Evaluation of SPI efforts in small and medium organizations

COLLA, PEDRO; MONTAGNA, JORGE

Denver, Colorado, USA

Workshop; 2nd SEI CMMI High Maturity Measurement and Analysis Workshop; 2008

Software Engineering Institute (SEI) - Carnegie Mellon University

In this presentation, the use of both a deterministic model based on mathematical transfer functions and a stochastic (Monte Carlo) scenario model are discussed. These models were used to predict the likelihood of net present value from the adoption of CMMI-based process improvement initiatives in small and medium software development organizations in Argentina. The nature of this modeling meant that the following healthy ingredients were not implemented: connect upstream activity with downstream activity and enable projects to achieve mid-course correc-tions to ensure project success. Rather than focusing on a single organization, the work centered on helping the broader community evaluate the likely outcome of process improvement efforts. The intended audience for the work included executives from small and medium enterprises and the SPI practitioners who support those doing business case evaluations. Since small and medium enterprises often have limited resources, another audience for the work was policy makers at the industry level who could pool their resources in identifying shared drivers and data collection. Government officials who were evaluating policies aimed at fostering wider adoption of process improvement efforts in Argentina were also part of the audience. Cooperation is necessary since most software organizations in Argentina are quite small, with a median of 20 people per organization.