Signal Over Noise: The Ultimate Algorithm for Twitter

Top Tweets from My Twitter Account, via my6sense. No Noise.

Twitter is one of the biggest success stories in social networking and the tech industry in general over the last five years. Since its 2006 debut, the microblogging powerhouse has converted from a small audience of early adopters to playing a major role as the pulse of the planet, the go to destination for breaking news as it happens, or a shared social experience around entertainment events, calamities and public interest. But even as millions and millions more have signed up for accounts, even its most supportive users see a gulf between the site and its potential - exacerbated by a continued problem with signal to noise. On Friday, well-respected blogger John Battelle explained his position, saying the site has incredible potential, but that it becomes impossible to get value from the stream. (See: Twitter and the Ultimate Algorithm: Signal Over Noise (With Major Business Model Implications)) He writes:

Many observers have commented on how noisy Twitter is: That once you follow more than about fifty or so folks, your feed becomes unmanageable. If you follow hundreds, like I do, it's simply impossible to extract value from your stream in any structured or consistent fashion.

As you know, even if you are extremely selective about the people and brands you engage with on Twitter, they are not always on topic or engaged in discussions that match your interests. Battelle asks for Twitter to "present that stream in a way that adds value to my life," recognizing the challenge that "It requires some Really F*ing Hard Math, crossed with some Really Really Hard Semantic Search, mixed with more Super Ridiculous Difficult Math."

Battelle suggests one approach is for Twitter to assign a rank or context to every tweet and how it relates to other tweets and accounts, which then links to advertising. If the noise is solved, then the company's path to revenue becomes clear.

At my6sense, we love Twitter. But we love making Twitter even better and smarter than it is today, and finding the signal in the noise, through learning from your own actions, what are your interests. We know that this process is incredibly hard, which is why our team of artificial intelligence scientists and engineers have worked diligently for years to find signal from the noise of social networks, including Twitter, Facebook, Google Reader, and more.

The Best News and Content Floats to the Top of Twitter

Our goal is not to find what is the most interesting content from your friends, or from the world, but just for you. In the last few years, some companies have approached the noise problem on Twitter with smart curation of content or what they refer to as filters. my6sense is not focused on curation and filters, but instead the ranking of the content that flows through your real-time stream, and surfaces the best for you automatically.

How do we score each tweet, as Battelle suggests? Based on a combination of keywords in the tweet itself, the story linked from the tweet, the author of the post, the source of the post, the sender of the tweet, the time of day, the duration between the tweet and your receiving it and many other factors. Some might call this proprietary. We call it really hard work, just like Battelle suggests.

Want to see the full potential of a Twitter without noise, where you get value from every top tweet? Add your Twitter to our Android or iPhone app, or add our dedicated Chrome extension to your browser, which brings the power of my6sense to Twitter.com. The more you use Twitter, the smarter we get and the better value you see. We know this challenge is hard, and we know every top social network is going to fight this challenge head-on to solve the problem themselves. But we've done it, and we invite you to give our approach a try.