How Software Can Help Move The Needle On Diversity

How Software Can Help Move The Needle On Diversity

We asked several participants in the upcoming Techonomy 2014 conference to write an article for us on what they are passionate about right now.

Even after spending my whole career in tech, I’m still taken aback by each new diversity stat that underscores the systemic and trenchant bias in our industry. One of my recent favorites: 76 percent of feedback given to women included negative criticism of a personality trait, while only 2 percent of men received such feedback. Or this from Fenwick and West: 45 percent of tech companies in Silicon Valley don’t have a single female at the executive level, compared with only 16 percent of the rest of S&P companies.

If you are working in Silicon Valley today, or as a professional in any region of the country, you know these numbers—you’ve heard the stats, you see it in your own engineering teams, you’ve read “Lean In” or a take-down of it.

In 2012 I started Unitive, a software company that empowers HR teams with technology to make hiring and promotion processes more bias-free. At the outset, I thought the hardest work would be convincing potential customers they had a diversity problem that needed solving. That was before Anne-Marie Slaughter wrote “Why Women Still Can’t Have It All” and renewed the conversation about what life is like for working women; it was before the “Lean In” movement turned the focus of that conversation to Silicon Valley; and it was before gender discrimination in the news snowballed with a few embarrassing moments for white males in tech; and, of course, it was before this year’s release of diversity data by Facebook, Google, and others. Suddenly diversity outcomes and how Silicon Valley is falling short is a favorite issue to pay lip service to.

Even better—today’s thought leaders are reaching consensus that diverse teams are smarter, return a stronger bottom line, and are more innovative problem solvers. A diverse team delivers a better product. As Robin Hauser Reynolds points out, the earliest airbags were designed by an all-male team with an exclusively male form in mind—and were fatal to women and children.

We’ve come a long way in giving this issue the attention it deserves, but change is elusive. One of the greatest challenges HR teams and tech giants face today as they try to achieve diversity is that they lack technology to do so.

Right now, the go-to strategy for improving diversity in a company is unconscious bias training. But we don’t have much research to show that such training is working. Here is a statement from a study by Harvard and Yale researchers on the merits of interventions for reducing prejudice: “We conclude that the causal effects of many widespread prejudice-reduction interventions, such as workplace diversity training and media campaigns, remain unknown.”

I am thrilled that organizations are willing to spend resources on something like removing unconscious bias. But I’m concerned that our leading innovators are using an outmoded approach to tackle such a pervasive issue. I’ve always solved problems by building technology, and I’ve tackled big problems in the past like online fraud and website security.

And so Unitive was born. There are many places within the hiring and promotion pipeline that unconscious bias can be mitigated, and through software we can pinpoint those biased decisions in the moment they are happening. That’s something unconscious bias training will never be able to do. Our products include many tools that help employers, for example, write more inclusive job descriptions, review candidates more anonymously, and exercise consistency and accountability during interviews.

There are three ways our software intervenes when bias comes into play. First, we restructure familiar processes. One example: When hiring managers review resumes, we know that they measure one attribute differently based on the rest of a candidate’s attributes, i.e., a candidate’s stellar experience may start to appear less impressive if she went to a non-Ivy League institution. With our software, hiring managers are forced to establish the criteria most important to them (i.e., experience versus where candidates went to school) before they look at a resume. Then, instead of reviewing entire resumes at once, they review all the resumes in sections: reviewing all candidates’ education, experience, and skillsets and scoring those components independent of the whole—and, of course, without consideration for whether the candidate’s name is James or Jamal.

Secondly, our software shows people where and when they might be biased. There are many “teachable moments” in the hiring and promotion processes. An example is when someone writes a job description that includes 20 requirements. Research shows that women feel they must meet all requirements while men do not. As a recruiter or hiring manager adds requirements, we can remind and reinforce that the more requirements they include in a job description, the fewer women will apply for the job. They can still choose to add a large number of requirements, but at least they are aware of the possible impact at the moment they are making the decision about the job description.

Finally, we can record when someone has disregarded the recommendations for making things unbiased. In the previous example, if a recruiter or hiring manager includes 20 requirements, we can track whether it has an impact on the candidates considered for the job. We are also able to see which of our strategies are most effective and which features need further iteration to achieve the most impactful outcomes.

The chief talent officers and other people we speak with about our software are extremely excited to see a totally new approach to the diversity problem. Training and other exercises to reduce bias and increase diversity aren’t moving the needle in the way people had hoped. The Unitive platform has the power to unlock resources our nation’s innovators need to create the most diverse—and therefore most innovative—teams.

Laura Mather will give a short presentation on diversity and moderate a session on data security at next week’s Techonomy 2014 conference. At September’s Techonomy Detroit conference, Mather spoke on a panel about inclusiveness in the tech industry.

Article originally published on


No Comments

Sorry, the comment form is closed at this time.