It wasn’t too long ago that the suggestion that social media data could feed financial services underwriting and products was, well, dubious. Yes, we paid it lip service — isn’t it cool that OnDeck and Lenddo are using social media to make loans?
The lip service ends today.
Like a slow-moving glacier that finally crashes a chunk into a fjord, social media data has achieved a level of depth — and history — that makes it impossible to ignore. In part, this is what justifies the valuation Affirm got yesterday, but I will get back to that in a moment.
Last night, I attended Jon Zanoff’s fintech Meetup in New York. Besides being a sold-out, well-done event, it offered demos from five fintech companies looking to unicorn. One was particularly interesting to me: Social Alpha.
Social Alpha mines social media data for investment ideas. It looks into social media to discern tradeable events, before announcements from the public companies. You are wondering, Should I really invest my hard-earned money based on Twitter traffic? That’s that “dubiousness” coming out.
But Social Alpha has gotten well past the “Twitter traffic for stock tips” stage. Here’s how Social Alpha describes it:
Sophisticated Computational Linguistic techniques are able to rapidly assess the opinions of thousands of analysts and investors in the matter of nanoseconds, and thus gauge the social pulse of the market. This social market-sentiment tends to be a leading indicator of stock price movements.
The key is how Social Alpha is able to filter data. One of the Social Alpha founders was asked last night how the company’s algorithms can distinguish being posers and legitimate purveyors of valuable stock information? “You can fool some of the people some of the time, but you can’t fool all the people all of the time,” the founder responded. This is especially true when you have volumes of Twitter data over years.
This depth of the social media data is at least part of the reason why Affirm got its massive valuation yesterday. Affirm, a lender at the POS, gauges risk by “taking into account data from atypical sources such as social networks.” At one time this would be absurd. The loan applicant has 10 Facebook friends, so let’s lend him money. But the depth of data available via social media — multiple social media networks, remember — is profound. Why is the fact that the borrower did a credit inquiry four years ago more valuable to an underwriting scorecard than the fact that the borrower’s movements can be verified to a GPS pinpoint for the last, say, five years and that none of those locations include a federal unemployment office?
The utility of social media data, in part driven by its ever-expanding history, can no longer be dismissed. Its value can no longer be minimized, and, in turn, its potential applications have expanded exponentially.
This is the Age of Social Media Data. Get used to it.