Gee, Male: Using My Inbox As a Window To Gender Bias

If you can’t see it, you can’t be it. That’s why role models and representation is so important in changing the ratios of tech as we seek to build a more inclusive community that resembles society, and not our industry stereotypes. Last night I attended an AllRaise event where table discussion included male and female venture capitalists noting the differences in gender breakdown of their inbound dealflow. You can guess how it skews, no surprises here. We track our own gender stats at Homebrew and while we invest in female founders 3-5x more than the industry average, we know that as a two man fund, we don’t fully represent the transformation that we hope to help assist.

If you don’t know her, you can’t fund her. A few years back Google Ventures’ partner Rick Klau ran an experiment on his Twitter follow list and noticed he underfollowed women because his phone contact list (what is often used to bootstrap social graphs) biased male. He took steps to fix it.

Rick’s project reminded me of studies to better quantify representation, namely the Institute on Gender in Media, which analyzes the speaking time that women get in films, tv as well as educating against stereotypes and other imbalances. More casually you have concepts such as the Bechdel test, which asks whether a work contains scenes where two women are talking substantially about something other than a man.

If you don’t communicate with her, you can’t work with her. My inbox and my calendar are the ground truth for where I spend my time and attention. Unfortunately they still gives me very little feedback to understand, or even improve, these allocations. What does this have to do with gender? Well, both Gmail and Google Calendar should be able to provide summary level stats of the amount of time (meetings, email) that I spend with men vs women. What percentage of my meetings don’t include a woman. What are the words I tend to use in communications with women vs me vs mixed gender threads. And a host of other data that could expose me to unconscious bias and help me change my behavior over time.

But where does the gender data come from? As I understand it, Google+ profiles support a wide expression of gender identities. As a Gmail user you already have a Google Account/Profile regardless of whether it’s publicly available or not. Letting people opt-in to contributing their gender data for this research could be an easy call to action. Especially if Google is thoughtful about pronoun usage. Even without user contributed data, I’m fairly certain Google could use names, Gmail text and your Google cookie to make some assumptions about gender. I’m not suggesting that they “prefill” your gender identity in a way that presents to anyone generally -or- that the gender interaction data I suggest tracking above is available in any individually-specific/PII manner – just in aggregate.


Gender research isn’t an area that I hold any expertise in – probably naive somewhat to be honest – but I’m curious if this type of data collection and presentation would be useful and what some of the design considerations should/could be in order to maximize effectiveness without creating any harm.