So element 9 is the dimensional report. Typically this requirement is met by a supplier furnishing 6 sample parts that are fully laid out to print. And typically, everyone wants these parts in spec. I don't like this. All it tells me is the supplier can, in fact, make 6 parts to print. It's not statistical, and it also doesn't let me know they can detect good from bad. The counter proposal would be - instead of doing a full layout on 6 good parts, throw some bad parts in there and correlate the measurements between the supplier and customer. It's not much more work (unless you just typically review the suppliers ISIR and don't actually check them). And what you get is better information. You detect that you both agree on what is good and what is bad via the correlation. An exercise you will go through under the gun the day the supplier decides to run too close to the limits The only down side is, you "6 Golden PPAP samples" aren't so golden - there's bad ones in there. Which could be a hangup for a certain class of supplier quality engineer. Why not do both? Well, for one, cost. And for two, as I said I don't think there's much value add looking at 6 perfect parts. So why do it? Thoughts?