The odds are that Facebook is the first entity you announce your relationship status to. But if you’re like me, wary of the extent of personal information you share on the web, you’re likely to be a passive user and consumer of news feed. If you’re like me, you’ll also be trying to train Facebook into recognising what you prefer to see – yes to mother-in-law’s posts, no to babies and food.
Turns out, Facebook takes that concern just as seriously. In order to improve the quality of News Feed and serve up the things you’d most like to see, (and hence, increase sales and ad revenues) Facebook has come up with an experimental algorithm that analyses your network of friends to identify your strongest relationships, and thus content from whom you’ll most like to see.
To do this, researchers Lars Backstrom and Jon Kleinberg came up with a new network measure, known as dispersion, which looks at the people who span the different social circles in your life, instead of examining the more common measure of embeddedness, which simply looks at the number of mutual friends two people may share (a weaker measure, considering that another college mate may share many ties with me as a result of being in a common social circle, without being my romantic partner).
Dispersion, on the other hand, is measuring a person’s connection to different social circles – your college mates, your family members, your colleagues, your book club friends – connections that echo your very own. This person often turns out to be a romantic partner, or just as likely, a family member. They may not rank highly in Facebook interactions – you may not stalk your cousin’s Facebook page as much as you would a new colleague – but their announcements/news would matter to you.
Looking at links through the lens of dispersion could help the company understand how “you are different than the typical user and how to adapt your experience to that group,” says Kleinberg. That could lead to more interesting — and personalized — suggestions. “Our online tools are failing us currently in that we group people and define groups by overarching things and miss other common ground,” he says. “It would be nice to enrich the set of dimensions along which people have things in common.”