That, and the ONLY analytics based feature I used was the "After viewing this item most people purchased" section WAS THE ABSOLUTE BEST for shoppers. It made me feel like I could look at a vacuum, but then be directed to the BEST one. It was like, I'm doing all this crazy research and stuff, but how could a billion other people's research help me here. It seems like they run experiments now where once in a blue moon you might see that section.

I know why they got rid of it - and Amazon has this problem all over the place, but some products just sit, while others, with really no difference just fly off the shelves.

This feature still shows up almost constantly for me... there are a few flavors (some of which appear contemporaneously on a single product page for me) including "after looking at this item people purchased..." and "similar items to this item..." and "people who bought this also bought..."

Wouldn't it be nice if you could just choose what information to get, unless of this dice-rolling and analytics and cookies approach?

True, I think like most market incumbents, Amazon has become lazier lately. The vacuum-obsession problem, the proliferation of counterfeit goods, and subpar third party sellers are all more recent problems.

But if you just bought 1984 for example, there's a good chance you'd also want to buy WE, Brave New World, Erewhon, etc. The algorithm made sense for the book store, but it makes less sense for the "Everything Store".

FWIW, I just got a better book suggestion from your comment than I ever got from Amazon.

I wonder if some of this is just....the very low average item quality on Amazon these days.

My recommendations right now have seven separate ten-dollar tire valve extenders, two screen protectors from the same brand (only one of which is for my phone- and that one I already bought!), a $1,000 drone i accidentally clicked an ad for last week, two more screen protectors from other brands, four Qi chargers, and five cases for my phone (I already purchased a case for my phone through Amazon).

Very persuasive!

Their book recommendations were generally relevant to your interests. Much like early Netflix, once they had enough data on what you liked they could correlate that against other customers with similar tastes.

I doubt it is. Recommendations can be great for certain kinds of purchases -- books, Amazon's original niche, being an obvious example: "people who looked at this book also tended to look at these books" can be a meaningful signal. But I don't think that's what helped Amazon to astronomic growth, particularly when we're comparing it to Sears.

The recommendations were useful to me while I was still browsing, not so much after I'd made a purchase. It made browsing a lot easier.

> I've never once acted on an Amazon recommendation

i have!

Yup.

I would sacrifice so much karma for more upvotes for this

Worth it