Abstract
In order to reduce the efforts of the IT system development projects while maintaining high quality, we have been engaging in standardization and improvement of processes for more than 10 years. In 2007, we set up quantitative project management processes based on CMMI high maturity process areas. These enable project managers to manage their projects in a more proactive manner by utilizing statistical prediction models and quantitative data. This paper describes a practical case about improvement of quantitative project management in our organization. In order to improve the effort prediction model by reflecting upon the unique characteristics of various projects, we examined estimation processes and developed a new model by utilizing stable tendency between the effort variance of a certain development phase and that of a whole project. As measures to reduce the work time spent on prediction, we established automated tools. These improvements are effective for institutionalizing quantitative project management in the organization.
