Are you seriously proposing Amazon doesn't realize what's going on with their review systems?

Knowing the problem exists isn't the same as having granular data on the scope and nature of the problem. That said... I'd bet they have some data.

No, as the sibling's post indicates I am suggesting they don't have sufficient data to bound or correct the problem. It's a problem of getting the correct data, then extracting signal from that data. People here seem to be trivializing that part.

Mostly because the dozen-or-so data-analytics services that AWS provides are all architecturally of the "scratching-your-own-novel-itch" shape rather than the "best-practices, well-known-standard" or "bug-for-bug-clone-of-the-thing-some-other-cloud-vendor-is-doing" shape. Amazon built DynamoDB, acquired Redshift, etc. to use them.

None of that tells me they have identified the correct metrics to collect in order to identify the extent of the problem, to identify the specific listings or reviews that are fraudulent, or to use in a system to suppress the fraud. They certainly have the infrastructure to do it, but that is the (relatively) easy part.

You think everyone on the Internet except people at Amazon knows about this?

I don't think "everyone on the Internet" has any idea as to the true scope of the problem, nor do I that no they have a scalable solution.

> People here seem to ascribe a level of omniscience to tech giants that I'm not certain they deserve.

I'd describe it as a level of basic knowledge about their own product. You don't have to be a god to analyze Amazon's data, especially if you're Amazon.

Assuming you know what data to collect and store. That isn't a trivial task. Observability on a large and complex system like this is hard. I don't think it is reasonable to assume it's easy for Amazon, and thus their inaction is based on exploiting this bad behavior.