Rankings and Social Signals: The Wikipedia Effect

As the web continues to become ever more conversational in nature, we’re constantly told that Google are increasingly incorporating social cues into their ranking algorithms. Nowhere is this trend more obvious than with the integration of Google+ data into the SERPs which, in an increasing number of instances, directly effects the personalisation of the SERPs. However, the importance of other social media signals – such as Facebook Likes and Twitter mentions – certainly isn’t proven and, in my opinion, doesn’t really have any effect at all on where a site appears.

This is what I like to call the Wikipedia effect.

The Internet’s de facto encyclopaedia of everything doesn’t even have social buttons embedded on the page. Yes, there are almost certainly other reasons for the lack of a Like button (privacy, anyone) but surely if they were that important to rankings, and therefore traffic, these would have been included?

But this page has 27,000 Facebook Likes and ranks first, I hear you say. Well…

Correlation != Causation

via http://xkcd.com/552/

Even though there’s a high correlation between socially popular sites and high SERP position, it doesn’t necessarily follow that all those social indicators are the reason for their success. In many cases, the momentum created by social sharing leads to a website inevitably building up a positive and in-depth link profile, which leads to high domain authority and PageRank, in turn meaning great SERP position.

Look at it this way:

via http://pubs.acs.org/doi/abs/10.1021/ci700332k

Do you really think that the amount of Lemons imported from Mexico made a blind bit of difference to the road fatality rate? It’s very unlikely.

Social Doesn’t Signify Superlative, at Least Not Yet

If you want your Twilight fan fiction website to pick up visitors from the endless collection of Social Networks it pretty much goes without saying that it’s important to have a lot of engaged Facebook fans and Twitter followers. Nevertheless, don’t expect this to help you rank on Google if your content is crap and you’ve just copied and pasted everything from teamedward.com.

One of the drawbacks of using social cues for ranking purposes is that they can be manipulated easily using sites like Fiverr. Despite Wikipedia’s lack of tools for tapping into the supposed power of social cues such as widgets, it still manages to dominate top spots for pretty much any search query you could ever think of. How? Content and links.

Wikipedia doesn’t want social signals to be used heavily in ranking schemes because they don’t matter for them and would be counterproductive. More traditional cues beyond social popularity such as the quality of the written word, along domain authority and PageRank have a far greater and more consistent impact than any social network could ever dream of having. Think about it: if social metrics ran search all you’d ever get would be videos of cats that like water.

The Future of Social Media SEO Integration

Now, SEO will increasingly depend on social signals… but not in the linear way that some have been expecting. Yes, social data from Google+ is going to be used more and more (after all, Google need the big + to succeed), but that’s about as far as I think it’s going to go. Social is all about generating traffic, readership and community. It lets you converse with your customers, generate new ones, gives you the chance to dream about going viral, and – if something works – the chance to show off a bit.

A final mention has to go to Facebook’s new Graph Search which is pretty much the epitome of why using social metrics to run a search engine just doesn’t work. Quite scary invasions of privacy are just the start. In the traditional sense it adds very little real world value… although might help with getting a date and is mildly entertaining (just like funny cat videos).

Matt Beswick

Co-founder of Aira and Blush. Marketing / tech geek. Husband. Dad of two. Enjoyer of bourbon, cookery, computer games, and things that go fast.