I realized my company had to change after talking to an African-American woman who’d just quit Redfin. When I asked her what it was like being black at Redfin, she simply said “lonely.”
I wish all the people engaged in culture wars and political-correctness debates could have been there to tell me what I should have told her then.
She didn’t judge the company based on our diversity efforts; she experienced it based on our actual diversity, as one of a small number of African Americans in our Seattle office.
It was easy to decide then that I had to change the company because I hadn’t yet realized how much I’d have to change myself.
This essay describes those changes, personal and organizational, but it isn’t aimed at the diversity advocates. It’s for the people in my seat, the ones accountable for delighting customers and generating profits, who will lose our jobs and our identities if our businesses don’t grow at a berserk rate.
Redefine hard core
The unholy power we draw on in our darkest hours as an entrepreneur is not our intelligence or creativity, but the intensity of our work ethic.
Our companies are fighting for their lives against giants with more money and people. Often there’s a billion-dollar prize at stake. Once our survival seems assured, we never quite get over what we went through, and we never want to get over it. What we fear most is becoming as complacent as the companies we’ve challenged.
But in many startups, this pressure to work hard is so overwhelming it becomes a form of aggression against anyone with other commitments, including people with children and aging parents.
It follows you into the bathroom and down into the parking garage. It’s why startups are often run by a cadre of young men, single or married to stay-at-home spouses.
When a Redfin team slipped a deadline, I used to ask why their offices were empty at 6 p.m. Our chief technology officer, Bridget Frey, has been militant in objecting to such questions on principle. “You want to know where they are?” she said. “They’re picking up their kids from daycare, just like you.”
Bridget suggested an alternative performance criterion: results. We invested in measurement systems that an entry-level engineer could access on her own at 9 p.m.
It took a year to warehouse data from across the business in one place, to figure out which metric to assign to each team, and to train the team on how to move it.
But it was worth it. Since the two of us embraced this approach four years ago, I’ve stopped asking why people aren’t in the office. Redfin’s traffic has accelerated from 26% year-over-year growth to 45% now.
In roughly that same period, the percentage of engineers in Redfin’s Seattle office who are female has grown from 0% to 33%.
Objective measures such as these are the ultimate refuge for people who have faced double standards and moving targets all their lives, but are also the happy place for CEOs eager to have someone else responsible for hitting a number.
I realized this at a dinner with a portfolio manager who was the mother of five. I told her that the macho hedge-fund world would be the last industry I’d want to work in as a mom.
“It’s the best,” she said. “If I worked for you, my career would depend on whether you liked me. Here, I’m judged on my trades. My partners make too much money from my trades to be sexist.” The most convincing arguments for diversity come from the capitalists, not the socialists.
Embrace bureaucracy (the good kind)
Since we at Redfin are coders and customer advocates, not traders, many Redfin promotions are still subjective, but that conversation got me thinking about how to make them less so.
For years I’d resisted writing down the criteria for hiring and promoting employees. I believed that check-box promotions, where people move up by doing everything right except generating results, create a confederacy of do-nothings.
I preferred promoting people based on the manager’s discretion. That changed when another woman left Redfin for a bigger job at another company, without realizing she was about to get promoted at Redfin.
Her manager never knew what the standard process was for promoting her, because there wasn’t one. That left the departing employee to worry about our basic fairness.
No one should have to worry. It’s your job as a manager first to be fair, but also to convince people of your fairness. Fairness, after all, is the starting premise of any collective activity, whether it’s a kickball game or a technology-powered real estate brokerage. And the only way to be fair about a promotion is to write down what it takes and refer to that regularly.
Redfin has documented the promotion criteria for every engineering role above entry level, and for all directors and vice-presidents.
We now require managers to make the case in writing for every promotion, and also, because the process should be rigorous, the case for rejecting or deferring a promotion. These documents overwhelmingly focus on results.
Does all this documentation sound like bureaucratic baloney to you? Well, bureaucracy was originally conceived in 134 BC as one of the foundational innovations of modern society, where Chinese mandarins were chosen based on a civil-service exam rather than by political or family affiliation with the king.
It helped an entire continent avoid the dark ages. Like all great ideas, it spread to the rest of the world. A bit of it should come to the technology industry too.
Train people to be strategic
Board members who decide someone shouldn’t be promoted view promotion criteria as bureaucracy in its worst sense. We often simply conclude “she just doesn’t have it.” I can’t count the number of times I’ve heard that phrase, about men and women, from me and others. It sounded innocuous enough, until one day I realized it was the sound of someone’s head hitting a glass ceiling.
But it’s also often a fair assessment. The would-be executive being passed over may chase off-base ideas, or get overwhelmed by the sheer number and ambiguity of projects he or she has to handle at a senior level. Promotion denied.
When our chief technology officer first heard that someone just wasn’t strategic enough, she responded with a question no one had ever asked before: “So what are we going to do about that?” My initial response: nothing. If there’s one belief boards hold sacrosanct, it’s in God-given talent, which it is their job to find, coddle and lavishly overpay. You either have it or you don’t.
But most engineers don’t believe in the kind of God who anoints people for corporate leadership. They believe in humans’ power to make things better, and that includes all things, even other humans. They abhor the waste of writing off one of your most talented employees.
So Bridget and I have worked on a program to train Redfinnians on the un-trainable: how to be strategic. We go through the calendars of up-and-comers to clear time for chasing big ideas.
We expose them to board-meeting materials so they’re ready to contribute, not just listen, the first time they participate. We turn their writing upside down to put their conclusions first. We teach them the business, explaining our technology stack and financial statements, even if it isn’t directly relevant to their jobs.
About half the time, none of this makes a difference. Talent still matters, and the would-be executive really has hit his or her ceiling. But half the time, the person breaks through. And this in turn has increased our money-making capacity.
Redfin could never have launched a mortgage business this year while also building an automatic system for scheduling home tours, because one executive was responsible for both projects.
Fortunately, we’d been developing another executive to take over on-demand tours. Since the new executive is an Asian-American woman, this increased the diversity of our executive team, but we think of that increase in diversity as a means to an end, not an end in itself.
Diversity as a product
We still have a long ways to go to take advantage of all the talent the world has to offer us. This is in part because for years, I focused on short-term results, hiring the folks nearest at hand, elevating the loudest voice in the room.
In Silicon Valley, we talk about the technical debt that piles up from years of hacking in quick-and-dirty features, but we also accumulate debts from such short-sighted business practices. To grow into a great company, with better talent than our competitors, Redfin has to set aside time to pay those debts too.
It has helped to approach diversity at Redfin like other challenges our engineers have to solve, first because it brings a relentless mindset to the problem.
When I hear that there just aren’t enough qualified candidates to diversify Silicon Valley, that the problem is too hard, that it’s not our fault, I try to recall another time when Silicon Valley tried to solve a problem, failed on the first or second try, then resigned itself to defeat.
We’ll increase diversity the same way we solve every other problem: by iterating fast, thinking long-term, trying out risky ideas, valuing progress over perfection, analyzing everything.
This is why at Redfin engineering has taken the lead on rewriting job descriptions, retooling career ladders and recruiting from new sources, developing a prototype for improving diversity beyond engineering in the same way we develop prototypes of our products. We never assume we’ve solved the problem, instead measuring whether we’ve made it better, then fixing another round of bugs to make it better still.
The crazy thing is that I still believe in all the utopian stuff that brought me to the technology industry in the first place. I think Redfin is a work of art. I want it to be a meritocracy and a sea of love, for everyone.
We’re still figuring out how to be a city on the hill, and I’m still figuring out how I need to change to make it so, but this is what has helped us get this far.