Updated Findings - Mediavine Sites Analyzed

[ClickityClick]

80% of the sites have MV. Most are decent size sites with 500+ posts, data from semrush.


Travel Niche:

218k -> 133k (no big social media following)

1,5mil -> 1,5mil spared by Sep, hit by Oct 1,2mil (no social media)

300k ->300k spared by Sep, hit by Oct -> 200k (no social media)

75k -> 75k (20k YT active, 15k insta, 20k fb good engagement, few k insta/twit pared by Sept update -> 60k after Oct spam update

270k -> 270k (5k fb - good engagement, 50k insta, 3k YT) spared by Sep update, -> 230k after Oct spam update

210k (no change in traffic) (5k fb - good engagment, 12k pin, 10k insta)

540k -> 160k (decent social media following almost 0 engagement, fake followers/subs?)

225k -> 97k (137k insta followers, 20k pinterest (follow back fake?, 167k tiktok)

80k -> 80k (spared by the update) (not huge social media following on other platforms apart from insta 13k)

100k -> 35k (no big social media following)

300k -> 90k (no big social media following)

150k -> 100k (no big social media following)

400k -> 140k (no big social media following)

170k -> 170k (15k YT channel, 100k fb) spared by Sep update, -> 130k after Oct spam update

Food niche:

15 sites

4 down more than 30% down (3/4 no social media following, 4th smalish few k social media following but active

2 down 30%-10%

5 unaffected (less than 10% down or few % up)

4 up

Travel niche all around lost way more than food, (probably because more food bloggers have big and active social media?)

The unaffected ones in the food niche all have total 30k+ social media followers/subs through all their social media channels, and active engagement.

Good luck trying to rank in 2024 without big social media presence and high engagement.

 

[JOoa0ky]

Now that I think about it, it's not so much "social media" per say but more so the comments/engagements that you get from users. Google seems to be lumping the user interactions on social media to be similar to Reddit/Quora. As we all know, those two got a big boost recently...

Perhaps it satisfies the "personal experience" checkbox with all these "real people" hearting, commenting and liking.

 

[ClickityClick]

Now that I think about it, it's not so much "social media" per say but more so the comments/engagements that you get from users. Google seems to be lumping the user interactions on social media to be similar to Reddit/Quora. As we all know, those two got a big boost recently...

Perhaps it satisfies the "personal experience" checkbox with all these "real people" hearting, commenting and liking.

Could be faked, I think the bigger part of the pie would be traffic from social media (verifiable Chrome users), then the rest things like comments, followers, likes.

 

[bernard]

I really don't think Google directly uses social media traffic, likes or Chrome traffic to rank results. Would basically be a completely new algorithm with massive data need.

I think Google gets datasets with social media data in it and then train a model with it, by running correlations (what ML is all about) between sites with good social media stats and then the already known algorithm factors (onpage and offpage).

These are two important distinctions.

Am I wrong? Do you think Google directly measures engagement and so on for each site?

 

[CCarter]

train a model with it

That'll be a smarter way of doing it, since then you can predict a website's future ability and boost it knowing you've got a model of what users "want".

Whether they use social media directly or indirectly - I again always think the best solution is to be active and engage with an audience where they are at. That's the end goal of what Google is ultimately trying to measure and get it.

You know they are looking for "Quality Brands". And "Quality Brands" have these characteristics in common. So you can spend time pretending to be a quality brands, blackhat method, or you can be a quality brand and do the things they do.

It comes down to this: a one-man operation trying to compete with a company with a paid marketing team, social media team, and marketing department, and general operations - the one man operation is going to now-a-days get crushed in the SERPs.

The golden age of a one-man operation going toe-to-toe with Walmart or Amazon are long gone. Maybe you can get some obscure wins here and there, but you're not going to go up against Amazon without being on these other platforms and that means having deeper pockets or are very very creative.

That's how I see it.

 

[bernard]

The golden age of a one-man operation going toe-to-toe with Walmart or Amazon are long gone. Maybe you can get some obscure wins here and there, but you're not going to go up against Amazon without being on these other platforms and that means having deeper pockets or are very very creative.

The english speaking internet is a big place and EEAT can work for you as well as against you.

When people write reviews/content and source someone (recommended for EEAT), they're going to link to someone, that someone can be you, if you just bother to take some pictures/do a quick vid and write from a first person perspective.

I just did a search for sleeping bags and found a pretty cool affiliate site that has done absolute A++ reviews. Every box ticked. Pictures, tables, clear language, obvious expertise etc. They even advertise as being affiliates on their About Page. Not a one man crew, but a one man owner, with a team of real people as writers/testers. They have 60.000 followers on Facebook but barely any engagement, a handfull likes on new posts. 4000 on Instagram, some likes, but no comments. They rank great. Support my thesis. They rank, because they make great content, content that people appreciate (with a follow). Not linking them, but they're easy to find, if you google.

Big slice of a smaller pie is what I see as easily the best way to go about "niche site" things now. We're back to micro niches, but this time, you need to be on point with your conversions and maximizing each user, which would mean hitting them up on multiple levels (email, social, video etc).

You might also want to not monetize your site at all for the first 6-12 months and just network. This also isn't a new tactic.

That's if you want to go this route. Everyone can decide. I'm not going to spend my time on earth testing gadgets, but if you're a gearhead, then go this route is my recommendation.

 

[Ryuzaki]

You gotta watch out for correlation and causation and confirmation bias. Some people use data to trick others and some trick themselves with it.

Sites doing well on social media is a direct consequence of trying to do well on social media but also having a high quality site to point users back to. Reddit and Quora rank well despite forums traditionally not ranking well because people tag "Reddit" to the end of their searches. Which is ironic that Google thinks this is the solution (Users didn't want Reddit, they needed Reddit because the rest of the results are garbage, just like Quora is).

Google is known to not use signals it can’t obtain on its own. There are too many sign-in walls and too expensive of API’s with limits to make it feasible.

What they can use is click flow data from the SERPs, analytics data for referral traffic (even then, app traffic registers as direct), and Chrome CRUX data.

You should set a hypothesis and test for it, but the data above doesn’t support the conclusion so it’s time to reform the hypothesis, if your looking for silver bullet causality, which you shouldn’t be here. You should look at big data and correlations, and this isn’t big enough data.

Google’s entire game right now is to save money by creating classifiers (we figured this out years and years ago in our discussion on blessed or cursed sites as a way to avoid "honeymooning" every page). It’s definitely about training models on segments of their index and then extrapolating it out to the rest. Which is why there’s always misses and tweaks after each core update. This one has some catastrophic failures.

One thing I can promise you is it isn’t solely about social media. That’s one tiny signal in a vast algorithmic ocean. But you can be sure that sites performing well on social media put forth the effort to satisfy other variables where Google “can” analyze deeply.

 

[bernard]

One thing I can promise you is it isn’t solely about social media. That’s one tiny signal in a vast algorithmic ocean. But you can be sure that sites performing well on social media put forth the effort to satisfy other variables where Google “can” analyze deeply.

The basis of Google has always been links to sort things out in where to originally rank you. The onpage stuff is where in the index to place you. If there's nothing under that index, you go first, otherwise links aka citations (scientific origin) deserves where you go.

Then when you go into the "virtual library", they sometimes put you front and center on the exhibit right inside the door. Do people then pick you up and like what they see? Do they ignore you? Do they pick you up and put you down quickly? That's what decides who gets to stay at those 1-3 spots.

Which links correlate to high social engagement? Vendor links because you featured their project in a cool way? Wikipedia links because you were the originator of a meta-meme?

Which links do not correlate to high social engagement? That's as much a tell as the reverse.

No links, few links at all, over a long time, that's a big negative tell. No editorial links, over a long time, an even bigger tell.

 

[Jesse]

Google is known to not use signals it can’t obtain on its own.

Genuine question because I'm pretty sure I don't know as much about this as you: Can they get this data from their control of Chrome and Android?

 

[Clint Butler]

Genuine question because I'm pretty sure I don't know as much about this as you: Can they get this data from their control of Chrome and Android?

They used to get social data directly from some top companies; I don't know if they still do now, with everyone competing for ad revenue.

 

[bartman]

I remember way back when there was a major update (can't remember the name!) Lots of theories around too many ads above the fold - and most sites reduced the ad load. Over time I have watched the ads increase again.

Some sites are almost unusable these days - I read a great article on Vanity Fair the other day and I had to read it 3 lines at a time - on desktop! Usually, I just close and go back, but it was a good read lol.
Here is the thing though - these sites overloaded with ads, they are ranking at the top.

I opened a page today (from search) - all I could see on desktop was the logo of the site and a full-width huge ad. Seriously, nothing else and the logo was just a shit looking name! No navigation links - nothing else at all above the fold.

I'm sure I'm not the only one who just goes back to search after they hear their pc whirring like mad and almost coming to a standstill! Bad signals for ranking? - seems not.

I have recently come across sites ranking on the first page and when I click on the result, it is behind a paywall. Good user experience? How many hit the back button? In what world is it good to have the search results return articles you can't access?

Things are definitely strange to say the least in the serps after the recent changes. Almost Every theory I have read can be debunked quickly with an example of a site breaking the rules or the theory. And what is it with Rediit - Google has lost the plot sending their traffic there. I find it a shit show for any niche info I look for! Just saying!