Is quant analysis just for baseball, or could venture capitalists also benefit?
After watching a trailer last week for Moneyball, blogger Dan Frommer wondered if the film’s lessons could be applied to early-stage tech investing:
"Tech is, after all, an inherently data-driven industry, but much of the talk around investment seems to be about relationships and personality. With the recent boom in startups and smaller investment vehicles, you’d think that data- and value-driven startup investing — looking for specific indicators and trying to exploit them — could be an opportunity."
He then asked a few VCs to opine, and their (paraphrased) responses ranged from “interesting idea, but not for us” to “you so crazy.” Two thoughts:
1. If you were going to pitch a fund like this, I’m not so sure Moneyball should be in the slide-deck. After all, the story’s (not so) dirty little secret is that it hasn’t actually worked for Oakland. The team Michael Lewis chronicled was successful mostly because the team had great starting pitching, not because it had a bunch of everyday players with high OPS (that’s also true of the 2004 Red Sox, by the way). Moreover Oakland is in third place, on its way to a fifth-straight year without heading a playoff appearance. Pretty sure the sequel will be titled: “I’m number 3.”
2. More to the point: There are at least two firms that have tried some version of this. The first is Correlation Ventures, which so far has raised at least $45 million for its debut fund (according to a regulatory filing). Not exactly what Frommer envisioned, since Correlation is a co-investor rather than lead, but it’s still analytics-driven venture. Still too early to determine is Correlation will be a division-winner or also-ran.
Then there also is Google Ventures, which does traditional due diligence but also runs prospective investments through some sort of algorithm to help it make decisions. The team has always been vague about the specifics of how this works, but perhaps it would be more surprising if Google's venture arm didn't use some sort of proprietary data analysis...