A Discriminant Model for Classifying Software Project Performance

Abstract

Project managers are concerned about completing the projects on time and cost. IT projects across the globe are notorious for their time and cost overruns. This paper presents output from a comprehensive study on software development risk and project outcome with respect to the projects executed by software companies in India. Based on the data collected from over 300 projects, the authors developed a discriminant model for predicting the project outcome category based on risk scores of a project. The discriminant models developed are seen to possess adequate prediction accuracy to be used in practice. The models can help the project managers in early detection of likely project failures and hence to initiate appropriate counter strategies.