Ever since Google began to include real-time results in the SERPs, from sites like Twitter and other blogs, speculation has been rife as to how they prioritise which tweets appear and which don’t. Whilst a search on a popular ‘hot topic’ can often reveal what seems like a fast stream of tweets, a corresponding search on a site like Twitter search reveals that Google are obviously heavily filtering the tweets and blogs they choose to include.
Well, it looks like the answer lies in a modified version of Google’s much-maligned Pagerank score, according to Google ‘Fellow’ Amit Singhal. In an interview over at Technology Review, Singhal explains it thusly:
“One user following another in social media is analogous to one page linking to another on the Web. Both are a form of recommendation. As high-quality pages link to another page on the Web, the quality of the linked-to page goes up. Likewise, in social media, as established users follow another user, the quality of the followed user goes up as well.”
So in a nutshell, it’s not about the number of followers you have, but the number of followers your followers have. Got that? It’s not your friends that count, its the friends of your friends! But this is apparently not the only weapon in Google’s arsenal, with tweets containing #hashtags also being awarded a lower relevance score (an interesting move, and not one that would seem to dictate quality very well).
The interview mentions a number of other ‘flags’ which might lower a tweet’s relevance (jn Google’s opinion) but the most interesting one to me is the plan to incorporate geo-location data, where available, to return results local to a user. Whether I, as a searcher, would find results from a person in the same city as me more useful on the subject of a major global issue is debatable, but I’m sure Google are already figuring out a new algorithm to decide that for me!