The tech industry’s woman problem: Statistics show it’s worse than you think

November 7, 2013
November 7, 2013

One of the most frustrating things about the tech industry’s woman problem is the paucity of reliable data on the number of women working in technical roles. Now, thanks to a public Google spreadsheet created by Tracy Chou, a software engineer at Pinterest, we have data on how many women engineers work at 84 different tech companies. To collect the data, company employees have been performing internal head counts, and most contributors have identified themselves openly, though Chou invites anonymous submissions via email. Contributions have also come from people who are manually counting the number of women on companies’ team profile pages. Chou has focused her efforts on women engineers, defined as “women who are writing or architecting software, and are in full-time roles.” Until now, there have been little data on how many women are among the prestigious and well-compensated ranks of engineers, as opposed to the many less technical roles within the industry.

The numbers, while preliminary, are revealing: tech companies employ an average of 12.33% women engineers. This is consistent with what I’ve observed over the course of 16 years working in the industry (12 of which I spent running a web design and development firm) and what I’ve heard from others. The numbers also map neatly to current figures on women computer science grads (pdf), which suggests the “pipeline problem” argument is legitimate.

Among the companies listed, gender diversity varies. A handful are at parity or better: Levo League, Hackbright Academy, and Yellowsmith—all companies, incidentally, with women at the helm—boast 67% women on their engineering teams. The Muse sits at 75% and Kabinet and Spitfire Athlete both hit 100%, though both have two-person engineering teams. On the other hand, 15 companies on the list are without a single female engineer: Treehouse, 37signals, and Causes.com among them.

While some of the smallest teams have 50% or more women, the numbers drop significantly once you look at engineering teams of 10 or more. When I broke down the data by size of engineering team, the averages looked like this:

Fewer-women-on-large-teams

My segmentation is somewhat arbitrary, but on average it looks like bigger teams have a lower percentage of women. It’s easier to get your percentages up when you’ve got a four-person technical team than it is when you’re hiring by the dozen.

Companies that participate in the counting do so as a signal to prospective employees that they are committed to diversifying their teams. A spokesperson at Mozilla—the largest company on the list, with a 500-person engineering team but only 43 women—told me that the project is “a reminder to keep pushing for more diversity.”

Until now, those of us writing about tech and gender have been making do with broad, US-centric data from the National Center for Women and Computing and the Anita Borg Institute, along with other sources we collect piecemeal. These data are tricky because they don’t typically differentiate between departments and roles within organizations: A woman in the HR department at Cisco will typically be counted as a “woman in computing,” whereas a woman software engineer at an investment company won’t. NCWIT suggests that women hold more than 25% of “computing occupations,” whereas my personal experience in the sector, which I’ve heard echoed by many colleagues, is that the numbers are significantly lower among software coders.

Even federal regulations have not provided us with reliable information: In March, when CNN was looking for data on women in tech, it was stonewalled by Silicon Valley giants whose size requires that they report diversity stats to the Department of Labor. While the government has that data, it won’t release it publicly, and most of the big companies aren’t talking. It’s unfortunate since data from these companies could be particularly valuable for benchmarking purposes given that their engineering teams are big enough to be statistically significant. When I tweeted earlier this year about that CNN story, one commenter suggested that sharing the data would result in a PR nightmare for the companies in question.

Marc Hedlund, VP Engineering at Stripe, has a different perspective: “If the first step is admitting you have a problem, I think in this context you have to say that very publicly for it to matter. Every company has a problem; your willingness to face it and work on it is what matters.”

Chou describes what motivated her to collect her own data: “While companies talk about their initiatives to make the work environment more female-friendly, or to encourage more women to go into or stay in computing, there’s no way of judging whether they’re successful or worth mimicking, because there are no success metrics attached to any of them […] As an engineer and someone who’s had ‘data-driven design’ browbeaten into me by Silicon Valley, I can’t imagine trying to solve a problem where the real metrics, the ones we’re setting our goals against, are obfuscated.”

While Chou’s project is a good start, there’s still room for more granularity. She told me she’d love to gather data on the backgrounds and roles of both men and women engineers at each company in order to learn whether there are patterns around junior/senior positions or traditional versus non-traditional training. I’d like to look at the gender ratios for each company’s executive team and board, and see if there is any correlation with the company’s track record on hiring women engineers. It appears from the current data that many—though certainly not all—women-led and -focused startups have higher numbers of women in technical roles.

Another helpful way to expand this project would be to add temporal data, to chart how companies fare over time. This would be especially helpful in combination with data about companies’ policies and programs (if any) for recruiting and retaining women.

Etsy CTO Kellan Elliott-McCrea said: “The best, and arguably only, approach to changing something is first to be able to measure it. You change, you measure, you change, you measure… It’s as important for organizational practice as software.”

He adds, “I’ve heard projects like Tracy’s described as consciousness raising exercises, but I think that undersells the value of data.”

Two of the tech industry’s great strengths are its relentless focus on data, and its orientation toward the future; it’s time for companies to use these strengths to tackle their lack of gender diversity. This project gives me hope that tech companies, who are strong proponents of company dashboards and other forms of tracking and sharing metrics, will rise to the challenge of analyzing what works—and what doesn’t—when it comes to changing the ratio on their technical teams.

You can follow Lauren on Twitter at @LaurenBacon. We welcome your comments at ideas@qz.com

Top News

Powered by WordPress.com VIP