M. Louisa Locke has a good post on marketing via KDP Select. I keep seeing these really black-and-white posts about that strategy. You know what I mean: In the Golden Age KDP Select was a license to print money, but since the algorithm change it doesn't work, not even a little bit.
I have to question both sides of that assertion: There were plenty of authors disappointed by KDP Select back in the day, and obviously my experience shows that it can work, if by "work" you mean "improve sales," not "make me a huge bestseller forevermore." It's all a matter of your expectations, I guess--I was actually happy with how the Facebook ads were working. Yeah, BookBub worked way better, but even if it hadn't, I still would have considered the experiment with KDP Select to be a success, both because it improved sales immediately and because it pointed out an approach that I could use to improve sales in the future.
I think a big part of the problem is that people are still looking for the One Magical Thing that they can do for their book once that will take care of everything forever. And you know, that One Magical Thing just doesn't exist. I think the BookBub ad worked because of all the effort I'd put into the book beforehand, and I expect to have to continue to put effort into marketing in order to keep sales going.