechanisms may be responsible for accelerating or slowing the adoption process and identifying which mechanisms dominate this process is likely to make an important contribution to the assessment of a particular innovation's prospects (Mansell and Steinmueller, 2000). Hence, there is a need to examine both organizations’ characteristics and the externally interactive mechanisms that affect the diffusion of corporate EC adoption, in order to identify the crucial factors that cause organizations to adopt EC. Additionally, most prior studies have determined particular factors that influence corporate EC adoption at a certain point in time. However, Rogers (1995) argued that, in terms of the time of adoption, different groups of adopters have distinguishing characteristics regarding their adoption of innovations, so that individuals can be classified into adopter categories, based on when they first begin to adopt innovations. The important differences between these groups of adopters suggest that the factors explaining the reasons for the adoption of innovations change over time, as the level of innovation diffusion increases. This implies that the empirical results from research based on the groups of early adopters and later adopters should be different, and that a longitudinal study of the factors affecting innovation adoption, in terms of the time of adoption, is essential. In spite of the fact that numerous studies have tried to identify the factors affecting corporate EC adoption, at a certain point in time, remarkably few research studies have focused on the substance of these shifts, at the different stages of the EC adoption life cycle. On the basis of these two considerations, this study investigates the diffusion of corporate EC adoption by considering the factors that are associated with corporate characteristics, as well as the external driving mechanisms which embed organizations within networks of business interdependencies, and by comparing the crucial factors that influence corporate EC adoption for early and later groups of adopters. Given these two approaches, the objectives of this study are: (1) to identify the corporate characteristics and the externally interactive mechanisms that affect the diffusion of corporate EC adoption, (2) to demonstrate that the effects that influence EC adoption change significantly along the adoption life cycle, and (3) to investigate the substance of these shifts for a number of specified driving forces in the field of corporate EC adoption. To achieve these objectives, this study uses a network autocorrelation model to develop a research framework for corporate EC adoption. In this model, egotistical opinions and behaviors are determined not only by reaction to various constraints and opportunities granted by the conditions of the ego (local effects), but also by the opinions and behaviors of others (interactive effects) (Leenders, 2002). With regard to local effects on the influence of corporate EC adoption, this study identifies the significant factors that affect EC adoption, which were cited in prior EC-adoption-related research. With regard to interactive effects, this study follows the approach used by Shih (2008) to investigate corporate EC adoption using two different social network models; the cohesion model and the structural equivalence model. The models examine the contagion effects of two actors that render them socially proximate, so that adoption of an innovation by one actor triggers adoption by another actor. This study then formulates specific hypotheses regarding expected changes in local and interactive effects during the EC diffusion process. To test the models and hypotheses, this study makes use of a sample of industry-level corporate EC adoptions, based on the data from input–output tables and industrial census data for Taiwan during the years 2001 and 2006. The data from different years enables explicit examination of shifts in the local and interactive effects on corporate EC adoption between two points in time. Subramanian (1996) reviewed literature related to innovativeness and classified it into two major categories. The focus of the first category of research, named ‘innovation process research’, examines the process of innovation diffusion in a market. Research in this field views those who adopt innovations earlier as innovators and labels later adopters as imitators (Gatignon and Robertson, 1989). This research field uses the assumption that categories of adopters consist of individuals with similar characteristics and similar degrees of innovativeness, at their time of adoption (Rogers, 1995). The characteristics of early adopters are distinguished from those of later adopters, in order to understand the different emphasis on R&D and marketing activities along the processes of adoption life cycle. The second category is described as ‘innovation variance research’. This field of research focuses on the association between the innovativeness of organizations and internal and external organizational factors, such as size, slack resources, structure, degree of specialization and environmental conditions (e.g., Damanpour and Gopalakrishnan, 1998, Tabak and Barr, 1999 and Nystrom et al., 2002). This field measures the strengths of the associations, determined by the amount of variance of the dependent variable (i.e. innovativeness) that can be explained by the independent variables. Although these two research fields have different focuses, most scholars agree that innovativeness is a multidimensional construct, because of the nature of the integral results of multidimensional considerations (Subramanian, 1996). Hence, this study integrates both the innovation diffusion process and the association between innovativeness and organizational factors, in order to identify the important factors that influence innovation adoption in the early and later stages of the adoption life cycle. In addition, with respect to the factors, this study examines both the local effects and the interactive effects on innovation adoption. The rest of this paper is organized as follows. “Research framework and hypotheses” section shows the research framework, and specific hypotheses are proposed in association with EC adoption by organizations. “Methodology” section introduces the measures of network autocorrelation models used to investigate the local and interactive effects on EC adoption. “Results and discussion” section empirically examines a sample of corporate EC adoption in Taiwan to test the proposed hypotheses, and managerial implications are discussed herein. The last section draws conclusions and proposes further studies.

The analytical framework proposed in this study explains the diffusion of innovation by considering both the local effects, resulting from corporate characteristics or factors, and the interactive effects (contagion effects), derived from the external influence of the networks of business interdependencies in which organizations are embedded. Additionally, an improved understanding of the shifting factors that affect the early and the later stages of the adoption life cycle reveals the dynamic processes underlying innovation diffusion. The new contributions of this paper to the study by Shih (2008), which examined only the effects of contagion on innovation diffusion at the early stage of adoption life cycle, are (1) to consider both local effects and interactive effects (contagion effects) and (2) to compare the influence of local and interactive effects on innovation diffusion at the early and later stages of the adoption life cycle. This study then empirically analyzed an example of EC adoption diffusion in Taiwan between 2001 and 2006. The theoretical contributions of this study are that (1) it integrates the two main approaches of innovativeness-related research (i.e. innovation process research and innovation variance research) and (2) it includes consideration of the factors that influence the adoption of innovation, not only the organizational characteristics and factors (i.e. local effect) but also the contagion effects (i.e. interactive effects). The specific findings for each hypothesis have already been discussed, but their implications are discussed, herewith. In terms of local effects, the influence of the factors of size, IT intensity, slack resources and industry competitiveness on the adoption of innovation adoption changes over time in the course of the adoption life cycle. Overall, the findings have two implications. Firstly, in the early phase of innovation adoption, the most crucial driving factors seem to be more concerned with the nature of the innovations, including the size of the organization, which increases the risk tolerance in innovation adoption, and industry competitiveness, which motivates organizations to adopt innovations as early as possible in order to enhance competitive strength or to avoid falling behind. Secondly, in the later stage of innovation adoption, the combination of important stimulating factors appears to focus more on issues related to the practical implementation of innovations, including IT intensity, which may cause organizations to postpone innovation adoption until the innovation becomes compatible with existing IT infrastructures and has been proven successful in the market, together with the slack resources, which offer organizations surplus resources to implement innovations. A similar conclusion was drawn by Waarts et al. (2002), in their study of ERP adoption, which revealed that, in the early phase, firms are primarily concerned with the nature of innovations, such as the company's attitude towards the innovation, the strategic importance of the innovation for the firm and industry competitiveness, while later adopters tend to concentrate on the practical implementation aspects related to the number of seats within the firm and the available budget. In terms of interactive effects, both the contagion mechanisms of cohesion and structural equivalence significantly affect the diffusion of corporate EC adoption. The extent to which the two contagion effects affect the diffusion process changes, as the adoption process progresses from early market to later market. The dominant contagion effects for early adopters and later adopters are different. Overall, in the early stage of the adoption life cycle, the most important contagion effect is structural equivalence, which defines the course of innovation diffusion based on similarity of network position and measures the degree of imitation between competitors resulting from the need to conform to prevalent norms within structurally similar actors (Shih, 2008). The contagion effect of structural equivalence is also related to the nature of the innovation, since the mimetic behavior associated with joint innovation adoption by competitors can be viewed as a strategic tactic to reduce or eliminate any competitive advantage established by competitors who have adopted the innovation. In the later phase of the adoption life cycle, the most crucial contagion effect is cohesion, which is associated with diffusion by direct contact and communication, and which measures the influence of innovation adoption by supply-chain partners on an organization that is considering adopting the innovation (Shih, 2008). The contagion effect of cohesion is apparently more concerned with practical issues associated with innovations because, in the later market, innovation adoption results from an organizational response to requests for adoption from its supply-chain partners in order to enhance the efficiency of business transactions via adoption of the innovation (EC). A similar result was recorded by Wu and Chuang (2009), in their research on electronic supply chain diffusion, which demonstrated that, in the early stage, innovation diffusion focuses on the improvement in perceived usefulness and relative advantage, while innovation diffusion at the later stage mainly results from the need to build good cooperative relationships between trading partners to facilitate their transactions. The overall findings of this research show that innovation adopters have different views on innovation at different stages of the adoption life cycle. During the early stage, adopters focus on whether an innovation has a strategic value that can enhance their competitive strength, or minimize the competitive advantage established by competitors who have already adopted the innovation. Given the uncertain nature of innovations, early adopters must also weigh the potential strategic advantage of the adoption of innovation against the potential disadvantages about failure. As the diffusion stage progresses, the strategic value of adopting an innovation become less important, but the problem of its practical implementation becomes increasingly important. Since an innovation generally unfolds from narrow to broad applicability, the compatibility of the innovation increases in many markets (Waarts et al., 2002). Competition shifts from the effective search for new business opportunities to efficient production and distribution during the ‘cycle of discovery’, from ‘exploration’ to ‘exploitation’ (Nooteboom, 2000). Exploitation begins when variety of content that emerges from exploration is distilled into a dominant practice. As a result of reduced uncertainty, demand increases and new providers of compatible products enter into the market (Gilsing and Nooteboom, 2006). Hence, during the later phase of the adoption life cycle, companies with high IT intensity and substantial slack resources demonstrate a high level of organizational readiness to adopt EC and consideration of the risks of failure becomes less weighty. Another practical issue is the external pressure on later adopters to adopt EC, in order to facilitate their business transactions, after supply-chain partners have already done so. Since certain effects in the analytical framework of this research exert a significant effect on the diffusion of corporate EC adoption, further research might benefit from including them in theoretical and empirical studies of innovation diffusion. However, some limitations should be noted when interpreting the findings of this study. Firstly, although the sample data set is census data that includes all available economic data, these data are limited to a specific economy (Taiwan) and to a specific innovation (EC). Brynjolfsson and Saunders (2009) found that the extent to which investment in information technology contributes to productivity and economic growth varies for different countries, industries and firms. The empirical findings may change, or other factors may become more important, when other information technologies or other economies are studied. Secondly, this study conducted an empirical examination of the research framework at the industrial level, rather than at the level of individual companies. The fact that the decisions by individual firms within a certain industry to adopt innovation may differ creates an opportunity for future research to examine the research framework on the basis of the empirical data for individual companies. Furthermore, it would be interesting to examine how networks of organizations could affect the decision to adopt innovation. Networks of organizations can be measured by either the transaction value of sales and purchases between organizations, or the actual flow of individual bites of information between organizations (for instance, systems and instant messaging). Then, social network analysis and text mining can help in the study of the relationship between whether a specific organization adopting innovations and its network position (e.g. centrality) or its network configuration (e.g. clique, blockmodel). Thirdly, this study limited the empirical data to the firms’ domestic operations and did not consider the impact of the global environment. Future research might include the effects on innovation adoption not only of the domestic environment but also of the global environment. Although these three considerations may modify the empirical findings in relation to some specific factors, they do not restrict the overall conclusions regarding the shift of the influences of effects during the diffusion of the adoption process. Instead, these limitations are to be considered as fertile areas for future research.