It's very fascinating. Since the medic update last year, it has popped in a couple of people in this thread who
knows what the update was about.
Some dude claimed he knew Google people who confirmed that it was all about links, another knew it was about anchor texts and now one guy knows it's about content.
I'm not bashing you
@blinggam, but is there any proof that what you are saying is correct? Anything you could share?
After Medic was launched, I used a custom content analysis script I had made, and compared my content to the top competitors.
I optimized my content by adding more similar densities, LSI, and “context words”. After doing so, the site went from #15 (pre-Medic) #44 (post-Medic) -> #10 and fully recovered from Medic even higher than it was before.
Sadly, on the March update, the site dropped to #30. After analyzing many competitors and talking to many high level SEO friends, we all came to the conclusion that the March update lowered the impact of on-page factors while increasing the value of links.
My personal theory is that people had learned to game the content algorithm by using tools like Cora and Google wanted to strip some of that power away.
I’m not saying that simply the person with the most PBN links win, I’m saying Google increased the value of links they deem to be “good” during the March update and lowered the value of on page content correlations.
I would now argue that this has been rolled back by ~50%, to where content is a bit more of a priority than pre-March, and links are a bit less.
—————-
Let’s take a health site I have. This site was #1 before Medic. After, it went to N/A. It has 100% health-TF PBN links with a few exact anchors.
I don’t focus in this niche anymore so I didn’t even bother trying to recover it.
After March update, it went from N/A -> #6. After “last” update in April, which is when I’m stating they partially lowered he signals on links, it went #6 -> #15.
—————-
I have been analyzing updates and managing many many sites for 7+ years. Admittedly, I can’t say with 100% confidence I know exactly what Medic targeted.
However, I can say with some confidence that (in my opinion) Medic is essentially a machine learning/AI correlations algorithm that attempts to rank “high EAT” sites higher by finding correlations between sites that are scored higher by quality raters. It just so happens that these content signals were seemingly high on the first correlation dump.
There’s new data coming in all the time, hence why we are in a never ending update cycle. Hence why John M said “the Internet is always changing”. Code language for there’s always new data coming in, and refreshes to signals happening based on what the data shows.
If my theory is true, they found some sort of previously unidentifiable link correlation and increased the weight in March, only to see that perhaps it performed a bit worse and now they’re testing different weights. When I’m talking about what changed, I’m speaking from a “broad view”. There are hundreds of signals, if they change a signal for links, it doesn’t have to mean “any/all” links, it could mean any subsection of links (for example, relevant links), etc.
If you read the quality rater guidelines, they directly say they want the algorithm to eventually be able to rank sites that the quality raters rate highly, higher in the SERPs. Why not find correlations between those sites and increase the weight of signals that match those correlations? That’s how, with unlimited processing power, you’d achieve what they are trying to achieve.