This post is more than two years old. As a result, it may contain information that's out of date or that I no longer agree with.
Posted 16 Jun 2015

In the last week I’ve participated in an wave of conversations about tech companies and diversity initiatives, spurred on by talking about things about sponsors of the Grace Hopper conference, and most recently, the Portland tech diversity pledge. As a result, I want to ask the one question that’s missing from every single announcement about these programs: Is there the slightest shred of evidence that any of them work?

I mean it. In the last ten years, I’ve observed a large number of tech companies’ behaviors, and rare is the company that has significantly made a positive change to the diversity of technical staff, and in particular their company’s leadership.

What I have heard instead is a common refrain: the pipeline leads to a meat grinder (or sewage plant, take your pick). That is, tech companies focus on increasing the diversity of their incoming hires, but for those who survive the hiring process, the situation inside remains extremely toxic.

Here’s my theory: by the time a tech company has managed to secure funding, they’ve already decided, implicitly or explicitly, who they want to be. That choice is reflected in all their actions and hires, and if the environment is dominated by white men, that’s how it will stay. I’ve heard plenty of companies worry about maintaining their culture as they grow. Far, far fewer talk about overthrowing it entirely because they’ve realized that culture is fucked.

True, it takes a brave, self-aware person to do that, but we’re supposed to be leaders! We thrive on disruption!

How are tech diversity initiatives measured, then? Primarily by the number of dollars spent. For example, last year Google spent $115 million on this but the impact on hiring and workforce demographics was small. Other companies put in big dollars too, for example at the start of 2015, Intel pledged to spend $300 million over the next five years. Google is often cited as a thought leader in this area, but as far as I can tell that means they’re good at communicating what they put their money towards.

One more thing: have you noticed how all the focus is on measuring the pipeline, and no one talks about real metrics for employee retention?

This is where the conversation about Grace Hopper comes in. After last year’s debacle, many of us started to wonder if the event merely functioned as a way for companies with bad diversity numbers to launder their reputation. What other answer is there, really, when sponsors are allowed to create an entire panel about women in tech made up of men?

Keep in mind too, that many of the external organizations companies give money to in order to support or promote these diversity initiatives are not-for-profit or tax exempt in some way. For example, the Anita Borg Institute which runs GHC (their most recent public filing with the IRS was for 2013 which means we can’t see any numbers that tell us more about GHC 2014 yet). A well designed program could then also serve to give the company useful tax breaks.

If none of these initiatives really work, what will? I hope you’ll find this idea stunningly obvious: fund companies started by people from underrepresented groups, and the organizations that support them. You may be wondering, how does that help companies like Google? Think of all the strategic acquisitions, “acuhires”, big tech companies make. How much seed money would it take to create a viable pool of small diverse companies to later buy out? I bet $150 million would make a pretty great start.

I was inspired to think about this because of my own experiences and those of several colleagues, but it turns out Intel is also on board. They’ve allocated $125 million to invest in minority-run startups. Of course, that’s only a tiny fraction of their profits (in 2014, they reported a net income of $9.6 billion).

And Google, if you like my theory and want to try it, I hope you’ll also purchase a few dozen subscriptions to my magazine.

Past | Future | Random