Nearshore Americas

Q&A: The Fascinating Process of Using ‘People Analytics’ to Eliminate Recruiting Bias

Data is driving everything. Although HR departments have yet to catch up entirely, the emergence of people analytics is pushing recruitment towards more tech and data-driven practices, allowing companies to chip away at the biases embedded in their hiring processes.

Chilean company Genomawork is one of the several organizations that’s combining tech and people analytics to inch closer to the promise of bias-free recruitment. The company –which has presence mostly in Latin America, with some clients in the US– has made a name for itself among its over 150 customers thanks to its “game-based” approach to data collection. Candidates play games, a method which allows Genomawork to sidestep some of the most common traps which lead to bias in data collection and analysis.

NSAM spoke with Fabian Martinez Navalon, Founder and Executive Director at Genomawork, to have a better understanding of how his platform works, what makes it different from other recruiting tools, what pain points are customers seeking to alleviate with such technology and the future of business once the full weight of the people analytics wave hits. 

NSAM: How does Genomawork function?

Fabian Martinez Navalon, Founder and Executive Director at Genomawork

Fabian Martinez: We use games to collect data related to cognitive, emotional and personality traits. Once the candidates have played, we have a better understanding of their traits in those dimensions. We have thousands of variables, which translate to 70 traits.

Our process is as follows: first, we invite workers from a specific company to play a series of our games. Starbucks, let’s say. Starbucks will share with us performance indicators for these workers. We train the algorithm to understand which are the patterns that relate to performance. Basically, we seek a correlation between cognitive, emotional and personality traits with job performance. With that information, we build an “ideal genome”, which is a set of traits which characterize the best performers for a specific job position. 

In the second step, candidates play our games. They’re fun and provide feedback. 

The third step involves the platform predicting which candidates have a higher probability of success for a specific position.

NSAM: What kind of games are we speaking of? 

Fabian Martinez: We call them games, but they’re actually cognitive experiments, scientifically proven and endorsed as measuring tools by hundreds of papers. We took this scientific basis and “gameified” the experiments. 

For example, there’s an experiment that involves balloons, designed to measure risk aversion. Let’s say I give you 20 balloons. You have to pump them with as much air as possible, because the bigger the balloon, the more you get paid. But those balloons can pop. Some people pump the balloons too full, which can cause them to pop, or maybe not. Some people would rather have none of them pop. We took this experiment, turned it into more of a game and now use it to measure risk aversion.

The games are short; two or three minutes, at most. And each one measures a specific trait.

NSAM: What traits are measured with these games?

Fabian Martinez: Risk aversion, multitasking capabilities, planning, altruism, extroversion, impulsivity, negotiation abilities and styles, openness to experience, capacity to learn through feedback, and many others, which cover the three dimensions mentioned before: cognitive, emotional and personality traits. 

NSAM: Genomawork was built on the idea that current HR tests don’t work. How are they failing?

Fabian Martinez: On several points. To begin with, when a person postulates for a job, there’s a lot of what we call “social desire” to be evaluated positively. If we’re in a job interview and I ask you if you’re a proactive person, you’ll most probably tell me that you are. The response is determined by your wish to land the position. There’s a bias which games elude, because you’re deep into the game and you’re not quite sure about which is the right answer. In a way, there’s no right answer. 

The second point comes down to the importance of the employer’s brand. Historically, companies held all the marbles. Today, a lot of talent is actually choosing where to work.

We work with big clients where many of their potential employees are customers. In Starbucks’ case, it’s highly probable that someone who’s postulating for a job with them is a customer. If the candidate had a good experience postulating, that’s good. A bad experience can cost customers to the company.

Illustration of Genomawork’s model of data interpretation

Lastly, companies have been using the same recruiting methods for a long time. If you look at how it was done, let’s say, in the 1940s in New York, recruiting is not much different from today. In contrast, areas such as marketing have changed a lot. Marketing used to rely a lot on guata [feeling]; now it is more data driven. Today, a data scientist can have a career in marketing. 

In HR, we are still relying a lot on feeling, with a lot of bias. But now we’re seeing a wave of people analytics which lands all of these variables. And what better way to collect data than games?

NSAM: How do you manage to avoid bias in your own data analysis?

Fabian Martinez: We have a social mission. We want to level the playing field in Latin America’s job market; we wish for people to shine for their talent, not for reasons that have no correlation to performance, like a last name, gender, social position or their network of contacts. There are studies which show that, over here in Chile, a candidate with a “nice” last name can earn up to 40% more than one with a more “latino” sounding last name. 

From the beginning we have worked to lower bias. We have an impact committee which meets every month to work towards a better experience for all candidates; for them to be evaluated in a matter that is fair. 

Over 800,000 people have been evaluated using Genomawork. We have managed to understand how populations with different traits behave and used that information to reduce bias from a gender and social perspective, among others. 

NSAM: What sort of pain points are your clients looking to alleviate when they come to you?

Fabian Martinez: Again, three key points. First, they seek to use data for hiring; to have processes which are bias-free. 

Secondly, they want to predict performance. The recruiting department rarely coordinates with whoever is in charge of measuring performance. With Genomawork, we managed to make this measuring process objective.

Companies have incorporated these bias-free models in quite a short time, and they’ve also grown very demanding when measuring performance

Third: betterment of their brand as an employer. Before, companies used traditional recruiting methods. Now, Genomawork betters the candidate’s experience. You can ask the over 800,000 people that have been evaluated with our platform. In a scale of 1 to 10, the average of satisfaction is 9.2. 

NSAM: Do you have plans to enter the US market?

Fabian Martinez: Our focus is Latin America. We’re leaders here in people analytics. But we do want to reach the 1,000 client mark. We’ve been opening up the field with clients from Brazil and the US, but we haven’t lost our main focus, which is in LATAM. 

NSAM: Are the pain points the same for customers in the US as in LATAM?

Fabian Martinez: They’re very similar. 

NSAM: How long do you think it will take technologies like Genomawork to become tools of common use among companies?

Fabian Martinez: The concept of people analytics is fairly new, particularly in LATAM. I would say that there’s a gap between what’s seen in the region compared to Europe and the US. There are companies out there with similar offerings. There’s iMetrics in the US and Arctic Shores in England, which do something akin to what we do. 

I would say that, today, changes happen in a dramatically faster fashion. People analytics is something that executives have been asking for lately as a strategy to combat discrimination in their organizations. Some countries have made the fight against discrimination their objective for the year; the OECD and IDB have put the challenge out there, aiming to end gender discrimination in particular. 

Sign up for our Nearshore Americas newsletter:

Companies have incorporated these bias-free models in quite a short time, and they’ve also grown very demanding when measuring performance, which is something that has happened for years in other areas, like marketing.

In HR, this wave of people analytics is coming. We [Genomawork] are right there, in LATAM’s core position. There’s no solution like ours, broadly speaking.

Cesar Cantu

Cesar is the Managing Editor of Nearshore Americas. He's a journalist based in Mexico City, with experience covering foreign trade policy, agribusiness and the food industry in Mexico and Latin America.

Add comment