Re-planning for a successful project schedule

Time-to-market or project duration has increasing significance for commercial software development. We report on a longitudinal study of a project at IBM Hursley Park. The focus of this study was schedule behaviour; however, we explored a range of related factors, including planned versus actual progress, resource allocation and functionality delivered. In the course of the 12-month study, evidence was collected from eight interviews, 49 project meetings, a number of project documents and a feedback workshop. The project leader considered the project to be a success, not only in terms of satisfying resource and schedule objectives, but also in the marketplace. Whilst many of the originally planned external commitments were met, it is clear that the project did not adhere to its original (detailed) plan and indeed there were no less than seven re-plans. These re-plans were mainly in response to mis-estimates in the original plan, rather than in response to the introduction of additional requirements (of which there were several) or problems with external dependencies. Furthermore, these re-plans suggest a distinction between the nature of the initial planning process and the nature of the re-planning process during the project. Attention is also directed at the implications these re-plans have for software metrics and cost estimation research