Why I Care About Diversity

Romain Lerallut
Code Like A Girl
Published in
5 min readSep 11, 2017

--

I used to be a strong believer in the “strict meritocracy” and “if you build it they will come” principles. Research, reflexion, and discussions have changed my mind. Here’s the process I went through, with links to more readings.

I am an Engineering Director at Criteo. As a white male in my late 30s, I belong to the majority population. I have plenty of role models to choose from. I don’t suffer from any form of discrimination. I am never the target of offensive comments or even harassment. And I never wondered if I got my job because of my looks. So why do I care so deeply about diversity ?

I’m an engineer. I like to measure things and optimize them. Good news : studies have shown that diverse teams are more efficient. Quoting HBR : « A 2015 McKinsey report on 366 public companies found that those in the top quartile for ethnic and racial diversity in management were 35% more likely to have financial returns above their industry mean, and those in the top quartile for gender diversity were 15% more likely to have returns above the industry mean. »

Or Scientific American. More science, less business. Same message.

Encouraging diversity not only improves team performance, it also broadens your talent pool. Not all of the best engineers come from the majority population. As a space geek, my own personal hero is Margaret Hamilton, who not only invented the term ‘software engineering’ but whose fault-tolerant software allowed the Apollo 11 lunar module to actually land on the moon despite user error. You can’t get much cooler than that, in my book. Soft landings are better than accidental lithobrakings.

Margaret Hamilton with the code of the Apollo Guidance Computer MIT Museum & NASA

From a very pragmatic point of view, encouraging diversity in teams has general performance benefits. But there’s more to the story. I’ll start with a personal anecdote, bear with me.

Some time ago, I was chatting about this topic with one of our (female) Product Managers. She was telling me that women should be managed differently from men. They were more sensitive, especially to negative feedback. They needed more encouragement and reassurance because they had lower confidence or even self-esteem. Therefore, according to her, a manager should adjust their style to better fit the needs of the women they worked with. It made a lot of sense to me at the time.

But then, on my way home, it got me thinking. Something didn’t feel quite right. Then it hit me. I may hide it well, but I also need a lot of reassuring. I also thrive on encouragement. I want my regular fill of ‘attaboys’ (BTW is ‘attaperson’ a word ?) and pats on the back. I too value collaboration and do not feel comfortable in highly competitive or even aggressive environments. I actually would like to be managed « as a woman ».

Then I understood why I care about diversity : making your tech workplace « diversity-friendly » also makes you work hard at rejecting aggressive, offensive, or excessively competitive behaviors, showing people you care for them and their success, growing them instead of belittling them. It may be a cause, a consequence, or an accidental correlation, I don’t care. When we work on diversity, we make a better workplace for everyone. *I* want diversity because it makes *me* happier.

Diversity builds stronger teams. Diversity creates a friendlier workplace. Diversity good. Monoculture bad. Fine words. What do we *do* now ? Well, if I had a perfect recipe, I would tell you. But here is some food for thought.

First of all, be aware that you have unconscious biases (yes, that’s a bit of an oxymoron, so sue me). A common bias is that we judge people in the majority population on their potential whereas we expect the minority population to actually demonstrate performance to get equally rewarded. We also decide outcomes first and then unconsciously find reasoning to justify our decisions (Yale study on job discrimination , Blind orchestra auditions). You can start by working on that.

Second, ask those minority people what they perceive. Maybe your biases are unconscious to you but it’s likely they will be visible to them. So ask, and listen to the answers. Brace for impact. (Note : by asking this, you are also reinforcing their feeling that they are « special » and « don’t quite belong ». Damned if you do, damned if you don’t).

Third, give your minority people a voice. Make sure that they are well represented in various speaking opportunities. Make them visible. Make it very obviously clear that their opinions are heard and valued. Start in your routine meetings but only after reading this on speaker perception imbalance and that on interruptions in meetings. Fun fact : you’ll improve even your all-male meetings.

Finally, help foster the creation of dedicated groups so that they can discuss their issues together. Hey ! Isn’t that discrimination ? A bit. So what ? It is a place where they are the majority population, for once. Some people will feel more comfortable contributing in front of peers rather than in a bigger group where they are the outsider. It can be very stressful to be in a minority all the time. Also, role models don’t appear by chance one fine morning. They need to learn to trust their skills and realize they can guide and inspire others. When you face a dearth of such key figures, you need a way to prime the pump.

On role models and peer groups : It’s hard to turn stereotypes more on their heads than supermodel (and Ruby coder) Karlie Kloss encouraging teenage girls to code with Kode With Klossy (@kodewithklossy)

Parting words : On September 28th, 2017, Criteo will be hosting the first meetup of the newly-created Paris chapter of Women in Machine Learning and Data Science (@WiMLDS_Paris) High-quality tech talks by female speakers, followed by workshops. Men are absolutely welcomed to attend. It’s just that the speakers will be women (contrary to most conferences or meetups on the topic). I plan on going and for once, I’ll keep my mouth shut and my ears open.

I would like to thank my colleagues Nicolas Le Roux (@le_roux_nicolas) and Cédric Roux for the work we did together on this topic some time ago and who contributed many of the resources linked in this piece.

Additional resources :

--

--

R&D VP, head of @CriteoAILab , RecSys fan. #MachineLearning at scale to Make Advertising Great Again. Learn on petabytes, predict in milliseconds.