Wednesday, April 1, 2009

Integrated SCAMPI

I am faced with a problem with determining which CMMI to use (CMMI-DEV, CMMI-SVC, or CMMI-ACQ) for an in-house IT Department that performs all three types of functions for the organization. I face the following questions:
  1. Is there an integrated SCAMPI for all three models held together? Or is the scope is simply determined by adding different PAs from different models? In this case, against what model will the ratings be announced?
  2. What about the cost paid to the SEI? Is it calculated differently for such a SCAMPI?
  3. What about exclusions if all PAs from these three models that are not fully applicable? Is there a way other than pursuing Continuous Representation?
  4. Can you recommend any research work done already on integrating the three model for designing and implementing the OSSP?
I will be greatful for your help and support...

As a first step you should hire an SEI-certified Lead Appraiser, preferably in all three constellations, to provide you the proper advice as to which constellation is appropriate for your organization.

I would only recommend that an organization use the CMMI-ACQ if their primary focus was acquiring products and services from vendors. The CMMI-DEV and CMMI-SVC both have the Supplier Agreement Management Process Area, so either constellation will work if acquisition is not the primary focus of the organization.

Here are my answers to your specific questions:

  1. It is possible to conduct blended SCAMPI A appraisals that cover more than one constellation. But your Lead Appraiser will have to work with your organization and the SEI on how best to perform the blended appraisal and determine your Capability or Maturity Level ratings.
  2. There are NO fees paid to the SEI by the organization for any appraisal. Any appraisal costs are negotiated between you and your Lead Appraiser.
  3. The determination of the appraisal scope is performed jointly with your Lead Appraiser when planning the appraisal. The appraisal scope specifies the representation and the Process Areas being evaluated.
  4. I am unaware of any reported results using blended constellations. Though Mike Phillips from the SEI has said that blended SCAMPIs are permissible. I suggest that you contact the SEI and ask for this kind of information, if it exists.

No comments: