Real Data: Does Social Affect Search?

I’ve heard (and read) a lot of people talk about the impact of social media or social networking websites on search engine rankings. I’ve heard a lot about this but haven’t seen a lot of REAL data. There’s a lot of “I think” or “It seems to me” and even an “I believe” but there isn’t a lot of data to back up these claims. I’ve even encountered an occasional “As you can see” but often times these claims confuse causality with correlation. Needless to say, I’m cynical about the claims that social affects search (but I would guard this by saying- “at this point”, if you would let me).

Below you can see the 31 consecutive days of Google organic search traffic for a website on which I have worked. In this data you see typical ebbs and flows based upon the days of the week (weekends are down and weekdays are up) but overall traffic is trending up. Looks pretty nice, doesn’t it:

Next is a graphic representing social interactions about this website. I need to explain this data a little more. A social interaction is either a like or a share or a +1 or a <insert social network interaction here> that relates to this website. Which websites is this data pulling from? The particular social networks that are sending their data directly to Google through their Analytics Social Data Hub program. These include:

Now this is an odd list in many ways. First of all it is missing some very important and very popular networks like Facebook and Twitter. That makes sense since these are likely competitors to Google in some ways. Still, why not include Tumblr (one of my favorites and a very quickly growing network) and others? The answer: they haven’t submitted their data to Google in order to participate in the program. Second, this list is weird because it is unclear exactly how an interaction on these networks would affect a website, let alone provide data to Google. Meetup.com is a great website but what does a “social interaction” look like on their site and how would it affect a website?

What is noteworthy about these social networking sites is that they are feeding data directly to Google and are therefore participating in their Social Data Hub. For the purposes of what I’m trying to say here that is significant because IF social affects search THEN these websites are the likely candidates from which these affects originate.

So- returning to the data- here are the social interactions from Google’s Social Data Hub partners over the same 31 days for the same website I mentioned earlier:

As with the natural search data, these social interactions ebb and flow very naturally (we were not doing anything to actively promote their social interactions during this time).

What can we conclude from this data? Let me push them together and maybe it will be clearer:

Do you see it? Does this graph show you how social affects search? Me neither. When social interactions are very high, traffic from Google organic search doesn’t seem to raise particularly high- in fact, when this site had the most social interactions the traffic from Google organic search was actually lower. At the same time, when traffic from Google organic search is highest, it’s when social interactions are lowest. There doesn’t seem to be any connection at all. Sure, there are sometimes when social interaction is low and search traffic is low- but those are likely due to the weekends when people are outside having fun, rather than searching the web.

Thus, from this data I feel comfortable to conclude: “at this point in Google’s algorithm I don’t believe there is any substantial connection between social interactions and organic search rankings.”  This might change at a later time (indeed many of the team at Ephricon believe it definitely will change), but for now it doesn’t appear to make any difference in rankings.  That said, social connections can potentially help improve your listing’s clickthrough rate if your network is large enough.