The life cycle effects of software process maturity

Rapid innovation and intense competition characterize the information technology (IT) industry. In order to survive, IT firms must develop software products quickly, productively, and with high quality. Of major concern are the inter-relationships, often viewed as trade-offs, between quality, cycle time and effort in the software development life cycle. The research today lacks significant empirical studies that rigorously measure these tradeoffs. We investigate empirically the relationships between software process maturity, quality, cycle time, development effort, and infrastructure costs for a major IT firm over a twelve-year period. We first confirm the positive relationship between software process maturity and software quality at all stages of the software life cycle. We find that improved process maturity has a cascading effect through the life cycle, magnifying quality benefits in later life cycle stages. We next perform a comprehensive analysis of the effect of process improvement and quality on all activities in the software development life cycle. We find that improved process and higher quality lead to shorter cycle time and reduce development effort. Next we examine the effect of process maturity and organizational inertia on infrastructure support costs. We find that improved process maturity significantly reduces infrastructure costs. We also quantify the cost of organizational inertia in responding to these savings, providing a measure for organizational redesign. Finally, we examine the effect of software process maturity and development quality on the nature of software defects. We find that higher process maturity reduces high severity catastrophic errors at a faster rate than low severity cosmetic errors. Process improvements that reduce catastrophic errors can have an immediate impact on reducing development time and costs. The benefits of process maturity propagate throughout the software life cycle, leading to substantial returns in quality, development costs, cycle time, and infrastructure costs.