Amazon is a net-negative overall but I don't think this understanding is correct:

> The only purpose of the ads is to get you to pick an air fryer that isn’t that one (or for the best air fryer, to keep you on track to buy the one you wanted in the first place). The ads make the search worse.

This is only true in extremely simply scenarios. In reality, most of the time individual consumers may prefer something other than "the highest ranked item in this category." Maybe another company has a feature that increases cost but adds utility, maybe there's a brand you trust more that makes a slightly less popular version, etc.

The problems of parasitic advertising still exist. People still see ads for products they will like less than the top rated search, but it's wrong to imagine that Amazon actually knows what that person would like best. If matching people to products was that easy you'd have all kinds of 3rd party recommendation services making very specific recommendations to make money off affiliate links. Instead we have a million "top 10" sites that work about as well as picking the #1 seller in an amazon category: not that well.

I think this comment is mistaken. The blog post is simplistic and doesn't cover the case you're describing of multiple good products for different consumers, but it's ultimately still right. Ads are not a good solution for this use case either because they don't exist to actually match consumers to good products, they exist to selfishly drive sales to their brand owners. That goal is parasitic to the service of product recommendation. Similarly with Amazon itself, their goal is to maximize some combination of sales revenue and ad revenue.

The solution to this is just what you described, a 3rd-party recommendation service. I don't think there's any reason Amazon should be able to make better recommendations than I can with a minute or two of research. Advertising giants only have a marketing profile on people, I think it's fairly limited in terms of how helpful it can be.

The problem with the "top 10" sites you mention is ironically that they are almost all actually advertisement sites, not genuine product recommendation sites. This means they get paid to place products and are essentially corrupt. I would guess the ecosystem is so universally horrible is because SEO costs money and requires them to aggressively monetize to exist. I've read that the Wirecutter is one of the only genuine sites and I find that it usually works fairly well at the price bracket they choose and for certain kinds of home goods.

You raise a very good counter to what feels like the main thrust of the guys article. It put to words something that felt "off" to me while reading it, that there was an assumption made that was never stated and was also incorrect - so thanks for putting that to words.

He seems to assume that there is 1 air fryer that everyone on Earth should be getting, and all others are strictly inferior. There are probably no products that exists where that logic is true. There is so much variation in cost vs quality, or entire feature-sets like you pointed out.

This bleeds over into incredibly "messy" areas too, like societal signaling. Someone might specifically want a more expensive version of something simply because it's more expensive. The author doesn't seem to account for any of this complexity, even the "simple" complexity of cost/quality/features.

The value that these types of ads can add is that Amazon can build up models that can "learn" and get better at matching every unique persons criteria and provide a way for vendors to get "matched" with that buyer. Obviously you can still critique this model, but to assume it's strictly bad feels like it lacks nuance/critical thinking.