Take a look at my paper posted on Data Envelopment Analysis. It might be an alternative.
Is this benchmark within one organization? How are you choosing which will be the ideal benchmark? I would think part of your thesis should focus on determining the factors for the benchmarking verification process.
KPI might be Change Controls issued during the process. Also, if these are done in classic Project Management fashion, I would think you should have clear resources and time expectations set out initially. So, maybe you could develop a %deviation from the initial expections as a KPI.
Given who the customer is, you might be able to have a short survey/ follow-up with them on how well the product meet expectations.
Anyway, just some random thoughts for you. Not sure if any of it helps.