It's still a black box though, isn't it? We don't know how things get flagged, we don't know how thoroughly things get read, etc etc. You're just inferring mechanisms that you admit we don't understand.
The fact that Amazon contains
(a) books violating content guidelines
(b) plagiarised books
(c) autopublished spambooks copied from Wikipedia in violation of the CC license
suggests that the review process is not catching everything. And thus either they *don't* employ enough people to vet things, or their vetting processes aren't working. I simply don't believe everything is getting read and approved.
We also should probably draw a distinction between how things work with KDP and how things work if you go via a third-party...
I was more addressing the point of "Amazon can't afford/wouldn't bother with a full-time gatekeeper it's all or nothing" point, but yeah, it's still a black box. That said, I'm making inferences based on LOTS of data from lots of different authors (I'm part of some other erotica/romance communities). It's like studying a black hole - you can never see the black hole, but you can see all the stuff around it, and based on enough stuff around it, you can start to paint a picture of what it probably looks like. The review process is there, but it's definitely not perfect or even consistent (I could rant forever about the lack of consistency). I never said Amazon was actually reading each and every book - books are all reviewed based on cover, title, and description. I did say in an earlier post that it's not about content, but about presentation. If certain flags are raised, then Amazon might scrutinize actual content (I have a book that I was unable to get unfiltered due to interior content - so they DO sometimes look at the interior). And yeah, this is all KDP goings on. I have no idea how third-party to Amazon deals work.