Maybe a better example where the search is for "e-book color" and the listing for the Paperwhite doesn't match either "e-book" or "color", so it's not ranking high because it's matching the search terms well with the titles but seemingly because of something else: https://i.imgur.com/BoUjPPv.png
Why wouldn't a Paperwhite return with "e-book"? And given the results it's way more reviewed and probably way more popular than the other listings there.
And you searched what could be translated to "sleeve" or "cover". The hard drive sleeve has way more reviews and engagement than the rest, by a massive margin.
I hate Amazon search for many reasons, but this seems to be a reach of examples. Show me an example where the alternatives have equal review counts but are pushed way below the Amazon listings or way more specifically excluding terms. If you had searched for an iPhone case and it returned a hard drive sleeve I'd think there was something funny going on, but not with these examples.
The search is "e-book color", if a product doesn't mention "color" or similar in it's title, description or metadata, I don't expect it to show up. No other black and white e-readers seems to show up, only Paperwhite is the black&white reader that shows up.
> And you searched what could be translated to "sleeve" or "cover".
The search engine seems to think I'm searching for either a laptop cover or smartphone/tablet cover, as every single result on the first page are in those categories. Except, for the Amazon Basics one, which coincidentally also ranks highest.
I'm not putting these out there as "This is definitely proof Amazon is fucking with the search rankings" but more like "These results don't really seem to match what I expect or what I get at other similar websites".
> Paperwhite doesn't match either "e-book" or "color"
> either "e-book" or "color"
A Paperwhite most definitely matches "e-book". Very strongly.
Most fuzzy searches have multiple weighting parameters in the algo. You're right, it doesn't match "color" very strongly, but the fact it has >7k reviews and "5+ million sold last month" compared to 31 reviews its match for "e-book" probably has a pretty massive weighting. It's probably 1,000x stronger in its natural weight in terms of overall sales and engagement to anything else related to "e-book". I mean, its 400, 50, 50, and an undisclosed amount sold last month compared to five million.
That item is the lowest cost of the options depicted, and has more reviews (which are proportional to sales) than the rest combined. I think most algorithms would put it on top.
The results for the search "funda" are all smartphone or tablet covers, or covers for laptop. Except for that one Amazon Basics product which is a cover/protector for hard-drives.
Below its ranking (scrolling, not visible in the screenshot) are products with exact same score + more reviews + "funda" in the title. Shouldn't those, by your understanding, rank above the Amazon Basic product then?
Maybe; depending on their age, the content of the product page (videos, comparisons, etc.), delivery time, and the rate at which those who see the listing purchase it (after searching funda). I am guessing that Amazon uses some sort of Bayesian+ algorithm to figure out what to list.
Maybe a better example where the search is for "e-book color" and the listing for the Paperwhite doesn't match either "e-book" or "color", so it's not ranking high because it's matching the search terms well with the titles but seemingly because of something else: https://i.imgur.com/BoUjPPv.png