One of the biggest areas of speculation, contention, and confusion within the SEO universe over the past six years or so has been whether (or how much) social media signals impact organic search rankings.
But even if Google isn’t directly using social share counts in their search algorithms, there ought to be some other explanation out there about why high share counts correlate with high organic search rankings.
Well, that is exactly what we’re going to research in this post.
Are social shares a ranking signal?
People have noticed the connection between social shares and ranking going back to 2010. But correlating rankings and social signals has been a bit of a cat-and-mouse game.
If you’ve done any SEO at all, you’ve probably noticed that the stories that rank well tend to have high social share counts.
These are your unicorns – the extremely popular magical pieces of content that drive a ridiculous amount of traffic to your site. These types of elite "unicorn" content drive 10-1000x better results than all your other content (the donkeys).
Why do top-performing posts often also have a high number of shares? What exactly is causing these observable correlations?
Some SEOs believed that Google was somehow factoring social share counts into the algorithm like links (though not with nearly the same amount of weight).
Social shares figured into Moz's Search Engine Ranking Factors 2015, albeit as a low factor:
"Always controversial, the number of social shares a page accumulates tends to show a positive correlation with rankings. Although there is strong reason to believe Google doesn’t use social share counts directly in its algorithm, there are many secondary SEO benefits to be gained through successful social sharing."
Indeed, there is a strong reason to believe Google doesn't use share counts as a direct ranking factor. Google has said so.
Repeatedly and emphatically.
Google doesn't use Facebook, Twitter, or any other social share counts as a direct ranking factor.
It's not shares, it's engagement
We need a new approach to answer these important questions. Maybe we’re looking at the wrong social metrics. Maybe we should be looking at social engagement rates rather than just the total number of social shares.
What percentage of total unique people who saw your update clicked on it and/or shared it?
Perhaps the relationship is that the social posts that get very high engagement rates (which leads to high numbers of shares) come from the same content that get above-average click-through rates in organic search results pages, which we know tends to result in better organic rankings.
But how can we test this theory?
A crazy new correlation study: Social engagement, organic search CTR, & rankings
So here’s my crazy idea: to compare social engagement rates with normalized organic click-through rates for 1,000 pages.
Previous studies have only looked at external-facing number of shares. But bots and other factors can easily taint share counts. Plus, studies have shown that many social media users share content without actually reading it.
How did I do this? I:
- Downloaded post engagement data from Facebook Insights (sharing and engagement data).
- Downloaded query data from Google Search Console (CTR and ranking data).
- Matched up the data. This was somewhat difficult because neither Facebook nor Google provided me with the destination URLs, so some custom programming was required.
Important note: You have to normalize your CTR for search based on position. Obviously higher average positions have higher CTRs than lower positions, so I’ve used my Donkey detection algorithm to compute the expected CTR by position to help determine whether the CTR is above or below expectations.
The results: Organic search CTR vs. Facebook post engagement
Here's what I'd consider a pretty strong link between higher social post engagement and higher organic CTR (and vice-versa):
Here, a 100% Relative Search CTR corresponds to a keyword/page achieving the expected CTR for organic search for a given ranking; 200% percent is double the expected search CTR; 50% is half the expected CTR, and so on.
What I found was that Facebook posts with extraordinarily high engagement rates – anywhere from 6 to 13 percent – also tended to have above expected organic search CTR.
Why? My theory: The same emotions that make people share things also make people click on those things in the SERPs. This is particularly true for headlines with unusually high CTRs.
The correlations were much stronger with unicorn content. The R-squared values were well above 0.5 – the model is stronger the more of an outlier you're pushing. Unicorns with high social engagement rates almost always had high organic CTR, and vice versa.
The correlations were substantially weaker with donkey content. The R-squared values were pretty noisy, around .1 to .4. Donkeys sometimes had high engagement rates, sometimes low engagement rates. The same was true with CTR, some high, some low.
So this research illustrates how high social engagement rates correlate with high CTR, and vice versa.
Really, the argument isn't whether social sharing causes organic search rankings or organic rankings cause social sharing.
It's about how engaging your content is.
Actual examples
Theory is great. But let's see if the theory matches by looking at some top-performing content.
Here are just three examples of posts from my company that have top organic rankings on Google and above-expected organic CTR. What was the engagement rate on Facebook?
This post has brought in nearly 500,000 visits from organic search. It had a 7.4 engagement rate on Facebook.
OK. Once is just a fluke.
This post brought in more than 250,000 visits from organic search. It got an 8.5 engagement rate on Facebook.
Two times? Could just be a coincidence.
This piece brought in 100,000 organic visits. It had a 7.1 percent engagement rate when shared on Facebook.
Guys, now we have a trend! All of these posts that rank well had 3x or 4x higher engagement than my average Facebook post.
I could keep posting more examples like these, but it would be more of the same.
Correlation or causation?
What is causing the correlation? There is one thing that makes me certain that the relationship between social engagement and organic click through rates is a co-dependent, causal relationship.
Machine learning.
Machine learning systems actually reward high engagement with higher visibility.
Higher visibility means higher organic rankings and more social shares.
To determine success, an algorithm looks at whether users engaged. If more people engage, that's a clear sign that their algorithm is showing this right content; if not, their systems will audition other content instead to find something that does generate that interest.
Here's a greatly simplified look at the role machine learning systems play in the Facebook news feed and Google search results. Basically, it's all about rewarding content that has above-expected engagement:
When a piece of content fails to beat the expected engagement, it won't get that same visibility, whether it's on Google, Facebook, or any other system that measures user engagement.
Whenever someone searches on Google for something, Google wants to return the best result. Out of all the potential results Google could show for any given query, Google must find what's most useful and relevant.
One way Google checks itself is to look at organic click-through rate (but not the only way!). Did users click on the result in Position 1, or did more people click on the Position 2 or 3 result?
Even though all three of these pages may answer a user's need, click-through rate is a huge clue about whether Google is providing the best answers in the right order for users.
Now let's think about Facebook. Whenever a piece of content gets hot, it means lots of people are talking about it relative to the number of people who see it, in a short period of time. Are tons of people liking, commenting, and sharing a post?
When this happens, Facebook's machine learning algorithm gives these posts or topics greater visibility. It becomes a virtuous cycle:
- Post gets lots of user engagement (shares, likes, comments).
- Facebook rewards the engagement by showing it to more users.
- Higher visibility results in the post getting lots more user engagement.
- Facebook rewards the engagement by showing it to more users.
- And so on, until the the social post is no longer new and engagement dwindles.
What to do?
Turn your best social stuff into organic content and vice-versa.
Since stuff that does well on organic social tends to also do great in paid social, it follows that your content that gets top organic rankings will make great content for paid and organic social.
Conversely, your content that gets tons of engagement on social media platforms (paid and organic) will likely rank highly organically for the topics that they cover.
These unicorns I've been obsessing about forever matter. Big time. Is your content a sparkly majestic unicorn or a boring old donkey?
At the heart of a unicorn is a truly remarkable, inspiring idea. Truly exciting ideas (not just ideas you think are awesome). Content with remarkably high engagement rates has high conversion rates and does incredibly well in paid and organic search and social media, because of machine learning systems that greatly reward remarkably high user engagement.
Conclusion
The old theory was that high social shares correlates with high organic rankings.
But really it's not the number of shares that matters. It's the engagement rate.
Remarkably high social engagement rates correlate strongly with high organic search CTR, which correlates with high rankings. Meaning, click-through rate matters a great deal. Think of it like an invisible hand that helps determine whether your content succeeds (thumbs up) or fails (thumbs down).
What’s happening here is that Facebook Ads, Facebook's news feed algorithm, Google AdWords, and increasingly Google organic search are all systems governed by machine learning systems that reward remarkable engagement with greater visibility.
High engagement rates and machine learning systems are the common factor that explains the correlation between SEO and social metrics.
What do you think? Do your very best-performing pieces of content get tons of social shares, have a high social engagement rates, and drive a ton of traffic from organic search and convert well?
Hi Larry,
I had missed out your post earlier, it's a great post and it endorses our views of importance of focusing on target audience and finding the right ways to engage them.
For general people Off-page is too focused on the high DA, high traffic websites and getting backlinks from them. However, your study shows that it's not only posting content but giving the signal to google that you are doing the right engagement with your target audience is what matters.
I thoroughly enjoyed reading your post, and would hope to see more of such informative posts from you.
Regards,
Vijay
Larry (the unicorn guy) you always have in-depth and impressive articles and this one can be added to that list. This is great on so many levels but the biggest being that you need to not just create content and share but make the engagement be the focal point of pushing it to the masses. The data you provided is the proof in the pudding that no matter the algorithm you are looking at more people that engage in the piece of content the more it will be shared and rankings will grow.
thanks tim. glad you brought this up. it really is the key point. we talk so much about great content so much, yet it's hard to quantify what is great content. is it content that converts well? generates tons of SEO traffic? reaches a broader audience? earns links? every once in a while it's possible to find a rare and beautiful unicorn that does all of the above and the key to being able to achieve all of them is having a ridiculously high engagement rate which will be greatly rewarded by the ML-enabled platforms that value high engagement. how many content marketers or SEOs out there are currently benchmarking their content against engagement rates? not many!
What are your thoughts on a boosted facebook post receiving high engagement? Will this likely see the same movement in search engine rankings as a non-paid for post on Facebook?
You're a unicorn, Larry. So much substance and engagement through the roof. Thanks for the insight!
thanks Virginia!!
I do some-what believe that social shares contribue to to the success of a link profile, however, not directly, but because the social share contribute to a more rounded, natural link profile. I think that's it's benefit.
Factoring in machine learning, the relationship between CTR and high social engagement is quite clear.
Thanks for the blog!
Interesting stuff Larry. However, I would argue that the main impact of social shares on organic rankings is that of any links received as a byproduct of network diffusion (originally reported via a Buzzfeed study). As Ronell Smith has previously stated on this very blog " if we simply optimize our efforts at creating and sharing content, links naturally occur." - it's these links that I believe have the biggest impact on rankings.
Hi Ben! thanks. yes, i've read the buzzfeed study! It's an interesting theory and I believe it to be directionally in the correct neighborhood -- unfortunately it's incomplete.
here's why: Moz and Buzzsumo teamed up and did this study of a million articles: https://moz.com/blog/content-shares-and-links-insights-from-analyzing-1-million-articles and incredibly found (among other things) NO overall correlation of shares and links.
So, we need to revisit the old theory to account for this data. For example, what's causing the network diffusion in the first place? I'm suggesting the missing link is high social engagement rates and high search CTRs that are greatly rewarded with increased visibility by their respective machine learning-enabled algos.
In short, high social engagement can lead to high CTR. Very helpful article, Larry.
Citation and google plus is the answer, i´m reading this post in a bus, i´ll have a closer look at home becouse i really interesting, we have more of 60K followers and i need them to improve my SEO.
I guess it's somewhat interesting that there's a higher correlation between FB engagement rate and Organic CTR. How I would consider making use of this:
Yes and focus on finding the real outliers. 3, 4, 5 or more times the expected engagement rate. The strategy doesn't work for donkeys. only unicorns!
Automated this...now to analyze data...
Simple answer is no, social signals are not a ranking factor at all at this stage.
You can also look at that from a different perspective: google has a registered patent which allows it to track and monitor brand signals in social media: https://www.seobythesea.com/2015/05/brands-entities-in-social-network/
Maybe it's not a causation nor is it combined in google's algorithm, but one could reasonably speculate that google won't bother to register that as a patent, unless it actually uses it.
As someone who has actually filed numerous patents over the last 7 years (like half a dozen of them), i'm absolutely certain that patents are not a great way to understand how a google algorithm actually works. here's a few reasons why based on my own experiences:
Maybe it's better than random guesses but a google patent is far from definintive proof imho.
Great post Larry, thanks for sharing. I'm imaging that in most cases you would convert your social content in to a blog post - assuming that the social content is not already ranking organically in Google under the social property... Could you add to that? Sounds like there's a few more things to consider like duplicate content etc... What are your thoughts?
think of social as more of a unicorn detector. if you post an idea that does great on social (eg; 30% engagement rates or something ridiculously high like that) then there's a good chance you've stumbled onto a goldmine of an idea that could be used elsewhere and would likely do great in terms of other channels (high conversion rates, great SEO, etc). Turn the idea into blog content for sure (not just one story - include many follow-up stories), video content, infographic, eBook, webinar, etc.) basically you find a truly remarkable idea, they are very rare and wonderful, so go all-in!
Ah I'm with you. Love those unicorns! Thanks!
Larry, I really enjoyed this article and will be sharing it with our team!
While I think it's undeniable that engagement positively influences rankings and helps unpack the correlation between social shares and organic performance, I tend to agree with Ben Wood's comment above--namely, that content that captures attention and engagement from such a large audience will also naturally earn links, and be rewarded with higher rankings as a result.
In your examples with WordStream.com, did you notice that either of the pieces in the three screenshots above had higher links than average, or even links that contained the title of the article (or some portion of it) within the anchor text?
Either way, this is a truly fascinating piece! Thanks for sharing!
thanks Richie. i just replied to the comment you referenced. bottom line: There's more going on than just greater social sharing resulting in more links. We know this because an analysis of a million articles from moz and buzzsumo showed no correlations between social shares and links. so, it stands to reason that there's more going on here. I'm confident the missing link are social engagement rates/organic CTRs and their big impact on visibility in the newsfeed / search rankings.
Great post, thanks. Another reason to make good content, if you haven't head enough reasons so far :)
and yet there is so much donkey content out there!
Think of it this way: you cannot simply erase everything you do not like just like the worst things in the human history cannot be simply undone. There are donkeys in the world, and you have to face this fact. The point is to try to breed unicorns in the generations to come. And this is exactly the way how to do it.
Great job, Larry!
Thanks Larry :-)
thanks andy :-)
Great article, Larry! It actually makes sense when you stop and think about it - higher engagement brings higher organic positions! I loved all those Sheldon's interruptions. :-D
Best,
PopArt Studio
I just want to confirm this statement: "One way Google checks itself is to look at organic click-through rate (but not the only way!). Did users click on the result in Position 1, or did more people click on the Position 2 or 3 result?"
It's something I strongly believe in but I've ready opposing viewpoints at other trusted sources (Stone Temple) as an example. Google has an expected CTR per position. If the results they are providing don't match up with what is expected you have to assume that they will make changes accordingly.
Thanks for the great research!
the only thing missing from the stone temple article was any data to back up the arguments. If you actually compile the numbers the conclusion is obvious
Really great article! Included it in our yearly roundup of best SEO articles. Keep up the good work Larry.
I disagree. What you're basically doing is using engagement as a measure of the quality of the content. Because you can take any other measurement in Facebook (attention/brand lift/whatever) that increases whenever the content is good and correlate that to CTR and it will always, always, come up with the same scenario. So it's a bit disingenuous to say specifically that engagement rate correlates to high CTR.
Because the truth is, it does not. The truth is, good content IN THE RIGHT PLATFORM = good CTR. Or good impressions/views/whatever.
Which is not groundbreaking. Marketers have known this and have been using this idea on TV/Radio/Print.
If you follow the suggestions in this post, you'll have pages full of memes. Because what has the highest engagement rate on Facebook? Memes.
You might want to clarify the post further, because as it is, it's pretty incomplete.
I believe that if it influences the positioning the mentions in social networks that are made of us to a greater or lesser extent, but what if it is clear is that we must work them and try to obtain positive comments which in a greater percentage or The less they will translate into the goal we are trying to achieve, a sale, a registration, a call ...
The post was worth to read,
yes, exactly the ration of social engagement matters more than social posting.
Thanks Larry for the great article.
Nice tips to think again about that content which can generate high engagement. Ultimately to get high CTR & Organic ranking in SEO Services.
Good day Larry!
First of all I want to introduce myself, I am new in SEO, not 100% new but I'm still learning facts and theories about SEO. When I read this post, some of them are familiar some are not. I want to know if social shares really affect your ranking? Do you have a summarize version of this post? (It's fine if you don't have it) Thank you very much... Your post is still very helpful.
Anton
Late to the party...
I've always found it hard to understand why people would believe Social signals would be a ranking factor. As you mention, Facebook and Google don't get along and aren't sharing their data, Google+ was a wasteland, thus there was minimal data to pull from and G only recently began indexing Tweets again, meaning they didn't have access to Twitter data.
The relationship between engagement and rankings makes more sense to me. Higher organic rankings tend to correlate with longer time on site, higher click thru rate and more interactions on the website - comments/shares.
I really enjoyed this article and the study you have done Larry. Great piece!
Nice article and I think the data approach is great. But... The content on the actual page is not what is casuing the CTR from the SERPs in all instances (your examples have answers snippets, and therefore the content is sort of on the page), but for a large % of the SERPs out there the only thing you have to impact on those CTRs from the SERPs are the title and snippet in the results. So, what you're actaully saying is,
"Create great 'content' to increase dwell time, and optimise the crap out of your page title and Meta description to increase CTR"
I believe social shares have no direct effect on organic Google rankings. Social shares increase interactions which I believe carry some weight in the Google algorithm.
Thanks Larry, I'm quite new in SEO and Social Media, but I really appreciate that you put all this information quite clear and easy to understand.
Let's see now if I'm able to do it!
Gracias
-Ricardo
Surely a crisp post!
Thank you so much for sharing.
Great Post Larry! As I noticed that google share different- different views on the relationship between both (ranking and social sharing) time by time. Google has 200 ranking factors and social sharing is a part of them, but not play a major role in ranking directly.
mainly social sharing generate traffic, if your web traffic is high and great engagement, then automatically your ranking is improving. so indirectly it's played a major role is ranking by different way.
Thanks for sharing great knowledge Larry.
The social engagement we will always give a better CTR, regardless of the niche we are, of course some will better be increased CTR, very good article Larry
Quick question.
"Since stuff that does well on organic social tends to also do great in paid social, it follows that your content that gets top organic rankings will make great content for paid and organic social.
Conversely, your content that gets tons of engagement on social media platforms (paid and organic) will likely rank highly organically for the topics that they cover."
Does this imply that promoting content on social will help boost that content's organic rankings in Google?
yes! the promotion of memorable, inspiring content (particularly video content) on social ads to relevant audiences will create a bias in people's heads towards favoring your brand. later, when these people finally do need your products or services, they'll either (a) do a branded search for your stuff, in which case you win. OR they'll (b) do an unbranded search for your products/services, but will be 2-3x more likely to click on your listing because they've heard of you in the past. People overwhelmingly click on the brands they've heard of before, resulting in higher CTR, which boosts organic search rankings.
Or... or... or...
What Google is already looking for in its existing algorithm is the same thing that causes high levels of engagement in social media. That is, great content.
The same thing that causes increases in inbound links and many other factors.
Great content drives great rankings... and great social media engagement.
Everything else is correllation.
put another way. that facebook news feed algo and google organic search algos have machine learning based components that reward remarkably high engagement content with greater visibility. case closed! :)
Interesting idea, tying Engagement Rate to CTR. You would kind of expect there to be a correlation but I'm not sure 3 instances can really be called a "trend."
hi the chart contained nearly1000 data points. The three examples were just concrete illustrations among them.
I see. I don't think I've been swayed in to believing that social plays a huge role, but aim for good engagement and you've got a good page, right? Also, I feel like CTR is only used in drastic cases, rather than "this ctr is 2% higher than it should be..."
i didn't say social shares plays a role (that was the old theory). i'm saying that social posts with remarkably high social engagement also happen to have remarkably high organic search CTRs, and that machine-learning enabled newsfeed and organic search algos dramatically reward high engagement/high CTR content with far greater visibility.
and YES. this mostly works for the outliers. the content that has 2x, 3x, 5x or 10x higher than average CTR / engagement rates (not merely 2% or 3% changes - that's just noise).
Of course, if people like something on social media there's a good chance that people that arrive on the article through the SERPS will also like it... so higher engagement rates. I'm still confused as to whether you're saying that social engagement does or does not cause higher rankings.
higher organic search CTR most certainly does increase rankings. and high social engagement is basically the same as high organic CTR.
Yeah I'd agree that CTR plays a role in the rankings, the outliers that is. I can see high social engagement as being correlated to a combination of high retention and click through rate.
I can see how this might CAUSE higher rankings within social media news feeds, but when it comes to Google, I'm not sure it even has access to engagement rates never mind using it as a ranking factor.
I can see a use for this, though. If I see a competitor has a post that has a high engagement rate, then I know it's a good article. I could create something similar, build some links and the world is my oyster.
pretty sure google just uses dwell time and CTR.
Hi Larry,
I have a question about "Did users click on the result in Position 1, or did more people click on the Position 2 or 3 result?
Even though all three of these pages may answer a user's need, click-through rate is a huge clue about whether Google is providing the best answers in the right order for users."
How do visitors, and therefore Google, know if the best answers are in the right order unless they click on all the results and compare them. Surely what influences the CTR is the snippet in the results and any previous experience the visitor will have had with the site? When looking at the latter for example when looking for business services here in the UK I will almost always ignore a particular directory (Yell) when it's in the results preferring to look at the websites for the businesses themselves, this is based on my previous experience of the directory.
yes. It's not just CTR that matters. Other user engagement metrics like dwell time / task completion come into play: https://moz.com/blog/do-website-engagement-rates-impact-organic-rankings (did the user find what they were looking for?
You are also correct that Brand affinity has a huge impact on both CTR and dwell time / task completion rates. Searchers overwhelmingly prefer the brands that they're familiar with. https://moz.com/blog/10-cro-truth-bombs-that-will-change-the-way-you-think (see #5)
Social Engagement Rates = (Total Unique People who engaged with your post) / (Total Unique People who saw your post)
If that's true doesn't it mean that Great Headlines and CTR matters more than what the content is on the inside? Does this mean Google is promoting clickbait?
Definitely CTR is part of measuring user engagement. the other part is dwell time / task completion rate. If people are clicking but bouncing away (eg: if you promise free iphones but no free iphones are to be found) then that would be considered negative engagement and would hurt you. Read this story for more info: https://moz.com/blog/do-website-engagement-rates-impact-organic-rankings
Hi Larry! Your post make me think about what Matt Cutts said in the past. To rank in Google we have to create "valuable content." I have to confess that It was too difficult for me to understand this, but you Unicorn theory is a patent of what Cutts said. I'm totally agree with you when you say that Facebook rewards a person when he/she has posts (in Facebook) that create engamement. Curiously these are the posts that (at least for me) are being show a lot of time rather than others with less engagement (likes, shares, conversiones, comments).
I think that is difficult to predict this behaviour, but my reccomendation is to make good SEO by content for the own website (Google) and good content for the fan pages. I've used tools like "similarweb" that tell us how many people find us because Social, and it is little. I know a lot of local celebrities that have a lot of engagement in Facebook, but NOT have a personal webpage where traffic can be derived... and that makes me think the following: not to put all the content in Facebook, either in Google. I think the best is to make an equilibrium, that way People can find us typing keywords and appearing in Google and making the Social a plus. What if Facebook decides to "cancel" the account of the celebrity? All the efforts are gone. I think is more difficult to be penalized by Google always we do White Hat SEO.
Nice Post Larry, I liked you inphographies! :)
Thank you very much for your input and working hypotheses . Undoubtedly the SEO and social networks are related, albeit inderecta way, because it brings more website visits and this if it is a crucial factor for SEO (especially if over your content have high rate of bounce rate).
Personally I think it also directly, because search engines also perceive the click rate on a publication of the social network. They are not stupid and increasingly evaluate everything.
Thanks for the post! It's important to remember that sharing is not everything. The key is CTR and engagement! :D
yes!! the social counts are kind of a red herring i think.
Game changer!
It's links... It's always been links... it most likely will always be links. Social signals are unicorns -- made up, fantasy signals people WANT Google to pay attention to, while at the same time, Google really only cares about the links.
that's definitely been true historically. but i think SEO requires a bit of "where things are going" kind of work, too.
Hi Larry
I understand ... In fact the position in Google is related more to the interaction and user engagement on social networks with the amount of information we provide. But ... regardless of the social network? If the interaction on Facebook or Twitter is the same, the results are translated in the same way?
yes. Posts with unusually high social engagement rates on one social network tend to also do remarkably well on others, like LinkedIn, Twitter, and Facebook. The heart of a unicorn is a fantastic idea. These fantastic ideas are social platform agnostic, and tend to get great CTR in organic search
excellent