Thursday, November 18, 2021

CMMI V2.0 Random Sample Generation - Explained!

What is this "Random Sample" about, and why do I need it?

New for CMMI V2.0, the "RGS" is a way to ensure that projects for your appraisal are not "cherry picked," and that ALL of your organizational unit is, in fact, performing at ML3.

First, a little history.

Back in the bad-old days, there wasn't a lot of guidance about what should be included in an appraisal.  For v1.0, Lead Appraisers could identify whatever scope they wanted, and also exclude whatever Process Areas they wanted.  This was OK for credible and ethical Lead Appraisers, but like any other business opportunity, the scammers showed up and were awarding ratings for companies that only showed one project (when they had hundreds) and excluded important PAs like Requirements Management and Planning!  One appraisal, conducted outside of the US, was for a 50,000 person "Organizational Unit" that looked at one 4-person project without Planning, Requirements, or Measurement.  Oy!

As you might expect, the Quality Department started scrutinizing appraisals more closely, even rejecting some, which caused people to ask "how many is acceptable?"  It's kind of a crazy question, given the unlimited number of combinations there are in the world, but backed into a corner, they said "three."

So - the magic number for almost all appraisals became three - as silly as that was.  At the same time, they clarified that ONLY Supplier Agreement Management could be excluded, and then only if the OU didn't use any suppliers (a pretty rare occurrence.....yet people do it still).

Enter V2.0

When V2.0 came out, the Institute had two goals.  One was to transform the perception of CMMI from a compliance model to performance model.  With the credible application of the new Practice Areas "Governance" and "Implementation Infrastructure," they've come close to that.  The other was to ensure a credible sampling.  That's where the RGS comes in.

Now, when we plan an appraisal we send ALL projects to the CMMI Institute (ISACA) and they send as a "Randomly Generated Sample" (RGS).  We have to send in the project list no more than 60 days prior to the start of Phase II, and accept in no later than 35 days prior to the start of Phase II.  The RGS will select a set of Projects, and then a set of Practice Areas for each Project.  In other words, no more "Focus" and "non-Focus" projects - just a random generation.  The RGS looks something like this:

As you can see from this truncated sample, the RGS, in this case, split the PAs across the two "projects" and assigned the more "organizational PAs to the bottom Support Function (EPG-like team). So, no project has ALL Practices Areas.  This creates a broader sample that requires ALL projects in the OU to be ready for their Appraisal.

For more on CMMI, Agile, and other frameworks, head over to