I was eager to try those because Amazon said ads with images other than products got better results.

I've tried both single images and multiple ones (which produce a slideshow). My results are worse than with the old sponsored brand ads.

I selected keywords carefully and checked daily to deactivate nonperformers and put some actual customer searches in the negative targeting area. Results didn't improve.

Is it a mistake to use the keyword targeting? Would product targeting be better?

It may also still be in the experimental phase. I've heard reports of some people not seeing it, even on very high-ranking books.

But since it's only likely to benefit high-selling books, wasn't the ranking enough? This seems a little bit like overkill.

It sounds to me like a keep Trad happy thing.

When they hit a winner, they want Amazon to acknowledge it with something special.

It may also still be in the experimental phase. I've heard reports of some people not seeing it, even on very high-ranking books.

But since it's only likely to benefit high-selling books, wasn't the ranking enough? This seems a little bit like overkill.

I wonder if this is only on amazon.com - my new release has unexpectedly sold over 200 copies this month

it says "100K+ bought or read in past month."

Slight scale imbalance.

100,000 sales/reads in a month to get that. The quoted book has high level orange tags, which I suspect comes first before the extra note.

I have seen this on some other kinds of products for a while.

From glancing at the discussion on FB, I noticed an alleged Amazon response that makes it sound as if the figure is more than just borrows and reads. It might, for example, include listens. (Amazon uses that data in its Amazon Charts system.)