I don’t have a pretty graph or parity plot to drop in here, but I do keep a very simple spreadsheet where I track vendor test we run alongside blind customer result that gets posted. Vast majority of tests are clustering right around label, and then you have a couple of outliers that everyone already knows about.
For example, on Mast P200 our internal result for that lot was about 197 mg/mL and a customer blind test came back at 153 mg/mL. On Test C50, our internal came back at 58 mg/mL and a customer test landed at 63 mg/mL. Those are the types of points that I have.