Why Search Engines Should Use Social Signals as Ranking Factors

It’s extremely difficult to test isolated hypotheses in search, given that so many factors contribute to ranking algorithms. Within any single factor there may be second order effects we can’t separate.

You know… is it the tweet itself or the link within the tweet (or both) that seem to have an affect on ranking a content item?

In addition it can often be extremely difficult to get a clear and consistent answer from the search engines (Google in particular).

If you’re new to social-signals-as-ranking-factors, then I would suggest you start by reading this post by Jonathan Allen, which collates the story right up to the end of 2010, including:

• Matt Cutts confirming the assertion that Google does use social signals, primarily referencing author authority, following reports that Twitter and Facebook signals factor into Google’s and Bing’s search algorithms.
• A patent filed by Google also suggests an interest and research in this area.

Bing states that Twitter author authority is a factor (in some way) and Facebook links shared with everyone are considered (though Facebook author authority isn’t).

Google states that retweets outside of the link are a signal, plus author authority; with Facebook links (on fan pages) treated in the same way. In addition, they state that Facebook author authority is treated in the same way as Twitter: “Yes we do compute and use author quality. We don’t know who anyone is in real life.”

Some pretty clear and exciting statements in there, since followed up by a less convincing re-confirmation by Cutts, followed by a confusing statement later that seemed to completely contradict that links in retweets are part of ranking. Check out this interesting SEOmoz test and discussion in the comments.

Confused by Google? I am.

However, common sense tells me that social signals — from author authority to number of re-tweets to number of “likes” — should (if not now, then very soon) be some component contribution to the Google algorithm.

1. Lessons From History

Regardless of how long we’ve been working in the SEO industry, most of us know that there were search engines before Google, and such search engines used on-page factors and keywords in content to determine ranking.

It wasn’t until Sergey Brin and Larry Page put forward “The Anatomy of a Large-Scale Hypertextual Web Search Engine” paper, that links were actually a lead component of search rankings. It seems almost daft to us now that there could ever have been an attempt to consider a pages’ relative authority without any link component — links being the very stuff of the web.

Given the functionality of social media sites and our ability to use buttons to make essentially qualitative assessments (I like this piece of content, I’m re-tweeting this piece of content), would it just make sense in terms of the evolution of the web to incorporate such data where it’s available (in some part)?

2. They Know Who You Are

“Yes we do compute and use author authority.” In this case in answer to a question about Twitter, but we already know from social search that Google “knows” who we are interdependent of third party data.

If I set up my Google profile and add “my sites” with nice XFN rel attributes, I’m telling them straight up that /NicholaStott on Facebook is /NicholaStott on Twitter, and Quora and LinkedIn and wherever else for that matter. Given that social circle results now feature interspersed throughout my SERP, I’d warrant that user feedback (i.e., positive CTR data) has informed that change.

Google knows who we are, knows who are friends are, and knows that we like and trust the stuff that our friends are sharing. Why wouldn’t our known interactions with content be taken into account algorithmically outside of social circle results?

Google is a business. Whatever increases value and relevancy of the core product is good for the business. I’d find it hard to believe that social circle data is completely off limits.

3. Product Development

Right now Google +1 is limited to my direct Google connections only (as opposed to my friends network a la social circle). Maybe there are technical or legal reasons that I can’t be shown what search results a secondary contact of mine has +1’d (or maybe this is in the pipeline as it is early days after all).

Once +1 rolls out on web pages, I’d argue that there will be greater awareness and desire among web users to set up a Google profile and also connect with others via Google.

Google already takes user feedback into account when it comes to algorithm updates. Panda explicitly incorporated user feedback elements and this was explicitly communicated to us from Google. Surely +1 feedback as contributory ranking factor is on the cards?

Enough hypothesizing and un-founded supposition from me — I’ll leave you to your own conclusions. However, if something smells right to me I subscribe to the Karl Popper school of scientific research. I’m cool with these theories until someone proves me wrong.