New Study: How Social Sharing Dominates SEO Ranking Factors

New Study: Social Sharing Now Dominates SEO Ranking FactorsThe latest research by search analytics provider Searchmetrics proves compelling evidence of what we’ve known for a while: search engine ranking positions are influenced as much if not more by social signals than by traditional forms of SEO.

The gap between social media, content marketing and SEO has been closing fast over the last couple of years, but it’s now at a point where anyone remotely interested in their website ranking anywhere on Google needs to take the sharing of engaging content through social networks extremely seriously. The graphic below shows what ranking factors well-positioned pages in Google’s SERPs have in common.

New Study: Social Sharing Now Dominates SEO Ranking FactorsFor me, a few things stand out.

First, Google +1s, Facebook interactions, tweets and shares on Pinterest are becoming increasingly vital in achieving good ranking positions. So much so, in fact, that aside from backlinks and basic SEO strategy, they should now pretty much lead the way in SEO strategy. Google is on a mission to make search dependent on genuine human factors and, understandably, it is placing social signals at the top of its priority list. So to rank well, websites must start to focus on quality content that people want to share through their networks.

Second, the often-stated mantra of SEO professionals of ‘keywords, keywords, keywords’ could become redundant. The use of keywords in domain names, URLs and H tags is less influential now than it was in last year’s Searchmetrics study, implying that, just as in the case of ‘backlinks, backlinks, backlinks’, this form of optimisation is relatively less important than natural language linking. Do use keywords in your content, but don’t expect them to drive rankings on their own merit.

Third, backlinks are the one ‘traditional’ SEO factor that remains significantly influential. Searchmetrics’ study fails to spell out the quality of backlinks in addition to the quantity, but it’s very, very safe to assume that when looking at the number of backlinks, quality must be taken into account. The days of healthy backlink profiles including 1000 link farms in Armenia are well and truly over.

In SEO terms, quality content shared through social networks rules.

The Searchmetrics study looked at 300,000 websites, 10,000 keywords and billions of backlinks and social shares. Correlations were calculated using Spearman’s rank correlation coefficient. You can download the full study here.

Like this post? Subscribe to FutureComms and get it straight to your inbox.

Posted by Paul Sutton

  • charliesaidthat

    Careful! This is a correlation study.

    It does not prove influence. In fact the reason it is shared might be because people found it at the top of Goog and shared it. 😉

    It is an interesting study – but also has been interesting in SEO circles as some of the stats like “Keyword in Page Titles” has a correlation of 0, only because they ALL have that, not that it isn’t important.

    Very very easy to take this data way out of context.

    • Paul Sutton

      Not sure I get your drift, Charlie. The study assesses the correlation between which sites (within the study) rank highly in Google and which factors they have in their favour. It’s therefore entirely reasonable to say that the most common factors are ‘more likely’ to have an affect on SEO. More to the point though, it compares the relative importance of different factors TO EACH OTHER. That doesn’t mean, like you say, that keywords are not important, but it does mean that social signals are relatively more important.

  • Sean McGinnis

    What Charlie said.

  • Paul Sutton

    Sorry guys, but I don’t buy it. That’s an interesting follow up post you reference Charlie, which addresses some interesting questions. But whichever way you cut the data, the fact remains that social signals are highly correlated with web pages that rank well, according to this study.

    That is not to say that these factors cause high rankings, you’re right. And yes, you have to make some assumptions to interpret this data, hence the differing opinions. As Searchmetrics say themselves, the only people that know that work at Google. BUT, if we assume that all/most optimised web pages have factors such as decent on-site optimisation of keywords, which would seem to be the case judging by the low correlation coefficients of factors such as keywords in H1 tags (for example), then it could well be the social signals that indicate the difference between the good and the best. And this totally fits with the way we know Google thinks and what it’s trying to achieve. So I’m not sure how this can simply be disregarded.

    It COULD be coincidence that best ranking sites in this study all had lots of social signals. But to believe that because it fits the way current SEO thinks is to bury one’s head in the sand, surely…

  • robertwent

    Charlie seems to have hit the nail on the head.

    How do you know that the results of all the sites positioned at 100-130 wouldn’t show exactly the same results when compared to each other?

    The follow up link makes no mention of measuring factors such as basic grammar, spelling and general readability and overall quality of content. Instead it chooses to dwell on the services sold by seo companies and present the data in a way that can very easily be confused with overall importance with a direct impact on ranking.

  • Sean McGinnis

    Paul – I take issue with the language you use to draw your conclusions. We agree that this is a correlation study and yet you state:

    “search engine ranking positions are influenced as much if not more by social signals than by traditional forms of SEO.”


    “Google +1s, Facebook interactions, tweets and shares on Pinterest are becoming increasingly vital in achieving good ranking positions.”

    Neither of these statements are supported by the study. Taken on their own, these statements are simply not true. That is my concern, that people who do not really understand SEO or the way that search works will focus all their efforts on these two things and thing that they are positioned for success.

    This study adds very little (if anything) to what has been known about ranking factors for some time. Moz did two studies within the last year and a half or so which also suggested a strong correlation between social shares and rank, but correlation does not equal causation.

    Even though both you state that correlation does not equal causation, the two statements I pulled out and highlighted above state that “search engine ranking is influenced” by social signals and that social signals are “vital in achieving good ranking positions.” Neither of those two statements is true.

    What is safe to say is that “of all the factors studied social signals had the highest correlation with search rank.”

  • Martin McGarry

    Belated response, but how does Google define that your Facebook page is directly associated with your website, as there are no strict authorship codes to implement like with Google+. I could effectively choose to associate my website with any popular Facebook page by pulling there likes boxes to my site and posting content on their page/timeline. I am may have missed the poiint entirely, but I’m struggling to understand how a rival such as Microsoft allows one of their Major brands to be a defining factor in Google’s algo

  • Pingback: SEO, PR or Advertising Agency: Which is Best for Social Media? | MediaStreet News & Opinions()

  • Pingback: Social shares for backlinks quality check (new metrics)()