This week Amit Singhal, who led development of real-time search at Google, gave us a little insight into how Google is ranking tweets as a part of their real-time search efforts.
Tweets are a main component of Google’s real-time search results, but up until now there was very little information available about the algorithm used to rank these microblog posts. One factor of their ranking doesn’t really surprise those in the SEO industry, as it is similar to Google’s PageRank technology used for websites in traditional search. Basically, Google judges the value of a website by looking at it’s link structure – how many links on the web are pointing to the site, and the number of pages linking to those linkers. The same is true for their ranking of tweets, just replace the word ‘link’ in the previous sentence with the word ‘follower’. Of course, quality comes into play as well – the better the reputation of any given follower, the better your reputation is as well.
“You earn reputation, and then you give reputation. If lots of people follow you, and then you follow someone–then even though this [new person] does not have lots of followers,” his tweet is deemed valuable because his followers are themselves followed widely, Singhal says.
He’s quick to point out it is not a popularity contest, but it sure seems that way. Granted, judging number of followers may seem a good indicator of the level of clout one person has, but this model doesn’t take into account ‘tweet’ of originality. Say a small time reporter breaks a big story and tweets about it from their personal account, and then CNN tweets it from theirs – CNN will get the credit, while the reporter responsible for the news gets lost in the shuffle. Also, social media is a different ballgame than traditional websites, and using this rank method may not transfer well. If I place a link on my website to another page it is a very clear sign that I do, in fact, endorse that page – which may not be the case for all the people I follow. For example, I may follow competitors to keep track of my industry when I would never link to their site directly. Hopefully as real-time search develops Google will come up with a better way to deal with the Twitter account reputation factor.
Google has also developed a way to determine which tweets should show up for more common terms, such as celebrity names or current event keywords. The idea being to weed out spam and low quality tweeting from appearing in the results. A common practice on twitter is to include hashtags – basically the # sign following by the topic keyword. It’s a great way of keeping tweets about the same topic connected, and to get your tweet seen by those following the conversation. Unfortunately the practice that began as a useful organizational idea has been abused by spammers, and Google has determined it could be a red-flag marker of lower quality tweets.
Also, Google is scanning to find the “signal in the noise” to help sort through the tweets from different sources and viewpoints to determine the most up to the minute, relevant tweets for these popular terms. They are essentially looking for a sudden influx of similar tweets, for example tweets that mention ‘hospital’ along with ‘Lady GaGa’, and consider that evidence that the particular event mentioned is most relevant.
While real-time search encompasses many things and Google has intentions to continue refining their real-time results, Twitter updates are currently a major component. It’s nice to have the ranking factors those in SEM assumed were affecting tweet placement confirmed by the source.