Any sports fans out there?
Maybe you’ve seen “Moneyball,” the Brad Pitt movie about Oakland A’s General Manager, Billy Beane…
With the A's struggling to win, Billy takes a radical approach to building his team:
Computer modeling. Predictive analytics. Data!
No more scouting for talent based on intuition. No more bias against pitchers who throw funny or look funny. And no more favoring overpaid “stars.”
Just crunch the numbers and predict outcomes.
In the case of Billy and the A’s, it worked. They started winning.
But can the same method be applied to venture capital?
Could a computer model identify the most promising early-stage companies? Could this be the secret to making millions?
One venture fund thinks so – and they’re backing up their bet with $165 million.
Today, we’re going to look at what this means for you.
"Moneyball" for Start-ups
“Moneyball” introduced the concept of predictive analytics to a mainstream audience, but a wide range of practitioners – from credit card issuers to public-market hedge funds – had already been using this method for years.
Until recently, however, venture capitalists kept their distance.
Sure, VCs crunched a few numbers and looked at some data – but historically, venture investing was a people-centric business. It was about who made the introduction; whether they trusted and respected the team; whether the founder was visionary.
Now, along comes a company called Correlation Ventures…
Their bold plan?
Invest $165 million into a portfolio of start-ups…
Using a computer program to decide which ones to fund.
Yikes.
More “Yes’s” More Quickly
To get started, Correlation built a massive database of venture financings going back 20 years. It captured everything from specific deal terms to founders’ backgrounds.
It took years to get all the data ingested and create a predictive model.
But now that it’s built, Correlation’s investment process is incredibly efficient:
Start-ups supply them with basic info. Correlation injects it into their model. Out pops a “Yes” or “No” decision on funding.
Many professionals spend months doing due diligence on a potential investment…
Correlation’s entire process takes two weeks, max.
So while a typical venture firm might make half a dozen investments a year….
In 2013, Correlation made a whopping 32 of them.
Date-Centric Versus People-Centric
Last week, there was a flurry of news reports about venture firms getting into data science. At first blush, it seemed like they were following Correlation’s lead…
Balderton Capital hired a dedicated data scientist. General Catalyst started searching for one. It became clear that Accel and Greylock had data professionals on staff to evangelize and teach.
But when we dug deeper, we found that many of these firms use data simply to save time, or to augment their traditional “people centric” approach. They use it to help rank their investment pipeline, or identify trends.
At the end of the day, when it’s time to write a check, they still rely on their experience, industry knowledge, and intuition to bet on great teams and visionary leaders.
Evolution – Not Revolution
Yes, venture investing is evolving.
More firms are using data to make life easier. At least one is using it to identify winners.
But if you dive into their predictive models, you’ll find the data they look at are the same inputs we write about in our guide, “The 10 Commandments of Crowdfund Investing.”
For example:
Respect Management (Commandment #2) – After they identify a company that looks like a “Yes,” Correlation gets to know management the “old-fashioned way” – they pick up the phone or meet them in person.
Be a Follower (Commandment #3) – Correlation’s model only surfaces companies that already have a strong “lead” investor.
Diversify (Commandment #10) – If Correlation were certain about batting a thousand, they wouldn’t be making dozens of investments a year. But as one of their founders said, “We’re not claiming to have a magic crystal ball. We’re tilting the odds a little in our favor with each investment.”
So don’t worry about investing fads or buzzwords like “data-centric” predictive models.
If you’re following our 10 Crowd Commandments, you’re taking the right approach.
Happy Investing!
Best Regards,
Founder
Crowdability.com
