Enhancing the CoBRA® hybrid software cost modeling method for supporting process maturation

Cost estimation is a very crucial field for software developing companies. In the context of learning organizations, estimation applicability and accuracy are not the only acceptance criteria. The contribution of an estimation technique to the understanding and maturing of related organizational processes (such as identification of cost and productivity factors, measurement, data validation, model validation, model maintenance) has recently been gaining increasing importance. Yet, most of the proposed cost modeling approaches provide software engineers with hardly any assistance in supporting related processes. Insufficient support is provided for validating created cost models (including underlying data collection processes) or, if valid models are obtained, for applying them to achieve an organization's objectives such as improved productivity or reduced schedule. This paper presents an enhancement of the CoBRA® cost modeling method by systematically including additional quantitative methods into iterative analysis-feedback cycles. Applied at Oki Electric Industry Co., Ltd., Japan, the CoBRA® method contributed to the achievement of the following objectives, including: (1) maturation of existing measurement processes, (2) increased expertise of Oki software project decision makers regarding cost-related software processes, and, finally, (3) reduction of initial estimation error from an initial 120% down to 14%.