Institute For The Future

Data Exhaust and the Future of Peer Pressure

About a month ago, I received several negative comments through Twitter about a blog post on the idea that people are beginning to threaten themselves with embarrassment online as a strategy to improve health by, for example, installing a sensor enabled refrigerator that lets its owner's network know when he goes for a midnight snack. At the time, I wrote about the concept as an example of "smart pain," which is to say short-term, relatively harmless pain designed to help us achieve longer term goals. But it's also an example of a separate phenomenon: The bits of data we're generating without thinking about it are becoming an incredibly valuable source of peer pressure that has a ton of potential to change our behaviors, mostly by making us very, very uncomfortable.

Consider a fascinating startup called Lenddo, whose mission is to mine through social networking data to determine credit risk—and claims, surprisingly credibly, to be doing so for social good. Their argument is that for people who don't have credit history, their social and web history data--are they connected to reliable people, for example--can offer a useful proxy to determine creditworthiness, and that this proxy score will allow for a dramatic expansion in microfinance.

On their blog, Lenddo's CEO argues, somewhat counterintuitively, that this kind of social media mining will open up a return to more holistic ways of understanding how risky people might be.

Bankers lost their way in the 1950s as the very nature of community changed.  As the country started suburbanizing, the automobile and television all converged to change how industrialized societies interact and people became less connected with their neighbors.  One of the results is that banks turned to faceless financial ratios and computerized credit scores as the core of their underwriting process.  Furthermore, lending based on personal relationships did not scale up well in the age of mega-corporations…

The rediscovery of character based lending all started with Professor [Muhammad] Yunus‘ desire to help the extreme poor in Bangladesh.  Yunus hypothesized that if members of the community were willing to vouch for each other it would be an effective way to distribute credit.  He was right.  It turns out that people’s willingness to vouch for each other is driven by character, and the result was the expansion of the microfinance industry.  Yunus’ customers did not have access to electricity, let alone a credit score, but thanks to Yunus they could access capital to improve their lives.  Today the microfinance industry successfully lends over $50 billion a year, and Professor Muhammad Yunus won a Nobel Prize for his positive impact.

Globally, the default rate in the microfinance industry is around 2%, outperforming most western consumer lenders.  Clearly, lending based on character works.  Yet $50 billion a year is not much in a multi-trillion dollar debt market.  That’s because until recently, microfinance lending was only used by the ultra poor, and had to be face-to-face.  With social networks like Facebook, Google+ and Linkedin, these media channels can change all that.

Much like microfinance, one of the keys to making Lenddo work is peer pressure. As when someone gets a loan through Lenddo, and then fails to pay, the system "will inform a person's network about a late payment, and friends' scores will also drop." In other words, if you fail to repay your loan, you aren't letting down a giant bank; you're directly harming your friends.

Talk about peer pressure.

Now, this idea—of mining social data for creditworthiness—is not particularly new, but it's still very controversial. The Economist recently highlighted an expert in using social media to underwrite insurance risk—who has actively started paying for junk food in cash, for fear of tipping off algorithms to his unhealthy habits.

When I first read about Lenddo, my gut reaction was similar—my instinct is that I'd rather not have my Facebook page factor into what kind of interest rate I get on a car loan, for example.

That said, assuming these algorithms work as intended—and to be fair, that's a big, big if—it's hard to understand the concern. The idea behind using peer pressure to ensure loans get repaid is thousands of years old; the idea of using your friends to help you stick with a lifestyle goal is at least as old as alcoholic's anonymous. Technologies are simply facilitating new kinds of peer pressure, such as the ability to automatically alert your network when you do something you shouldn't, like have a midnight snack or fail to repay a loan.

All of that said, I still don't want to have my Facebook page analyzed before I get a car loan. Which is a simple way of saying that even though I can defend the concept of using my Facebook page to determine credit-worthiness, I still don't like it. Or, in other words, we're facing an uncomfortable future: One where we have an increasingly useful tool—in the form of social networks exerting peer pressure—whose potential lies in its ability to make us feel uneasy.