In the world of venture capitalists, not everyone is Peter Thiel. The Silicon Valley investor reaped 1 billion dollars in 2012 when he cashed in his Facebook stocks, turning a 2,000 percent profit from his initial $500,000 investment. Stories like Thiel’s may be inspirational, but they are by far the outlier. The start-up world sees thousands of hopeful companies pass through each year. Only a fraction of those ever return a profit.
Picking a winner, the elusive “unicorn,” is as much a matter of luck as it is hard numbers. Factors like founder experience, workplace dynamics, skill levels and product quality all matter, of course, but there countless other variables that can spell heartbreak for an aspirational young company. Successful venture capital firms claim to know the secret to success in Silicon Valley, but it can still be a harrowing game to play.
Humans just aren’t very good at objectively sorting through thousands of seemingly unrelated factors to pick out the subtle trends that mark successful companies. This kind of work, however, is where machine learning programs excel. Two researchers from MIT have developed a custom algorithm aimed at doing exactly that and trained it on a database of 83,000 start-up companies. This allowed them to sift out the factors that were best correlated with success — in this case, a company being acquired or reaching an IPO, both situations that pay off handsomely for investors.
In a paper published to the pre-print server arXiv, they say that their algorithm picked successful companies 60 percent of the time — double the rate of most venture capitalist firms. It did so by incorporating data on the founders themselves, the executives and advisors, such as education levels, and whether they had been involved with a successful company before, as well as information on how various companies progressed through the multiple funding rounds that sustain start-ups. They based their algorithm on a series of equations normally used to describe the chaotic movements of particles in a fluid, known as Brownian motion, and essentially attempted to isolate which variables mattered the most.
They found that one of the biggest predictors of success was how start-ups moved through rounds of funding. And it wasn’t the slow and steady companies that were hitting it big, it was the ones that moved most erratically, pausing at one level of funding and then rocketing through the next few. How this plays into start-up success isn’t completely understood at the moment though.
They also found correlations with more traditional factors as well, such as the level of experience among founders and executives. These are things that most venture capital firms already take into account when choosing start-ups to back, but it seems that the algorithm could optimize for these things much better than humans. It’s another area of finance where artificial intelligence has made significant gains in recent years. Algorithms are already trading stocks and performing market research for corporations. It may soon be deciding which companies get off the ground at all.
The researchers say that their algorithm could be applied to much more than just nascent tech companies. The same principles that allow it to pick a handful of winners from a crowd of duds should also apply in areas as diverse as the pharmaceutical industry and the movie business, where just a few successes can pay out billions. These are fields where the top players are lionized for their ability to sniff out winners and reap the substantial rewards. As with factory workers, bank tellers and telemarketers, the robots could be coming for their jobs as well.