Analyzing Mobile Commerce Data to Identify Urban Tribes

Last December, I attended a series of workshops on Mobile & Social Commerce at MIT Sloan School of Management. this is a summary of the workshop on identifying "Urban Tribes" by analyzing Mobile Commerce data, presented by Professor Alex Paul Pentland, Director of MIT’s Media Lab Entrepreneurship Program. 

 Prof. Pentland described an experiment his team did with people's mobile phones. “Today,” Prof. Pentland said, “five billion people are running around, carrying with them these little sensor packages.” You own one; I own one. Some of us own more than one.

And every time we activate these little sensor packages (i.e., use our phones) we spew out data: When we dial a number, it is recorded who we are calling. When someone answers our call, it is recorded how long we talked to them. As our phones connect to the nearest cell sites, a simple process of triangulation identifies exactly where we are. In fact, with GPS chips, our phones can reveal our location even when we are not on a call.

With other people’s phones doing exactly the same thing, it is easy for any person with access to the data that these devices spew out to know who we are, who we talk to, who’s standing near us, even who will be standing near us at a certain point in each day, where we go on vacation, what places we pass by every day, what places we actually stop in.

Professor Pentland's team combined data from mobile devices about who the people participating in the study called and who were in their physical proximity; using the recordings from the bluetooth transmitters/ receivers in their phones; and then used a statistical technique called factor analysis to build people's social graphs. In a paper published last May, ­Pentland and his group showed that cell-phone data enabled them to accurately model the social graphs of about 100 MIT students and professors. More importantly, they could also precisely predict where subjects would meet with members of their networks on any given day of the week.

Just using the anonymous location data, you could also identify a group of people who don't know each other, but show an identical pattern of behavior, i.e. work and live in the same neighborhoods, eat in the same set of restaurants, shop in the same set of shops etc. 

Now if you still can’t imagine how this data can be used to move businesses up, Prof. Pentland gave a pointed and very simple illustration: “Using information gathered from people’s phones, you can tell that this group of people (1) goes to a lot of coffee shops, and (2) they all pass by this certain area where there are no coffee shops.” What you have you gained in this instance? For one, you have just found a place that would be good for putting up a coffee shop!

Of course, that’s just the tip of the iceberg. Prof. Pentland added that a person who knows how to use this data can, in fact, also predict your product’s conversion rate, tell you where 95% of your buyers are, what and how your promotions should go, etc., etc.

This is the kind of resource businesses have out there right at this very moment, totally ripe for the picking. Now the question is, are you using it? If not, imagine the things – the profit and the cost savings – you are missing out on.

 Yash Talreja, Principal, The Technology Gurus