Evaluating The Benefits Of Management Configuration

The configuration management system with the seven project managers gave the SPI group detailed information about Bruel & Kjaer's software process and many suggestions about how to improve it but the group needed to systematize all the problems, issues, and opportunities raised during the interviews. The group discussed the lessons learned from the interviews on two occasions: after the first interview and after the last interview. Immediately after the first interview, a discussion took place in the group. Each interviewer or observer in turn stated what they saw as the important lessons, and then these lessons were discussed, categorized, and documented in the minutes of the meeting. Afterthe last interview, the SPI group sought to reach a common understanding of what the major lessons from each interview were. Before the meeting, each member of the group had gone through the summaries of the interviews and had collected a list of five lessons to be learned from each interview.

Each interview was then analyzed through the presentation and discussion of these lessons. There were different opinions of what would constitute a major lesson, but for each interview there was a certain sense of consolidation of the views held by the individual members of the group. After going through all the interviews in this way, a conclusion spanning all interviews was discussed. No common conclusion was reached, and the task of compiling the analysis in a short report was assigned to one of the external members of the group.

The first version of the report was a three-and-a half page summary of the major lessons across all interviews based on the discussion in the SPI group. The purpose was to provide a common conclusion that would serve as motivation to the project managers. The report further explained what was seen as the most important motivations and barriers to improvement that were observed during the interviews.

The SPI group commented on the initial report. The internal members of the group had noted that some of the projects had reported serious control problems during the last phase before delivery therefore suggesting that "project conclusion" be added to the list. An additional improvement area, "software reuse," was added at the request of the internal group members, but this cannot be justified from the analysis. After some rearrangement of improvement areas, the final report contained the following seven improvement areas:

•    Descriptive process model PARA Risk management
•    Experiments with prototypes
•    Software requirements specification and requirements management
•    Project tracking and control
•    Project conclusion
•    Software reuse/PARA

The report was now a five-and-a-half page document written clearly with the project managers as the target audience. Each item on the list of areas of process improvement was explained briefly.