On Diversity, and why you shouldn’t build your villa in a rice field
I don’t need to defend diversity, but the way I see it, only hiring people you went to school with/look like you/go to the same gym as you is the same as only going on dates with people from your village when you live in a city of 7bn people — boring & unlikely to ever help to expand your world view. I’d go as far as to say that it’s actually even worse than that, because having a boring love life won’t kill you, but having a company that doesn’t innovate (due to limited POVs), will likely not make it.
Diverse teams build better products, and in an effort to ‘fix’ their diversity problems, I noticed that more and more companies use algorithms in a very simplistic way that will, in the long-term, leave them worse off.
As we rely more and more on algorithms to make decisions for us, we have to be aware of the fact that the outcome is only as good as the data they are being fed, and when it comes to algorithms deciding who should be hired and who should not, understanding the context behind the data is crucial. As a society, we are aware of the fact that people are biased (but we’re not accepting of those biases), and so we have trust minimisation systems in place — the legal system, for example — to reduce the vulnerability of participants’ potential for exposure to harmful behaviour. The same system does not currently exist for algorithms, and although they can be used to get rid of some human biases, they can also introduce or exacerbate others.
I read about a big tech company that is ‘developing an algorithm-driven digital platform that scores and prioritises candidates based on the profiles of its top performers’. I’ve read about another company that is building a recruitment service that will help companies do something similar.
I see quite a few problems with this approach:
- First of all, you have to really understand the context behind why someone is a top performer in your company; someone could be a top performer because they have the right manager, the right team behind them, they’re working on what they love, etc. I will give myself as an example — I went from being ‘literally the worst’ in a company to ‘amazing’ in another — I am the same person, but the context was different. Also, you have to be mindful of what ‘top performance’ looks like in your company. If you measure the time someone spends at their desk or how punctual they are, you’re doing it wrong, because these metrics really are no indication of anything at all — except for maybe them having a shorter commute to work, or being slower to get their work done because they’re overwhelmed, or aware of this being a metric and just gaming the system 😄. In one company where I was considered ‘not great’ at my job, I was expected to be at work at 9am on the dot — even 1 minute late and my manager would chastise me for days. As a result, I started resenting my job (even though I loved what I was actually doing), and became less and less motivated to get anything done, until I quit and joined a more flexible startup.
- Secondly, unless you’re a truly diverse and inclusive company, the data you’re feeding your algorithms will be biased. Once bias enters your system, feedback loops will reinforce that bias instead of getting rid of it. An example — let’s say that an algorithm decides that you should hire candidate A, who is very similar to a top performer in your company, and not candidate B, who is close (definitely qualified, but she’s self-taught instead of going to Oxford like 60% of your company); you hire candidate A (after a few rounds of interviews, of course), and don’t give candidate B the chance to at least interview, because you never actually see her application. The algorithm receives the feedback that the candidate suggested got hired — which reinforces the bias — and although candidate A turns out to be great, candidate B could have been even more successful, but there’s no way of knowing that. An unbiased algorithm would have given candidate B a fair chance, too.
- Thirdly, you’re never going to be a wildly successful company (and a great one to work for) if you hire the same person over, and over, and over, and over again. A team of Steve Jobs’ would have NEVER built Apple. Also, as with any other investments, past performance is not indicative of future results, and people are no different.
- Fourthly, algorithms cannot work alone, and they cannot be the only deciding factor as to who you hire and who you don’t, but they could be a great tool of a highly successful recruitment team. I believe (and it’s one of the things I’m working on) that you can have truly unbiased, off-the-shelf, AI-powered chatbots that take into account the importance of diversity to help build amazing teams 🙂
Why you shouldn’t build your villa in a rice field
About a week or two ago, the coworking space I’m working from organised a seminar on the topic of sustainability; one of the things we discussed was the problem of accessibility to food, particularly in Indonesia, and although none but one of us had any idea what we were talking about, we all had an opinion on what to do to fix it (from having fewer children — yep, really 🙃 — to GMOs, to permaculture, etc.). After discussing it for a while, one of the attendants put her hand up and said something that I think about a lot, because it’s so simple and yet so effective: ‘Stop building your villas in our rice fields’. That’s it. Nothing else, nothing more. Turns out that she’s a farmer who now works with an NGO to help other farmers who have been displaced by increasing prices of rice fields because — you guessed it — they’re a favourite spot to build villas due to their gorgeous, Instagrammable views.
So, in conclusion, when making your hiring decisions, please be aware of your biases, those of your team, and those of the algorithms that are now being sold to you as the holy grail of recruitment. Ask about the data that has been used to train these models, ask for feedback from the candidates who are being interviewed by your team, and do everything you can to truly understand what you could be doing better. Otherwise, worrying (or claiming to worry) about diversity and inclusion while not taking the necessary steps to change things would be as hypocritical as worrying (or claiming to worry) about starving Balinese farmers while building villas in their rice fields.