Laura McGee has spent her entire career thinking about diversity and business. At one point, she helped lead the Trump-Trudeau Council for Advancement of Women, working with the prime minister and president to build a plan to grow the North American economy through diversity. During that time, she kept hearing from CEOs that they cared about diversity and wanted to improve, but that they had “no data and no metrics.”
That was when she decided to build Diversio: a platform that makes data collection, as well as acting on it, “super simple.”
The company just announced a $6.5 million Series A from First Round Capital, Golden Ventures and Chandaria Family Holdings. Protocol spoke to McGee, Diversio’s founder and CEO, about how it’s using data and AI to help companies get closer to its DEI goals.
Here’s how it works: The platform starts by collecting data through an anonymous four-minute pulse survey that goes out to all employees. This survey collects information on race, ethnicity, sexual orientation and disability as well as mental health. It also asks participants about everything from whether they feel heard by their teams to if they’ve experienced workplace harassment.
Diversio then combines that survey data with data around recruiting and hiring, as well as representation in different roles and teams. Questions that this data can answer include, “How do different departments do? How are different levels represented? How is talent advancing from a racial and gender lens?” As McGee pointed out to Protocol, often companies are good at bringing diverse employees in, but they are bad at keeping and promoting them.
What’s unique about Diversio is how it benchmarks this data against the company’s dominant group, which more often than not for tech company leadership is heterosexual white men without a disability. It then analyzes the data for significant differences in experience between that group and nondominant groups to help the organization locate its biases and specific DEI issues, both as a whole and within specific departments and roles.
For example, one question might be, “Do you feel that your opinion is valued by your team?” The average response rate for people of color might be 6.5 out of ten and 8.0 for the dominant group. That 1.5 delta helps pinpoint where the inclusion issues might lie.
Where artificial intelligence comes in is the platform’s free-text analysis. Diversio has trained its natural language processing algorithms on free-text data to identify what it calls “inclusion pain points,” or specific drivers of behavior or working conditions that influence the scores. One example might be that, in employee survey responses, many people talk about time off being a problem, with a significant difference between groups in that perception.
This startup is using data to improve employee diversity.Image: Diversio
The final piece of Diversio’s offerings is its recommendation engine, which spits out potential solutions tailored to the company’s specific DEI issues. According to McGee, the database currently includes about 1,600 recommendations, which Diversio gathered by scraping “every program and policy intervention we could find globally for diversity and inclusion.” One example of a solution: the 40% rule, in which job candidates must be at least 40% women, to make sure there’s equitable opportunity.
The idea is that companies can pick and choose the solutions that sound right for them, implement them and re-survey their employees as often as quarterly or once a year.
According to McGee, some of the most common issues experienced by the 400-or-so companies currently on Diversio include mentorship and sponsorship for non-dominant groups, as well as underinvesting in neurodiverse employees and employees with disabilities.
In terms of solutions, she’s observed that one of the most effective has been tying a portion of executive compensation to the company’s inclusion scores. Tying compensation to inclusion “really motivates action,” McGee said.
As to what the influx of cash is going towards, McGee said she sees it being allocated 50% towards growth by way of sales and marketing and 50% towards investing in the tech, from improving the NLP to refining the recommendation engine.Interestingly, Diversio itself has used Diversio the platform, another example of tech dogfooding in action. McGee said she didn’t fully understand how diverse the company was until the survey results came back and so many people self-identified as having a mental health disorder, for example. “Candidly, I didn’t fully appreciate the value of the product until I saw the results,” McGee said.