Skip to Main Content
The perception that software projects have unusually high failure rates has fueled the debate on software process. In the 1990s, the Standish Group estimated that the total economic toll of cancelled or overrun software projects could reach several tens of billions of dollars in terms of wasted effort and opportunity costs. Worse yet, even when delivered on time and on budget, software rarely has any significant salvage value when it fails to meet user needs. This article argues for an economic basis for rationalizing process selection decisions. It demonstrates how, under conditions of uncertainty, learning and flexibility affect such decisions. Uncertainty is a critical driver in process selection because it's ubiquitous in software development and it determines the degree of flexibility and learning needed to maximize economic value.