The Development and Empirical Analysis on a Two-stage Risk Based Performance Model of Customized Information Systems Development Projects
WANG Qiu-zhen,MA Qing-guo
DOI: https://doi.org/10.3785/j.issn.1008-942x.2007.04.015
2007-01-01
Abstract:With information technology playing an increasing role in economy, companies have grown more heavily dependent on the successful delivery of information systems (IS). However, the failure rate of IS development projects is high. Advocates of software risk management claim that a key question to reduce the chance of failure of a project is how to deal with the uncertainties of software development. This paper develops a two-stage risk based performance model of IS development projects to study the relationship between project inherent uncertainty, management practices and different dimensions of project performance. Two-stage risk refers to performance risk in development stage and performance risk in implementation stage, which respectively denotes the difficulties of estimating project outcomes during the different stages. Four objective uncertainty factors such as relative project size, technology complexity, development team skill, and client/user experience constitute project inherent uncertainty.In this paper, two dimensions of project performance: process performance and product performance are considered. Based on the dataset collected from 181 customized IS development projects of software houses in Hangzhou City, this paper carries out an empirical analysis on the research model with LISREL 8.7, a structural equations modeling package. The results reveal that performance risk in the development stage and performance risk in the implementation stage are two important intervening variables, which mediate the effects of project inherent uncertainty and management practices on both process and product performance. While project inherent uncertainty increases performance risk in different stages, management practices such as project planning and control, internal integration, user participation and top management support reduce such risk, as suggested by the risk-based approaches. Moreover, the research finds that project planning and control and internal integration have a direct positive impact on process performance, and internal integration, user participation have a direct positive impact on product performance. Further, top management support has a strong positive effect on use participation. In addition, the key factors which influence performance risk are found to be different in different project stages. For performance risk in development phase, key influential factors are project planning and control, internal integration and user participation. For performance risk in implementation phase, key influential factors are user participation and top management support. However, we have not found evidence for significant positive influence of user participation on process performance. The lack of support for a significant effect of process performance on product performance indicates that projects which come in within time and cost budgets may deliver poor systems. The implications of these findings are discussed along with directions for future research.