This is way overkill,
but contains a discussion of some of the structures and processes they have and how they see them.
I imagine PG has some of the same statistical resources. It would be interesting to see
what PG has considered important enough to measure, and how they interpret the measurements.