Skip to Main Content
Software development estimates are inaccurate and overly optimistic estimates are major contributors to project failure, despite the fact that every completed project is a rich source of information about performance and estimation. Modern development processes promote risk management, the realization of architecture first, the decomposition of the project into iterations, and the assignment of requirements to these iterations. When a project adopts these forms of best practice, it achieves a high degree of technical control and easier management. One difficult project management task is to accurately determine the effort required to complete the project. This article discusses a use-case-based estimation model for determining project effort. This technique calls for looking at the relationship between estimated and actual data to improve future estimates. Using a simple set of metrics, it is possible to generate a credible model for project estimation. The model described here works best in an iterative development process, allowing comparisons between successive iterations.
Date of Publication: July-Aug. 2004