Can data predict a future founder?
Can data predict a future founder?
That's the question that Bloomberg Beta and Mattermark set out to answer in June, driven by the desire of Bloomberg's new venture capital arm to develop relationships with these future belles of the ball.
"Part of it was driven by my own personal experience," says Bloomberg Beta head Roy Bahat. "No one wants to talk to you until suddenly everyone wants to talk to you. It's nice to get to know people before we want anything from them or they want anything from us.
Leave it to Mattermark, a company that analyzes data like LinkedIn profiles and tweets to deduce startup performing, to take on the seemingly impossible the task of identifying these future founders. It started by modeling out a couple hundred well-known venture-backed founders, the "whales" and the "unicorns" according to Mattermark co-founder Danielle Morrill. Morrill created a database listing attributes of these people like their age, college education, what job titles they've held, and where they worked before, and if they have experience working on a venture-backed startup.
The company analyzed the trends in this data set, and compared the findings against a pool 1.5 million people currently working in tech. Mattermark initially came up with a list of 70,000 potential future founders, and then further narrowed it down to the 350 most likely to take the plunge and start their own companies.
Morrill claims that these people are 25-times more likely than a random individual working in tech to one day found their own companies.
Here's where it gets weird: Today, Bloomberg Beta emailed these 350 people -- none of whom have ever founded a company before -- to say, basically, "Hey we think there's a good likelihood you might found a company in the future. So, we want to meet you and invite you to a selective year long series of networking events."
Ah, to be a fly on one of those 350 people's computer screens.
Bloomberg Beta wouldn't release the names of the chosen ones, given that these people had no idea they were being compiled into a master list.
The very existence of this weird data experiment is fascinating in-and-of-itself, but the insights into the makeup of the likely future founders are also intriguing. Mattermark's blog post covering the study offers the following traits:
- 38 percent are over 40 years old
- Only 15 percent have a Computer Science degree
- Management consultants are more than 2x more likely to be venture-backed founders than engineers
- 43 percent of these potential future founders worked at a venture-backed company immediately before founding
Morrill believes the surprising results cut through stereotypes and media hype about what types of people are more likely to become founders in Silicon Valley. "In the industry we love the theme of the unlikely founder and we tend to lavish attention on someone who is young and is capable of building a world class company," Morrill says. "That doesn't mean there aren't more experienced candidates, but there's less of a surprising tale to tell [about them] so they get less attention."
That said, Morrill also points out that conventional wisdom has its place too. The most predictive collection of traits for a future founder were, "Stanford graduates with Computer Sciences degrees who are currently working (but not founders of) a venture-backed startup."
Of course, this experiment is far from rock solid, a fact that Bloomberg Beta's Bahat readily admits. "It assumes the founders of the past will look like the founders of the future," he says.
There's so many factors that go into someone's crazy, masochistic decision to start a company -- and to put up with all the hell that comes with that. There's simply too many unknown variables for a data study based on a handful of attributes to reliably predict the future.
"The majority of the people on this list wont become founders because it's hard to become a founder," Bahat says.
So why would Bloomberg spend the time and money compiling this list and staging networking events for these people if it's all a crapshoot anyways?
"Even if these people don't start companies it's an awesome list," Bahat assures Pando. "We'll be getting to know people who might go on to become executives at startups. I'm convinced investing in people early who show promise is a good use of time."
[illustration by Brad Jonas for Pando]