اثرات عدم اطمینان پروژه و مدیریت ریسک بر عملکرد پروژه توسعه IS : دیدگاه فروشنده

The structural contingency perspective has been widely used in information systems development (ISD) project risk management research. This paper develops an integrative model to explore the moderating effects of uncertainty on the relationship between risk management and IS development project performance from a vendor perspective, rather than the client perspective that is mainly employed in the literature. A survey-based research design is used to collect data to test the proposed model. The results reveal that project uncertainty can moderate the effects of project planning and control on process performance and the effects of user participation on product performance. More specifically, the results indicate that project planning and control makes a greater contribution to process performance when there is a low level of inherent uncertainty and that user participation makes a greater contribution to product performance when there is a high level of inherent uncertainty. The results of this study contribute to a more acute understanding of the contingency approach to ISD project risk management.

With information technology playing an increasing role in the economy, companies have grown more heavily dependent on the successful delivery of information systems (IS). However, information systems development (ISD) project failures are common. The Standish Group Chaos Report for 2009 indicated that 44% of software projects were unable to be delivered on schedule, within budget, or with the required functions, and that 24% of all software projects were cancelled (Standish Group International, 2009). Effective ISD project management has received considerable attention from academics and practitioners. A key question for researchers is how to deal with the uncertainties of software development (Zmud, 1980, McFarlan, 1981 and Wallace et al., 2004) or, in other words, risk identification and management. One branch of IS research discusses risk management, project success, and the relationships between the two from a contingency perspective (e.g., Nidumolu, 1995, Barki et al., 2001 and Jiang et al., 2006). The contingency approach considers project success to be dependent on how well the project as a whole is able to deal with uncertainties in the project environment. With the exception of Barki et al. (2001), contingency studies of software project risk management do not consider uncertainty profiles or risk management profiles from an integrated perspective. Moreover, most of these studies focus on in-house development projects, where developers and users are members of the same organization. However, companies are increasingly outsourcing all or part of their IS activities to external vendors (Lacity and Willcocks, 1998), including IS development. Outsourcing may give rise to additional or different risks from the perspectives of both the client and the vendor (Taylor, 2007). In this situation, client and vendor share the responsibilities for managing outsourced IS projects. As system vendors absorb considerable amounts of risk, an integrated framework is needed for managing the risk in software development from a vendor perspective (Dey et al., 2007). However, prior research on risk management in outsourced ISD projects has paid little attention to the vendor's perspective (Taylor, 2007). The two parties involved in outsourcing may have different perceptions of risk, risk management and project success because of the differences in their goals and structures. Accordingly, the contingency relationships found in prior research need to be examined to determine whether they also apply to the study of the outsourced projects from a vendor perspective. Thus, this paper attempts to develop an integrative contingency framework to describe the effects of project uncertainty, risk management and their interaction on project performance from the vendor's perspective. This paper is expected to advance our understanding of the risk management of outsourced IS development projects and to provide system vendors with a set of guidelines that may be helpful for the effective risk management of outsourced ISD projects. The remainder of this paper is organized as follows. The next section discusses the theoretical background to ISD project management and reviews the existing literature. Section 3 describes the research model and the resulting hypotheses. The research methodology and the results of our model test are reported in 4 and 5. Section 6 discusses the results and the implications of the study. The final section outlines the limitations to the study and the opportunities for further research.

Our results showed that project inherent uncertainty had a direct negative effect on process performance and product performance, that project planning and control and internal integration had direct positive effects on process performance, and that user participation and internal integration had direct positive effects on product performance. However, the direct relationship between user participation and process performance was not supported, which is consistent with the view of Yetton et al. (2000) that user participation tends to increase budget variance by encouraging suggestions for changes to specifications and the empirical finding of Nidumolu (1995) that increased interaction between users and IS staff does not necessarily lead to a project that converges well (i.e., improved process performance). User participation is necessary for project success and participation in the requirements analysis stage can decrease the risk of there being insufficient requirements. However, too much user participation may have a negative effect on project success and delivery time. Given the novelty of the technology involved and the corresponding uncertainty about requirements, clients/users will continually change their requirements, which can lead to conflict and the product being delivered late and over budget. Therefore, managers need to be aware of the potential trade-offs between too much, and extremely limited user participation. The results of the full model with the interaction effects revealed that inherent uncertainty can moderate the effect of project planning and control on process performance and the effect of user participation on product performance. However, interaction effects between inherent uncertainty and internal integration were not found. More specifically, the negative path coefficient from the interaction term between project planning and control and inherent uncertainty to process performance indicated that project planning and control makes a smaller contribution to process performance when inherent uncertainty is at a high level, which is opposite to the finding of Barki et al. (2001). The positive path coefficient from the interaction term between user participation and inherent uncertainty to product performance indicated that user participation makes a greater contribution to product performance when inherent uncertainty is at a high level, which is consistent with the results of Barki et al., 2001 and Nidumolu, 1995. Barki et al. (2001) focused on in-house development projects where developers and users are members of the same organization, rather than the vendor's perspective employed in this study. It is more convenient and effective for project leaders to use standards, plans and formal mutual adjustment through a hierarchical structure in in-house projects. However, it is difficult for vendors to implement formal plan and control mechanisms, as project uncertainty increases in outsourced projects where the developers and users belong to different organizations. Therefore, enhancing the communication and coordination between the development team and the clients/users through informal coordination mechanisms is more important for high-risk projects. This study provided some support for the moderating role of project uncertainty using a PLS modeling method, which is not consistent with the results of Nidumolu (1996a), who used multiple regression analysis (MRA) to test moderating effects. Although the overall interaction effect size (f2) is comparatively small, Chin et al. (2003) emphasized that a small f2 does not necessary imply an unimportant effect. Even a small interaction effect can be meaningful under extreme moderating conditions and, if the resulting beta changes are meaningful, then it is important to take these conditions into account. Barki et al. (2001) found that high-risk projects need higher levels of internal integration than low risk projects. However, this study did not find interaction effects between inherent uncertainty and internal integration. The software development process can be interpreted as a knowledge-intensive system, incorporating the expertise and skills of many different people over an extended period. Given the nature of the software development process, communication and coordination between development team members is necessary for project success and is viewed as a conventional mechanism by system vendors. Moreover, internal integration also includes the integration of technological and business process knowledge. Compared to the development teams in the in-house projects of client organizations, the vendor development teams have the advantage of understanding business processes because of their rich development experience. Due to the previous reasons, the project managers of vendor firms are likely to consider that a high level of internal integration does not play a very important role in high-risk projects. The results of this study reveal that these risk management factors make different contributions to different dimensions of project performance for projects with different levels of inherent uncertainty, which has important implications for practitioners. It implies that proper management strategies should be applied according to project type and key performance criteria. When process performance is a key performance criterion, projects with high inherent uncertainty call for lower levels of formal planning and control. On the other hand, when product performance is a key performance criterion, projects with high inherent uncertainty call for higher levels of user participation. User participation is a client-related factor that cannot be controlled by the project manager, but it can be influenced. Project managers must take reasonable steps to ensure that they have the support and commitment needed to deliver a successful project. Accordingly, project managers require skills in relationship management, trust-building, and business politics. In addition, communication and coordination between development team members need to be strengthened, independent of the performance criterion and the level of project inherent uncertainty.