Comment by rightbyte
9 months ago
This article is really interesting. So, essentially each of the "top publishers" are some sort of SEO ring with different websites, linking each other I guess, or at least writing the same thing, to make Google's scraper believe it is high quality content? Or something similar.
I don't understand the details really, but I have suspected something similar a long time.
Basically yes. I used to write regularly for a large finance site. You've read articles from this site, they show up on HN regularly and its one of the top 100 sites in the world and bounces around in place among the top 3 finance sites.
Anyway, point is when I wrote for them, we had a list of places we could link to and places we couldn't. A lot of them where other sister sites the parent company owned. If we needed a source we first had to try to find it there. If it wasn't available there then there were another list of sites we were allowed to link to, which was basically top 100 sites.
Then there were sites we were NEVER allowed to link to. I remember one of those sites was Reddit. But there were many others. Then anything that fell in the middle was something we linked to if it was critical for the article, otherwise we wouldn't link at all.
So yes, its basically an SEO circle jerk at the top. Which is why you see the same 100 sites in all search results.
Do you have any theory of why Reddit was banned? To prevent some "astroturfing" blackbox algorithm to flag the site?
Yeah here's a very detailed report on what's been going on written by a friend of mine.
https://detailed.com/google-control/
No, there's no SEO ring or anything like that.
The top publishers are just genuinely the sites that people click on and link to the most. There's no objective definition of "high quality content", there are just the links that people click or don't click when presented with search results.
Google puts the links people click the most at the top. And people tend to click on results from sites they recognize, because the internet is filled with a ton of crap, and publications whose names you recognize are at least usually indicative of some kind of minimal level of quality.
That's all that's going on.
It's worth noting it's not just "people preferring well known sites" it's "Google preferring well known sites as part of EEAT"
You are explaining one component of SEO, which is click-through rate. Google will A/B test certain sites one spot higher or one spot lower to see if click-thru rate is positively affected. So if SiteA gets 80% of clicks while in spot #2, but SiteB gets 84% of clicks while in spot 2, then SiteB moves up to spot #2. The cycle continues as sites move up and down a few spots for fine tuning.
However click-through rate is only one component of SEO. The biggest and most significant component of SEO is backlinks, which are ranked by quantity and quality of the link. So if google trusts SiteA a lot and SiteA links to SiteB, then google starts to trust SiteB. If other trusted sites also link to SiteB then the domain reputation grows for that site. Then there is the same ranking on a per-page basis as well. So if one page in particular is very well-linked to, then Google starts to link that page higher and higher for relevant results. This is the largest and most significant aspect of SEO. The click-through rate is a fine-tuning algorithm once you get to the top, but it alone isn't going to help. Google will never test SiteA with SiteZZZZZZZ on page 12.