Almost one year ago, Google released its diversity data, kicking off a wide-ranging industry conversation about why tech companies and venture capital firms are so unrepresentative of the racial and gender make-up of the rest of the U.S.
Dozens of other companies followed Google’s suit while the issue of gender in the venture world has been on the front burner with the Ellen Pao and Kleiner Perkins verdict last Friday.
But Mitch and Freada Kapor have been here more or less all along. Mitch, who co-founded Lotus Development Corporation, which sold to IBM for $3.5 billion in 1995, and Freada, who has a Ph.D and has been active on the issue of diversity for years, run a venture firm Kapor Capital, math and science educational programs for students of color and are in the process of building a new center in Downtown Oakland around their ideas.
They shared some lessons from their work in an interview I did with them several weeks ago:
What would you say has changed about the diversity conversation since you started doing this kind of work?
Freada: What’s better is that we’re talking about it and we see it as an issue. Prior to Google releasing its numbers last May, it was nearly impossible to get any major tech company’s attention. And it still is nearly impossible to get any major Sand Hill Road VC firm’s attention, except Kleiner Perkins which has been sitting in court.
The numbers have gone up and down in terms of white women in computer science. But the numbers have never been anywhere but low for underrepresented people of color.
I’m hoping that what we have now that we didn’t have before is a new framework. I don’t mean that as gloss. I mean it as a deeper more nuanced and sophisticated understanding.
Not only is there a pipeline problem —
Actually, I was working on a story a month or so ago about East Palo Alto and I remember a woman who runs a well-regarded non-profit there telling me, “We don’t have a pipeline problem. We have an investment problem.” It struck me partly because the idea of a “pipeline problem” is repeated over and over. Yet it feels like this is also a way of abdicating responsibility, saying it’s the education system’s fault or whatever and therefore there’s not much really that any individual or company can do about it.
Freada: There is some substance to the pipeline problem issue. But I talk a lot about a “leaky pipeline” problem, which takes many forms — some of which have some layer of bias involved.
You have to look at all the drop-off points before you get into who gets to be a freshman computer science student. If you look at big school district level data, if you’ve failed 3rd grade math and you hit a certain absentee level in your freshman year, you can predict with almost certainty that a student is going to drop out of high school.
So what is it that we’re doing at 3rd grade math level? If someone is failing, can we do something to course correct? To write off a kid at third grade is just heart-breaking. These seem like points where we ought to be rallying, intervening and flipping the argument over its head.
But even with the pipeline problem, I do also think that you have to have an understanding of how bias operates.
What kinds of advice would you give to growth-stage tech companies thinking about diversity?
Freada: My first piece of advice is to stop thinking about yourself as a meritocracy.
Because if you believe you’re a meritocracy and you have numbers like everybody who has released their data, the implicit message is that Caucasian men are better than everyone else. Because that’s who is overrepresented. It’s endorsing a notion that these other people aren’t as good, and therefore that’s why there aren’t a ton of them.
Mitch: It’s a remarkable fact about people that we’re not consistent all the time. At one moment, a person can acknowledge how much they would like to have a diverse mix of employees. Then in another moment, they can also believe that it’s a genuine meritocracy where the best people and ideas get funded. To an extent, that’s what they like about Silicon Valley, its meritocratic nature. That is not completely wrong. It’s just wrong when it comes to matters of race and class.
Freada: Right. There was an MIT, Wharton and Harvard Business School study where they took a business plan that had already won a business plan competition. So we knew that it was investable.
Then they ran a controlled experiment where the same business plan videos were narrated by either male or female voices. When the same pitch was narrated by a male voice, the judges chose it 68 percent of the time.
I’ve been doing a lot of customized talks on hidden bias to help companies understand how — despite their best intentions — they may be making biased decisions anyways.
I try to come up with all kinds of practical steps to mitigate bias at the end of the talk. There are things you can do that are as simple as removing names from resumes.
Or a Rooney rule, [which the National Football League adopted more than a decade ago and requires teams to interview minority candidates for head coaching and senior football operations jobs. There’s no requirement that they hire these candidates, just that they include at least one in the interview process].
Or changing employee referral networks. When I ask if companies have employee referral networks, most say yes. But if I ask whether they’d have a structure for referrals of underrepresented groups, people often think that’s discriminatory and unfair.
They don’t think that having an employee referral system in a company with skewed numbers is by itself unfair. They think that getting to diversity that way is unfair.
Yeah, I understand. There was a profile of Tristan Walker last year in Fast Company and the journalist pointed out that if you’re recruiting from your friends, and they just happen to be more white or more male because people naturally gravitate towards people who are like them, that may be structurally racist even if there’s no deliberate or malicious intent to exclude people.
Freada: One of the things that I think doesn’t work, and we know this from earlier decades of work are things that can sounds like or feel like affirmative action. For a variety of reasons, those are seen as unfair. What happens is the candidates brought in through those mechanisms face all kinds of career limiting backlash
If you look at the law firm world, I remember in the early 1990s, every African-American associate that I interviewed at a top-tier law firm told me that affirmative action was the worst thing that happened to their careers. The automatic assumption is that you don’t “really” belong here. You’re not “really” qualified to be here. You’re only here because of affirmative action. I worry about seeing the same reactions today.
Mitch: There’s still a lot of lazy interviewing at tech companies like how many jelly beans does it take to fill a 747. Not only do they not get at anything, they are quite biased. Google has started to disavow them.
Freada likes to talk about the idea of “distance traveled.” Where did the candidate start from? What hurdles have they overcome?
How do you think about the term “culture fit” as a rationalization for including or excluding people?
Freada: I think there’s a lot of confusion about what’s a “need to have” and what’s a “nice to have.” It’s important to look at the dynamics of culture fit early.
One of my very favorite quotes is from Twilio’s Jeff Lawson. They, by the way, have one of the few diversity groups that is co-led by women engineers. To have engineering deeply involved in company diversity and inclusion efforts is critically important to getting it right and not having it be a side annoying thing.
Anyway, Jeff said something like, ‘When would it be a good time to think about diversity? When I’ve got 1,000 white, male engineers?’
It’s brilliant in its poignancy. Everyone has business fires to put out and if you don’t think about diversity as part of building a great company, you’re never going to prioritize it.
To understand the culture, you have to ask who succeeds here and why? That’s how you can begin to see if it might be inherently biased on the basis of gender or race.
Who do you think is a role model on this front?
Freada: Nobody’s figured it out. We have no one to point to in terms of demographic balance. But when we talk about companies making real efforts, Twilio is doing it in a very sincere way that’s completely consistent with their values. It’s also on their .org side.
It’s baked into the company as opposed to an afterthought. With their developer conference Signal, they reached out to us to figure out how to get a more diverse audience and put together a panel on it. It intersects all parts of their business.
What programs do you think are actually effective? There are lots of programs that Google and others have started to fund like Code2040, and then lots of regional or locally-focused organizations. Which ones are actually good?
Freada: A couple things. I’m trained as a researcher. I believe in rigorous evaluation. All of the groups ought to be looking at outcomes. The second thing is that we’re building a map of this “leaky pipeline” because we’ve got a bit of a herd mentality. There is way too much money going to hackathons teaching privileged girls how to code without any tie-in to anything else.
I understand the appeal of hackathons but it’s got to be connected to something — especially if you’re dealing with kids who are going to be the first in their families to go to college. You’ve got some responsibility to help them at different points of the pipeline too.
Then when I think about other organizations like Code2040, I think about the foundation world subsidizing some of the companies with the biggest market caps on the planet. They’re placing philanthropic dollars that are training underrepresented kids to be placed at Facebook, Google and Twitter. All of those are public companies which already have multi-billion dollars market caps.
Mitch: Of course, if you’re a startup and you’re hiring, you’ll want to hire people who have been at Google, Facebook and Twitter. Having those credentials is really important.
But what Freada is speaking to is when there are training programs for a cohort to be more inclusive, our view is that the host institution, whether that’s a Google or Facebook should be at the very least matching those philanthropic contributions if not going above them. They’re not making enough of an investment in developing these parts of the pipeline. And a lot of that money is coming from other philanthropic sources.
You run your own program, SMASH, for students of color in math and science. What did you learn from that?
Freada: We thought that if you could do a demonstration program to show that there’s a ton of talent and that if you created the opportunity, the kids would come. That was the point of SMASH.
We’re in a stage where there’s room for a lot of approaches. But what we want to talk about with a lot of well-intentioned people who aren’t starting their own programs is how to plug in the holes along the way. We’re mapping out all of the drop-off points so that as opposed to being the 400th person who funds a girls coding program, we can even out the dollars.
And then, we think it’s equally important to expose tech workers to these kids as it is for these kids to be exposed to math and science.
When we get engineers to come in and volunteer at SMASH’s hackathons, they all have the same reaction. They’re blown away by how smart and excited these kids are.
And we think, what? Did you really not think that black and brown boys and girls wouldn’t be as excited about a hackathon?
What about at the other end of the pipeline on the venture side?
Freada: What we’re trying to do at Kapor Capital is to show that you can invest in these communities and do well.
Mitch loves to talk about entrepreneurs building things that scratch their own itch. We talk about entrepreneurs like Frederick Hutson and Ana Roca-Castro who have built businesses.
Mitch: If you locked up John Doerr, Bill Gurley and Vinod Khosla in a room for a year, they could have never thought up what has become Pigeon.ly [Hutson’s startup] or Roca-Castro’s Plaza Familia. Their lived experience doesn’t include federal prison.
But both of these are first-rate businesses in terms of impact and in terms of what they’re going to be able to do financially. If you don’t have those folks at the table, those ideas aren’t going to get created, the beneficiaries of these services aren’t going to be helped, those jobs aren’t going to be created and the investors aren’t going to earn a return.
Are you going to share your returns to prove it?
Mitch: It takes seven to 10 years to prove out a venture strategy. We’re in about year four. I expect we will for better or for worse.
We are running the entire portfolio with the intent to fully reveal its internal performance. Even though it’s our money, we run it now as if there are external limited partners. I don’t mind making five-year forward-looking commitment to this.
How do you think about strategy in terms of being confrontational versus being more flexible in getting the industry to change? (It’s a classic tension in activism.)
Freada: I think there’s room for outsiders to be demanding and shaming. That’s an important role.
Mitch: But change can happen at other levels. Look at Y Combinator. Look at the regime change from Paul Graham to Sam Altman. There was this changing of the guard and now there’s a significantly greater commitment to diversity. It’s reflected in cohorts and in the non-profits and they’re blogging actively about this.
Y Combinator, which is like the Harvard or Stanford of accelerator programs, could have stayed as it wanted to. But they haven’t.
Look at Jessica Livingston. Look at how she’s speaking now. You won’t find her making speeches like this years ago.
The other thing to remember is the time scale upon which change happens. Because we’re in Silicon Valley, we’re so used to this accelerated Moore’s Law timeline because we’ve seen it so many times.
But when you look at for instance, the civil rights struggle, Selma was the culmination of a half-century of work. The NAACP was founded in 1909 and it took a half-century to get to Selma, which is the mid-way point. When we talk about what has and hasn’t changed, we need to think about the time frame upon which human institutions change.
It’s fine to be impatient. It’s one of the dilemmas of being human. But in the case of the big venture firms, they’re probably not going to change until the existing general partners retire.
Freada: People change more slowly than the the technologies they invent.
What do you think about other groups like the Rainbow Push Coalition, which has more of its historical roots in the civil rights movement. They went before the big tech companies at their shareholder meetings and demanded that diversity data be released before many companies actually did it.
Mitch: I’d say it’s necessary, but not sufficient. This playbook was used on Wall Street decades ago and it’s not clear that Wall Street is engaged in lasting change.
Freada: History is also being re-written about these tech companies that started releasing data last year. There is a whole crop of companies like HP, Cisco and Intel that have been releasing it all along.
Their numbers don’t necessarily look any better. Many companies have had had cheap diversity programs over the past 20 years. It’s a worthwhile question to ask about what’s broken with existing approaches to diversity.
I am an optimist. I wouldn’t be doing this work if I wasn’t. There are all of these young companies and people who are very eager to learn and start employee resources groups to focus on diversity.
But if you look at historic diversity programs, regardless of industry and company, they were fueled by a concern about being sued for discrimination. That’s not a positive motivator. That’s not a good business reason.
Last year’s effort on diversity numbers is a good thing. But it is still coming in response to external pressure. Journalists and outsiders have been persistent in asking for the numbers and what companies are doing to fix it.
But it needs to come from within. Twitter ought to be looking at who is using its product and how to mirror that customer base inside.
If you look at the Pew Internet studies on who uses products like Twitter, Pinterest or Facebook, do any of these companies look like their user bases?