This story was included in the 1Q 2018 edition of PitchBook's Private Market PlayBook, a collection of our most popular reports, original datagraphics and feature articles covering the key trends in VC, PE and M&A.
Money never sleeps. Stock trading that once took place under a buttonwood tree is now beamed using microwaves and lasers to computer-based algorithmic traders that seek edges counted in milliseconds or less.
But where investors are supposed to live and breathe innovation—funneling money to new and emerging companies—things haven't significantly changed: folks sitting around the table, with decisions based largely on "soft" qualitative factors and the confidence that they can predict which companies, technologies and industries will thrive.
Changing the game
The predominant consideration in a VC deal is the initial appraisal of the management team; first impressions, in other words—a quaint concept in an increasingly quantitative world.
But the old way isn't without justification: Unlike equity or bond investors, who have the benefit of years of well-organized, widely available data, startup investors don't have much to go on. In the vacuum that results, factors like founder personality, team dynamics and personal experience are heavily weighed rather than income statement data or operating metrics like units sold, information that often just doesn't exist.
Cognitive shortcuts try to fill in for the missing data points. Which allows bias to creep in if say an underrepresented founder tries to raise money, since they don’t fit the archetypal startup founder image— the Stanford-educated, STEM-degreed, white-male tech bro.
Yet a growing number of investors are innovating with data-driven strategies. Right Side Capital Management, Social Capital, EQT Ventures, Nauta Capital, e.ventures, Hone Capital, GV and Correlation Ventures, to name a few. They are early adopters of a new strategy, using machine learning and data to automate the process of selecting portfolio companies, eliminate implicit bias from the equation and quantify gut feelings.
Right Side Capital
San Francisco-based Right Side Capital is using its quantitative approach to concentrate on geographies outside the Silicon Valley and New York startup ecosystems, an advantage of not having to physically meet with prospective founders and thus allowing them to serve underfunded areas, avoid the competitive Valley scene and, it is hoped, secure better deal terms.
Managing director Dave Lambert explains the firm’s harsh assumption, based on historical data, is that the startups they fund are likely to fail. So instead of trying to focus on the right industry vertical, pick winners and beat the odds, RSCM makes small investments of between $100,000 and $500,000 (at valuations of $3 million and under) and diversifies aggressively to offset the high assumed failure rate.
Since 2012, they have invested in over 850 companies based on that philosophy. "It's too difficult to predict the future with too many variables and too many potential pivots in areas where the founders have far more expertise,” Lambert told PitchBook. In his mind, if these founders feel confident enough to make the sacrifices to roll the dice and be entrepreneurial, who is he to second guess?
He found irony in the fact one's ego wants to "believe it's an expert in everything" when, in truth, the beginning of knowledge is admitting what you don't know.
Perhaps the most well-known adopter of the Capital as a Service (CaaS) model is Palo Alto-based Social Capital. In October, the firm, led by star VC Chamath Palihapitiya, launched a data-driven operating system for early-stage investing.
Investment partner Ashley Carroll told PitchBook that the experiment resulted in a much higher ratio of underrepresented founders, evidence that the traditional VC process is perpetuating bias. Of the 3,000 companies the firm evaluated, just several dozen received funding. But the startups selected represented 12 countries, 42% were women and the majority were nonwhite.
Proof that the CaaS model works, at least when it comes to funding minorities, according to Carroll.
In a Medium post outlining Social Capital’s embrace of CaaS, Carroll described it as a no-frills model.
“No hoops, no $7 artisanal coffee chats, no designer pitch decks, no bias, no politics, no bullshit,” she wrote. “Just the best teams with the best ideas, the best execution, and the best metrics funded on the merits of their achievements, not the status characteristics of their founders or the exclusivity of their professional networks.”
EQT Ventures in Europe takes data-driven investing to a new extreme. The 2-year-old VC firm, which is part of private equity group EQT, uses an AI-driven data platform called Motherbrain to help it make investment decisions. The firm’s €566 million fund backs companies at all stages—except seed—with €3 million to €75 million checks. So far, the firm has invested in 22 startups.
Analytics partner and former VP of analytics at Spotify, Henrik Landgren, said Motherbrain could’ve identified Spotify and Uber as unicorns in the companies’ early days. He believes letting software play a key role in crafting one’s portfolio is “the next evolution of VC.”
“The better data you have, the better algorithms you have, the better investments you'll make,” Landgren said. “It's so easy to get caught up in a feeling and make decisions based on that. That's one of the big reasons we have Motherbrain, we can take bias out of that decision.”
With a career in analytics—Landgren built Spotify’s analytics team—he’s understandably passionate about data-driven investing: “You should of course use your gut-feeling to make great decisions, but you should train your gut using data,” he wrote on EQT Ventures’ website.
So where does the data come in exactly?
Relying on data anonymizes and standardizes the pitch process. Putting an end to the traditional VC-founder pitch meeting that results in a VC saying, "Hey, I liked them and they have drive." That’s all "soft, lovey stuff," Lambert said, which is difficult to quantify.
But it’s not impossible. For instance, are they technical founders? Do they have significant domain expertise? Do they have previous startup experience? And have they managed people and budgets before? Assign a score, crunch the numbers and leave the cognitive leaps of faith behind.
Not a panacea
Believe it or not, there are still some things humans are better at than their robotic counterparts: identifying strong team dynamics and building relationships with founders that can, in turn, reveal additional information.
EQT’s Landgren said EQT Ventures relies on Motherbrain to handle tasks humans aren’t so good at, like identifying strong investments without bias. That way, the humans involved can focus on what they are, in fact, good at.
James Newell at Voyager Capital, an investor in RSCM, strongly believes that traditional VC still has a role to play since a "full data" approach cannot be relied upon because hands-off "robo-investing" removes the ability for investors to add value. And the ability to measure everything from body language to personal stories and unique circumstance—things that are not easily inputted into a decision algorithm.
Moreover, Newell's skeptical that many early-stage data-based decision making "black boxes" are in fact bias-free. How does one determine which factors to weigh? And which qualities to rank higher than others? Giving a higher score to Stanford graduates, for instance.
But Voyager's investment in RSCM is a strong vote of confidence that there is indeed a role for data-driven VC. Newell admits that he is "in the business of innovation and will look at every potential tool to make better decisions" and wants to avoid being a "cautionary tale of funding innovation but not using it." Yet right now, the applications are limited in his view.
The future of VC
As much as the data proponents want to streamline and expedite the unicorn hunt, the structural nature of VC investing makes full quantization unlikely. Founders aren't just tapping investors for capital but also mentorship, access to expert networks and help scaling growth and achieving profitability.
Thus, in our view, the future of VC is going to be a mix of both old-fashioned gut feel and new-tech AI-enabled platforms. A full robo takeover isn't likely nor desirable. Early-stage data isn't robust enough. And many factors (such as founder chemistry, motivation, grit) are hard or impossible to quantify. But the universe of capturable data points is expanding, and VCs are developing a better understanding of their biases and cognitive shortcomings. Gut alone just may not be good enough anymore.