How We’re Thinking About Diversity At Homebrew: Start By Measuring It

This is an unusual VC post – we’re asking for help around a new metric we’re tracking but it’s in an area where we’re not experts: underrepresented founders.  We’re hoping to learn from those who have spent more time on issues of diversity and inclusion.

“What’s on your dashboard?” It’s one of the questions we ask startups during the fundraising process and post-investment to help them collect the information needed to run their business. Since Homebrew considers itself a startup – albeit one which writes checks instead of code – we too have a dashboard, primarily focused on dealflow since we’re too early in the fund to yet see results from our seed investments. Satya recently posted data from 2013 around our investment process and we hope it helped provide additional insights into how early stage investors think about opportunities. While most funds don’t publicly share to the same level of detail, the information we presented was pretty traditional.

Satya and I believe in “measuring what matters” (as the management cliche goes) and our dealflow sheet includes data such as how the company reached us, reason we passed, names of founders and so on. We plan to augment these basics with more information about how the company fared over time so we can look backwards and see whether we were good pickers. There is one aspect of our tracking that we held out from Satya’s post because it felt worthy of a standalone discussion: the gender and race of founders.

Why do we want to track this personal data about the people pitching us? Not to factor into an investment decision and not to opine publicly as the voice of the “Enlightened VC.” It’s because Satya and I believe that we can realize outsized returns for Homebrew only via meeting a broad set of founders and if we’re not seeing diversity at the top of the funnel, it needs to be addressed. This is an economic rationale for us. We understand there are also moral arguments to be made here but from the standpoint of investing we’re focused on an empirical truth: diverse teams perform better.

Of course this applies to the entire team composition but if our goal is to see a wide range of highly promising startups in our dealflow, starting with founder diversity seems reasonable. Here are the questions we’ve asked ourselves:

1. Is It Actionable? Yes. If we saw a lack of, for example, female founders working with Homebrew, we’d diagnose and correct. Perhaps we should mentor at female-focused incubators or speak at Women 2.0 events. Maybe there’s something about our brand positioning that’s making Satya and me seem less approachable – for example, we currently lack a female investor on the Homebrew team. Or it could be there’s some hidden bias we hold that causes us to screen out talented female founders during the evaluation process. To date we’ve funded a company with two female founders but not yet any underrepresented people of color founders. We believe there’s an opportunity to be more successful here and being diligent about measurement is a start. We can’t fix bugs in our system unless we have a bug tracker.

2. Is It Accurate? Decisions are only as good as the data you have. Should we ask the founders directly, letting them know why we’re collecting this data, or just collect it passively? Additionally, gender and ethnicity are only two aspects of diversity, why have we deemed them important compared to say, religion or socioeconomic background? Our current thinking is that recording something is better than nothing, and that while there are many types of diversity, it makes sense to focus. So we selected two attributes where there’s a large amount of research on the severe underrepresentation in tech (for example, see LPFI/CODE2040’s infographic). Without data Homebrew is left to fall back on anecdotal information and our flawed short-term memory.

3. Is It Marginalizing? To be transparent, this is our biggest concern. Would our attempt to look at this data actually produce unintended consequences? One friend suggested our efforts should be aimed in an opposite direction via a variation of the “blind” auditions of female orchestra musician study: would it be possible to first evaluate pitches without knowing the gender or ethnic backgrounds of the founders (which are often apparent via the “Team” slide of a pitch deck, or of course, from in-person conversation)?

Or perhaps we just shouldn’t be talking about this publicly at all. Because we operate as transparently as possible, creating a process that’s intended to be hidden from entrepreneurs feels wrong. So we’re left wondering whether this data initiative will end up marginalizing those we’re tracking, which is not our aim. We don’t want to start immediately categorizing founders we meet with by their gender or heritage. Rather we want to review our metrics over longer periods of time to see whether or not we’re meeting with a diverse group of entrepreneurs.

4. Is There Legal Risk? Our lawyers advised that if we track this data, we should clearly note why, in order to mitigate any misinterpretation down the road. That is, there’s a chance that if we track, say whether the founders were male or female, someone might use this data in charges of discrimination. The risk feels low to us and we’re comfortable bearing it.

So Satya and I are putting this out there for comments, especially knowing there are individuals who have worked their entire careers to understand diversity and hidden biases. We’ve already gotten some private feedback from a number of folks within the tech industry but want to open up to others. We’d appreciate additional thoughts, especially from other investors. Thanks!