Recently I was on a conference call with a group of Techstars mentors and founders, including many women and people of color, discussing a potential Techstars statement on diversity. One of the founders made a crucial point about affirmative action, that whenever a company launches an initiative to increase diversity, it has a potentially dangerous side effect of undercutting the perceived competency of the beneficiaries. It can appear as though they didn’t have to work as hard as others to succeed. As a black man himself, he knows that isn’t true, but it still weighed heavily on his thoughts about the issue.
In startup land, we like to tell ourselves a bedtime story that it’s all a meritocracy and we’re strictly about finding the best people. Hacker News is filled with comments like “who cares what race or gender they are; just pick the best people.” If only it were that easy, my little brogrammer.
Nearly sixty percent of all VC investments fail to return their capital. Professionals, who have often spent decades in technology, can’t pick the loosest definition of a “winner” even half the time. It’s not because they suck at their jobs. Rather, while it’s generally easy to weed out the unqualified, it’s damned near impossible to accurately predict which among the qualified are the most qualified. All of my favorite investors say the same thing — we can’t pick winners, so we just try to improve our deal flow.
And anyone who has ever hired anyone knows that often there’s no quantifiably “best” candidate. One person graduated from Stanford and has three years of experience. Another graduated from a state school and has ten years of experience. Yet another dropped out of college, went to a code academy and has a couple of well-respected repos on Github. So we start looking for ephemeral reasons to choose between them. This is where something called ingroup bias comes into play.
Ingroup bias is “the tendency to favor one’s own group, its members, its characteristics, and its products, particularly in reference to other groups.” We pick the person who went to the same college as we did. Or who grew up in the south like we did. Or just looks most like us. Because when unpredictability creates risk, we take false comfort in the familiar.
One notable characteristic of ingroup bias: the “favoring of the ingroup tends to be more pronounced than the rejection of the outgroup,” — in layman’s terms, it’s different from active racism. Unfortunately, it tends to have the same chilling effect.
A report done by CB Insights in 2015 that is often cited when discussing the issue of race and VC funding found that just one percent of funded startup founders were black. The fact that the last reliable report done on this issue was five years ago should also tell you something.
It’s much the same when it comes to women. In 2019, according to data provided by Pitchbook to Fortune Magazine, funding for teams of all-female founders hit an all-time high — of 2.7 percent. Women of color receive less than one percent.
But statistics show that companies led by underrepresented founders are better investments than those that aren’t. Private tech companies led by women have a 35 percent higher ROI, and, when venture-backed, 12 percent higher revenue than startups led by men. Women-founded companies in First Round Capital’s portfolio outperformed companies founded by men by 63 percent.
So where’s the disconnect? Remember, nearly 60% of all VC investments create negative returns, so why aren’t we seeing more women and minority-led startups get funding? Why aren’t VCs simply requiring women and minority founding team members? Especially in a time where diversity is a high social value, where companies of all kinds are working to overcome unconscious biases? And more importantly, how do we solve it in a way that doesn’t undercut the very people we’re striving to empower?
I have a proposal. And, frankly, I’m shamelessly ripping off someone else’s idea.
One of my favorite entrepreneurs, Ethan Austin, runs the Techstars Western Union accelerator in Denver. Over his years with different programs, he has found the same pattern over and over. It’s really easy to reduce the applicant pool from 1,000 to 15 to 20, and it’s almost impossible to further reduce that pool to just 10. In his experience, there’s always two or three absolute “yeses” — they have a combination of team, market and product that look like obvious winners. But after that, you’re left with a group of 15 or so companies that are equally good and essentially unsortable — there’s no real, quantifiable way to determine which ones are marginally better than all the others. His strategy, in the event of such a tie, is to select the companies who have worked harder to get where they are. The tie goes to the runner.
(In baseball, there is an often-cited interpretation of base-running rules that ties go to the runner. As with almost any rule in all major sports, there’s a fair bit of controversy surrounding whether or not that’s true — made more complicated by the fact that there is no rulebook in baseball that explicitly says those words.)
Unsurprisingly, a lot of those ties go to companies led by underrepresented founders because they have likely faced more adversity and worked harder in their lives. Startups are fiercely challenging, and the ones who stick it out are the ones that win. Ethan knows that underrepresented and minority founders have had to stick it out their entire lives.
To be clear, he always picks the best team possible. Always. The. Best. Team. Possible. No one has ever had their score bumped above a more-qualified team based on race or gender. If one team is batting .350 and the other is sitting below .300, he takes the .350. But when both teams are batting .325, Ethan backs the runners.
But ingroup bias doesn’t just show up at the selection stage. It also appears at the top of the funnel, with deal flow. Women and people of color often don’t get funded because they never even get to have a conversation. When a white male VC asks a white male CEO about potential investments, the likelihood is really fucking high that most or all of his recommendations will be companies led by white men. The chances of me (a white male) having a conversation with a founder who graduated from Howard University are next to none.
And that’s on me.
Anyone who has worked with me in the last decade has heard “Eric’s Theory of Sorting Algorithms”. It states that, when building a subjective list with nearly infinite options — let’s say, restaurants in Los Angeles — returning the objectively best options is an impossibility. I may want Chinese food, but what would truly hit the spot is a burrito. And since there’s no way to A/B test, the answer is unknowable. So instead of shooting for “the best,” we should just make sure to return a list of interesting results.
Simply put, there is no way for Yelp to give me the absolute 10 best dinner options in LA tonight. But if Yelp only returns, say, vegan smoothie joints, I’m going to suspect the site has a bug.
As a community, it’s time to start hiring and funding more than vegan smoothie joints. We need to, as Bryce Roberts recently said on Twitter recently, “Make the hire. Write the check.” And even before that, we need to improve our deal flow. We need to take the meeting.
In said meeting, we need to let go of the bias that tells us the person sitting across from us who doesn’t look like us is more risky or less predictable — because chances are they’ll outperform the people we think we know.
— Eric Marcoullier and Megan Hanson
There’s been an explosion of writing about inclusion these days and it’s tricky to add Megan’s and my voice to the mix. I’m torn between the competing goals of 1) shutting the fuck up and listening to people who need to be heard and 2) being an ally and using my (limited) status to push for change.
In the end, I was simply too inspired by Ethan Austin’s strategy to stay silent. The very first time I met him, some 18 months ago, he taught me that the tie goes to the runner and it’s a phrase I’ve returned to over and over again. Thanks for being an inspiration, man.