New Aspects of the CMMI SCAMPI Method
- From Discovery to Verification
- Focused Investigations
- For EIA 731 Users
shrimp scam·pi an appraisal whose scope includes only a very small number of the CMMI process areas
chicken scampi an appraisal whose lead appraiser comes from within the organization being appraised
In this chapter, we briefly review several of the major factors that drove the development of the SCAMPI appraisal method. For readers who have some familiarity with legacy appraisal methods, such as CBA-IPI and the EIA 731 method, this may be useful as you transition to the SCAMPI method. For all readers, as you work toward an in-depth understanding of the SCAMPI method, keep these factors in mind along with why things are the way they are and how you can best make use of it.
2.1 From Discovery to Verification
The integrated CMMI models combine the materials from three legacy models; the disciplines addressed include software, systems engineering, and integrated process and product development. Therefore, it is no surprise that the size of the integrated model, an amalgam that covers all three areas, is larger than any one of the single-discipline legacy models. With a larger model, there is more to review during an appraisal. One of the primary objectives of the SCAMPI development was to maintain the quality of the legacy appraisals but at the same time keep the cost of a CMMI SCAMPI appraisal lower than the combined cost of multiple single-discipline appraisals. Several methods were evaluated by the Product Development Team to address the expanded size of the CMMI models; the goal was to keep cost within reasonable limits and at the same time retain a rigorous benchmarking capability. These two objectives were in competition on the opposite ends of a teeter-totter. It was a challenge to keep these two objectives balanced during the development of SCAMPI and it remains so at present.
Figure 2-1 Balancing Low Cost and Rigorous Benchmarking
The CBA-IPI assessment method, which was used with the CMM for Software, was considered a rigorous benchmarking method. However, it was based on the "discovery" of evidence by an assessment team that the goals of the model were indeed satisfied. This was a very labor-intensive process. The team was responsible for finding evidence that each activity was satisfied through a review of documents and interviews. The developers of CMMI and its appraisal methods determined that the time required discovering that all of the CMMI goals were satisfied in the style of the CBA-IPI method would be excessive with a large model like CMMI.
In addition, some believed that the CBA-IPI method allowed for too much variability. Given the same organization, how consistently could different CBA-IPI teams reach the same conclusions on the model satisfaction? The developers of the method considered consistency of appraisal findings to be an important objective of the new CMMI appraisal method. They strove for a low-cost benchmarking method where there was a high probability that different SCAMPI teams would reach the same findings for an organization.
The basis of the solution to these challenges was to create a "verification" appraisal method rather than a discovery method. With a verification method, the organization undergoing the appraisal is responsible for providing traces from the model's goals and practices to the evidence generated by the processes that they use. The appraisal team then verifies that what the organization has provided is valid for each practice. While this reduces the appraisal team's effort, a SCAMPI appraisal is still neither easy nor quick. There are ongoing challenges to further reduce the cost of applying the method while maintaining the benchmarking capabilities of the method.