Diversity is key

I've always been hesitant to write about diversity. I'm a white cishet man living in the European capital of startups: I'm the definition of privilege. And yet I can't imagine writing a book about leading developers that ignores diversity.

When I find a problem hard to grasp, I sometimes think about it in terms of opposites: it's a technique that helps me define the main characteristics of the problem at hand. Understanding why our industry has such a broken hiring process and scores so badly at diversity was one of these hard problems for me. So let me try to break it down for you the way I understood it.

The opposite of a diverse team is a team of people with the same background, speaking the same native language. The most tragicomically common version of it being a team of "X white males aged 25 to 30". Monoculture teams are far too common in our industry. I talked to many people leading such teams and their thinking showed clear signs of misleading assumptions. I'm going to examine the assumptions that occur most often. Let's start with a question that we all know the answer to so well it may feel rhetorical:

Why do we build teams?

The answer is obvious: we want more. We want to build more features, we want to write more docs, we want to fix more bugs. We always want more.

The assumption is that more people get more things done. We also expect bigger teams to maintain or even increase the quality of the output. At the same time, we know that teams come with overheads of communication and synchronisation. So what offsets the increased costs? The underlying assumption is that we believe in collective intelligence.

We believe a team of people is smarter than just one person. When we say "we need a bigger team" what we really mean is "we need a smarter team". The problem is that increasing the headcount doesn't equate to a smarter team. Monoculture teams of "white male kids reading Hacker News in their free time" produce poorer results than diverse teams.

A smarter team is a diverse team.

Monoculture teams can't be as smart as diverse teams.

Then why is there so little diversity in the industry? I fail to come up with an answer that doesn't imply that we, as an industry, aren't as smart as we think we are.

If we want more of everything, then we want smarter teams. We want to increase our collective intelligence. So there is no better strategy than striving for diversity: it dramatically increases your chances of getting a smarter team.

Diversity isn't great because it's cool. Diversity is great because only diverse teams have a chance to be great at design. If increasing diversity is not your answer to building teams, then you are wasting time and money. It's cheaper and faster to hire freelancers.

Another misleading assumption I often run into is cultural fit. Most people hire or fire based on the belief that cultural fit means team members must be alike and think alike. That isn't cultural fit though, that's cultural homogeneity.

You don't need people to be alike, you need them to fit in with each other. It is ironic how we transform the meaning of cultural fit into something much closer to its opposite, while the name makes it clear: cultural fit is about fitting together.

Think about it the way you'd think about putting a band together. You don't need 4 guitarists and no drummer. But you also don't need to hire all jazz players. In fact, a band works great when they have a common musical goal and all their members have different musical backgrounds.

Perhaps the most complicated assumption is that diversity is just about gender. I'm not sure why this is so, I guess it's because one of our industry specialties is making our spaces hostile to women. Most companies have an awful gender distribution and they think it's good enough to reach a two-digit percentage of women in their teams. That isn't nearly enough, and thinking only about the gender distribution isn't enough either. We have to get rid of this assumption. It's hard to do because the problem is hidden well behind the gender issue, which is also complex enough on its own. But you can't ignore that there are more dimensions to this problem than just the gender. There is age, technical seniority, country of origin, religion or lack of it, sexual orientation, personality, area of expertise. I'm sure I'm biased in many ways and I'm missing some dimensions! The point, though, is that there are many dimensions and they all matter. Each dimension brings something to your teams: a way of thinking you didn't consider, a new attitude towards the problems your teams need to solve.

Diversity is the currency for intelligence.

Every time you increase the diversity of your teams by one or more dimensions, you are buying intelligence. You are making your team smarter because you are adding something new you didn't have before. Think of hiring as an addition exercise: "What would this candidate add to the team?" is a better perspective than "what can this candidate do for us?".

In one of the companies I worked for, we built a team of 30 people in less than two years and had almost perfect gender distribution across various disciplines. We scored pretty well on many other dimensions too. Many people asked me how we did it. It's simple: you have to care about diversity every single day from the first day. Simple doesn't mean easy though. The more diverse you want your teams to be, the more inclusive your processes have to be: job ads, screening, hiring and firing processes, written and verbal communication. It requires constant vigilance over your biases and your assumptions. The more people you work with, the harder it gets to keep the focus for a leader. More people means more biases and more assumptions too, which is an interesting catch-22 situation. You have to work with your existing teams every day because you need to stay diverse. As I said, simple isn't easy.

I want to close this paragraph with a trivial example I use all the time. It's about a day I'm running an analysis on the bug my teams had worked on in the past few months. After some crunching, an obvious number stood out. More than 85% of our frontend bugs happened on Windows browsers. I was puzzled at first but then something really obvious hit me. My whole organisation used mac laptops. All our frontend developers used Chrome. We were a monoculture! You know why I use this story all the time? It's because it happened in the team I just mentioned. We thought we were a diverse team. Well, we weren't if we had considered the OS/Browser dimensions!