mind-the-gap-ethnicity-pay-gap-featuredI know first hand that tech has a diversity problem. As a computer science major and a career data scientist with a PhD, I’ve been the only woman in many classrooms and meetings.

My experience is not surprising, unique, or unknown: there is a very public conversation about the lack of diversity in the technology workforce. This is a well-known issue. However, what surprised me as a new hiring manager was how institutional this problem is and how challenging it can be to make progress.

I work for a startup company that was co-founded by a woman. Many of our leadership positions are held by women (Head of Product, Head of Data Science, Chief Customer Officer, etc.). We regularly have conversations about diversity issues and our shared frustrations with our industry. But when I was promoted to head of data science, I quickly learned that the best intentions are not nearly enough to build a diverse team.

I am very familiar with pipeline issues, one of the many reasons offered for lack of diversity: there aren’t enough women qualified for technical roles because they drop out of the pipeline at various stages, from girls who opt out of math classes, to qualified technical college graduates who elect to pursue non-technical careers. I have taken part in unconscious bias training; I’ve attended research talks that show how word choices (or even bulleted lists) in a job ad can either encourage or discourage women and minorities from applying. I’ve been interrupted, had my ideas ignored until they were restated by a male colleague, and I’ve been asked all manner of illegal questions during job interviews. All of this is to say that while I was well aware of the challenges and issues, I was sure that by virtue of being a woman in my new position a pool of highly qualified women and minorities would materialise with very little effort.

There are many ways that a lack of diversity can be reinforced in the hiring process. One inexpensive/quick way to build a team is to rely heavily on referrals, which often serve to reinforce the demographics of the people who are already on the team because people are likely to know other people like themselves. The wording of the job ad can scare away diversity candidates. Having too many requirements can scare away candidates who are too intimidated to apply for jobs where they may not meet 100% of the stated qualifications. There is also a tendency to incorrectly associate skills and experience that are not necessary for a role. For example, there are many successful data scientists who have PhDs in physics, however, a PhD in physics is not required to be a good data scientist.

When I had my first opening, I wrote my ad for a Senior Data Science Manager and waited for all the resumes from highly qualified women and minorities to pour in. And I waited. And waited. Meanwhile, I received many applications from overconfident standup comedians, sandwich delivery drivers, and data science students (my ad suggested minimum qualifications of a PhD or equivalent plus years of work experience). I started working with a recruiter and I was soon interviewing many math and physics PhDs who struggled to communicate clearly and did not bring additional skills to my team. I sought advice from friends, one of whom called me out for my preference for PhDs at all (I wasn’t ready to listen). I talked to colleagues at my company about recruiting strategies.

What finally motivated me to move in a different direction was attending a talk that reinforced the same tired recruiting and team building strategies that have been shown to be problematic, leading to hiring the same non-diverse workforce our industry has been hiring for decades. It finally clicked that instead of accepting the status quo and letting myself off the hook because recruiting a diverse team is hard (it is), I needed to take some bold steps. I rewrote my ad and reset the level (from senior manager) to encourage candidates of all levels to apply. This required that I be creative in envisioning how a more junior candidate could contribute to the larger team. I changed recruiters to one that was on board with my hiring objectives, and who committed to sending me resumes from more diverse candidates. Once we opened the recruiting funnel, we started to see many great candidates who had very different backgrounds. We were excited about what many of these candidates could contribute to the team.

My team and I have not “solved” broad diversity problems, but we have moved in a positive direction. While we were optimising our candidate pool to better reflect the population, we also increased the quality of our candidates. For each open position, we had more, highly qualified finalists than we had before.

About the author

Rhona TextorRhonda Textor, joined True Fit as its Head of Data Science in 2015, having pivoted to retail tech from her role in handling applied machine learning for national security at Microsoft.

Rhonda, who leads a team of 7, has been central to the management and modelling of the data that built True Fit’s Fashion Genome and its product roadmap, which supports 17,000 retail brands and processes data from 150million shoppers who are registered users.  This data is then used to help shoppers find clothes and shoes they love and keep, helping retailers close the ‘loyalty loop’ and in turn, retain customers to improve customer lifetime value.

Rhonda has been instrumental in the development of 2 new products during her tenure at True Fit, the latest of which will be launched in 2021 and will combine the data power of the Genome with visual search and outfitting capabilities, in an industry first.

WeAreTechWomen covers the latest female centric news stories from around the world, focusing on women in technology, careers and current affairs. You can find all the latest gender news here

Don’t forget, you can also follow us via our social media channels for the latest up-to-date gender news. Click to follow us on Twitter, Facebook and YouTube