Can algorithms predict the next billion-dollar companies better than human venture capitalists? It seems possible, at least based on a formula created by CB Insights and The New York Times.
Back in 2015, the companies published a list of 50 startups that would eventually becomes “unicorns,” or those valued at $1 billion or more. It identified candidates using CB Insights’ Mosaic algorithm, which analyzes the health of a startup based on various data including strength of market, financial performance, and overall traction — a “FICO score for startups,” as described by the investment data firm.
Fast forward to today, and 48 percent of the companies on the 2015 list are now considered unicorns. “At the risk of sounding immodest, that is pretty good,” CB Insights wrote this month. “And if we were a venture firm, this kind of hit rate would make us legendary.”
That’s why it’s worth giving the 2019 list a look.
The 50 future unicorns hail from various industries and the median company has about $111 million in total funding. A majority are based in the U.S., with 22 from California, five in New York, and two in Massachusetts.
There is just one from Seattle: sales automation startup Outreach, which raised $65 million this past spring, announced it was moving into a new headquarters space this summer, made its first acquisition, and was the only Seattle company to crack the top 25 in LinkedIn’s Top Startups list for 2018.
Outreach CEO Manny Medina said the company more than doubled its revenue in 2018 and met all goals and metrics. Outreach now has more than 3,100 customer accounts and 50,000-plus users. It employs 315 people and plans to reach 450 by the end of 2019.
“This upcoming year we will make more investments in scaling the business efficiently and prepare for an IPO a few years out,” Medina told GeekWire. “This includes continued investment from our product to support, measure, and automate customer facing workflows. Our job is to make all sales reps great and drive higher revenue efficiency for their companies.”
The 5-year-old sales engagement platform uses machine learning to help customers such as Cloudera, Adobe, Microsoft, Docusign, and others automate and streamline communication with sales prospects.
Medina, a former director at Microsoft, originally launched a recruiting software startup called GroupTalent in 2011 with his co-founders Andrew Kinzer, Gordon Hempton, and Wes Hather. But the entrepreneurs pivoted in 2014 to focus on building tools for salespeople.
Chris DeVore, managing director at Techstars Seattle — Outreach was a 2011 graduate of the accelerator — said the company is a good example of why he focuses on investing in people over ideas.
“Outreach is one of my favorite stories,” DeVore told GeekWire last week. “The business they set out to build wasn’t working, but because they stuck together as a founding team and kept adapting and learning, they figured out how to find a productive thing. But that wasn’t because of where they started or the early metrics. It was because as humans, they were so committed and resilient and so gritty that they figured it out.
“And that’s really what you’re betting on,” DeVore continued. “It’s a 10-year journey and it’s never always up and to the right. There are always setbacks and near-death moments. It’s the human capacity for resilience and persistence every time that will turn a bad investment into a good one.”
While it’s a safe bet to invest in tenacious and dogged founders, CB Insights’ track record with its Mosaic score shows how data-driven formulas can drive smart investment decisions.
That strategy has worked well for firms such as Seattle-based Lighter Capital, an online revenue-based funding vehicle that uses proprietary technology to figure out which companies to back. Lighter Capital has invested in more than 300 companies across 500 deals since 2012 and plans to invest in close to 200 startups this year, CEO B.J. Lackland told GeekWire last month.
A recent PitchBook study found that 38 percent of venture capitalists use data to source and evaluate investment opportunities.
“Our survey shows strong adoption of data to inform investment decision-making and a growing appetite to increase usage,” Steve Bendt, vice president of marketing at PitchBook, said in a statement. “While the majority of respondents believe VC investing will always involve the human element, there’s enthusiasm to explore how machine learning can automate traditional VC.”