Deducing software process improvement areas from a Cocomo II-based productivity measurement

At the SMEF2006 conference, we presented our experiences with the set-up of a measurement environment using the COCOMO II-model for software development projects in a company in the banking and insurance area. The set-up was part of a larger research project on managing efficiency aspects of software factory systems. One year of measurements later, a database of 22 projects is obtained. In this paper we will present our conclusions and findings after these first measurement results. The effort multipliers in the COCOMO II-model represent the factors that have a linear influence on the amount of effort needed for a project. As such, they are a management instrument that gives an indication which parameters need attention within the company in order to improve the productivity. In this paper, we discuss a new kind of report we constructed in order to visualise the influence of the different effort multipliers. The goal of the report is two folded. Firstly, one can check whether the factors identified by the COCOMO II-model as effort multipliers indeed have an influence on the effort and therefore on the productivity of a project in this company. And secondly, one can check whether the amount of influence identified by the COCOMO II-model is comparable with the influence we detect in the company. As we will show, even though there was only data available of one year of measurement, useful interpretations could already be given to the results as well as indications about which areas of the software development process need to be focused on in order to improve the productivity.