Essays on managing distributed software development.

Abstract

Prior research shows that despite advances in communication technology, distributed software teams face several challenges in bridging space over time and achieving high performance. However software firms are increasingly adopting strategies to leverage global resources and distributed software development is becoming pervasive in the software industry. It is important to understand how the effect of distance could be mitigated in distributed software development and how software firms embarking on distributed development could achieve desired results. In this dissertation I examine the role of structured software processes and process based resource allocation models for managing distributed software development. First, I examine the impact of investments in structured processes deployed by a software firm that operates at the level 5 of the Capability Maturity Model. I analyze data collected from a field study of forty two projects completed by the firm to build and test an econometric model of distributed software development. I find that investments in structured software processes improved distributed software project productivity and decreased software failures. However, qualitative data I collected in the field study indicates that the widely accepted and influential CMM software process model does not adequately address the specific needs of distributed software teams. Then, using an action research approach, I identify twenty four key process areas that specifically address the social needs of software teams and arrange them in an evolutionary process maturity framework. Implementation of the process maturity framework at a leading enterprise software development firm indicates that teams with higher process maturity ratings according to the framework had increased levels of customer satisfaction scores and decreased number of failures. Another key inference from the implementation of the distributed process maturity framework at the firm was that top management must clearly articulate the goals for process induced organizational changes they pursue and employee support at all ranks is necessary for successful adoption of structured software process models. Finally, to address the challenge of resource allocation in distributed software development, I developed an analytic model of offshore software development that accounts for labor cost differences, distributed process costs and the development methodology employed for software creation. Applying the model to project data from a leading offshore software firm, I find that significant improvement in project performance could be achieved by utilizing the heuristic developed in the analytic model.