Switch Ventures, which uses predictive analytics to find high-performing entrepreneurs to invest in, has bolstered its team with a machine-learning engineer.
Joining the San Francisco firm as data scientist is Jason Victor. Victor most recently headed data-science, machine-learning and AI initiatives at expansion-stage firm Lumia Capital.
Prior to joining Lumia in 2017, Victor was a quantitative analyst for hedge funds. In 2011, he founded Cloudmetrx, which provided risk analytics to the trading industry.
His hire, effective in late March, comes as more venture firms are exploring AI-enabled platforms to help make investment decisions. Firms also are combining the data-driven approach with old-fashioned gut instincts on making deals.
Healthcare-focused Foresite Capital recently raised $668 million for its investment approach, which uses data to calibrate decision making. Other data-driven firms include Right Side Capital, SignalFire and Social Capital.
The Switch data team scrapes the Web in a large-scale data-collection effort to build tens of millions of data points on more than 150,000 founders. Using predictive analytics, Switch then tests thousands of hypotheses and builds models to predict which founders will outperform.
The models of predicting investment success are very powerful, says Paul Arnold, founder and sole general partner of Switch. “This is what the future will look like for a lot of venture firms.”
Before he started investing in companies, Arnold was an early employee and head of operations and accounts at the ecommerce platform AppDirect. Prior to that he was a consultant at McKinsey & Co.
“I had a theory that you can predict entrepreneur success, and that was how Switch Ventures came about,” he said.
He launched Switch with a proof-of-concept fund and is currently raising Switch Ventures II. The fund target is undisclosed and Arnold declined to discuss the details since he is in fundraising mode. He is currently investing from the tail end of Fund I and a separate opportunity fund.
Thus far, the hypothesis is proving out. Last year, Cambridge Associates ranked Switch as the second-highest performing U.S. venture fund for its vintage year across three core measures of IRR, TVPI and DPI.
Arnold did say he would take referrals and meet with companies that are not identified by his data-crunching model, but about 70 percent of his portfolio is based on the data analytics.
To date, the firm has invested in 34 deals in just over three years. The data-driven approach has brought Switch to a range of tech sectors, from consumer to B2B, including insurance, media, enterprise software and consumer SaaS.
Portfolio companies include digital-health company Aaptiv, subscription-based sports-media publisher The Athletic, insurtech company PolicyGenius and public-records-request-software provider NextRequest.
Exits include marketing-analytics company BrightFunnel, which was acquired by Terminus, and meetings-productivity startup Do.com, which Amazon Web Services purchased.
Switch was also a backer of Mattermark, which operated a data platform for the venture industry before it was bought by FullContact.
Photo of Jason Victor courtesy of Switch Ventures.