Increasingly, HR departments are becoming people analytics departments. More than 70% of companies now say they consider people analytics to be a high priority, according to Harvard Business Review. Firms in nearly all industries are exploring how data can be applied to improve recruiting, hiring, training and development, performance management, and more.

At the recent Wharton People Analytics Conference, representatives from two startup software companies talked about using people analytics to achieve important corporate goals: one, to promote a more diverse workplace, and the other to boost employee happiness, productivity and retention.

Battling Unconscious Bias: Blendoor

“People suck at judging other people,” Stephanie Lampkin, founder and CEO of Blendoor, stated in a video she played during a talk at the conference. The video showed her pitching an innovative software solution to a panel of investors. “Yes, even you — judging me right now — are subject to unconscious bias,” she told the investors, citing studies that have demonstrated the prevalence of hiring bias. For instance, if a resume is sent out with the candidate’s name changed from a female one like “Emily” to a male one like “Greg,” there will be a doubling in response rate.

Blendoor aims to combat bias in the hiring process using enterprise people analytics software. The company “aggregates diverse talent from multiple sources to broaden [the] talent search,” and then applies “blind review and analytics to mitigate unconscious bias from source to hire,” according to the company’s website. The aim is to help firms successfully build and manage a diverse workforce.

Lampkin said she is passionate about fighting unconscious bias in HR. Describing herself as “a 4’11” gay black woman,” she quipped: “No one’s more sensitized to solve this problem than [me].”

Profiled in Forbes, Fortune’s “40 Under 40,” MIT Technology Review’s “Innovators Under 35,” and The Atlantic, Lampkin spoke further about how she doesn’t fit most people’s image of a tech company founder. She was born to a Washington, D.C., single mother who experienced periods of homelessness and struggled with drug addiction. But Lampkin’s luck turned when her mother moved to be near her aunt, who was studying computer science at the University of Maryland, College Park. Lampkin learned to code by age 13, was a web developer by 15, and went on to earn a BS in Management Science & Engineering from Stanford in 2006.

“My first image of a computer scientist was someone who looked like me, not a guy in a hoodie and flip-flops.”–Stephanie Lampkin

Lampkin noted that because of her aunt, “my first image of a computer scientist was someone who looked like me, not a guy in a hoodie and flip-flops. That later became really, really important in my trajectory.”

She added to her laurels by earning an MBA from MIT’s Sloan School. But in the job market, she felt marginalized and underestimated. She spent five years at Microsoft but felt she wasn’t receiving adequate career growth opportunities. And when applying for an analytical lead role at Google in New York, Lampkin said she was told, “Sorry, we don’t think you’re quite technical enough. But we’ll hang on to your resume in case a sales and marketing position opens up.” She later learned that out of 55,000 employees at the time, Google only had about 12 African-American women in technical roles.

After being rejected by Google, “I took my ‘non-technical’ self and built the first version of the [Blendoor] app,” Lampkin said.

Unconscious bias is a costly challenge for many companies, Lampkin said — it negatively impacts employer brand, attrition, quality of hire, and of course, diversity. “Diversity is a hot topic, particularly in Silicon Valley where I live,” she said. “Many of these companies have been spending a lot of money trying to solve this problem over the past five years, [but] with minimal results.” In her view, the problem stems from trying to promote diversity without first confronting unconscious bias.

Blendoor offers its clients three types of services, according to its website. First, firms can gain access to a large repository of diverse talent. Second, they can view qualified candidates in terms of job fit, with bias-inducing information hidden from view (name, photos, age and the like). Third, they can receive analytics about how their own organization is progressing in terms of eradicating bias: The software tracks how qualified candidates with different demographics perform at every stage of the HR funnel. Lampkin said the idea is to “demonstrate transparency to recruiters about their decision-making over time.”

On the subject of attracting investors to her startup, Lampkin noted that only 2.2% of all venture capital money last year went to women-founded companies, and only 0.02% to African American women. “So ultimately there are about 40 black women in the entire world ever to raise a million dollars of venture capital,” she stated. This in itself reflects unconscious bias, she noted, “but [that statistic] drives me. It validates a lot of what we’re trying to do.”

Changing Culture a Nudge at a Time: Humu

True organizational culture change is notoriously difficult to achieve. But the startup Humu promises to help companies do just that. “We’re leveraging the power of small behavioral changes to create radical organizational transformations from the inside out and the bottom up,” said Stefanie Tignor, a people scientist and analytics lead at the firm.

Humu was co-founded in 2017 by Google’s former SVP of people operations, Laszlo Bock, who is now the startup’s CEO. Bock’s original vision for Humu was “to make software that helps managers and co-workers act more human,” according to The Washington Post. (Bock is also connected with Blendoor as an angel investor, Lampkin noted during her talk.) The young company counts among its clients Sweetgreen and Teach for America.

“We’re leveraging the power of small behavioral changes to create radical organizational transformations from the inside out and the bottom up.”–Stefanie Tignor

According to Tignor, Humu takes a unique approach to people analytics. Most HR people generally measure the basic milestones of employment such as salary increases, base salary, promotions and performance ratings. They may also gauge employee perceptions — for example, through engagement surveys. But they don’t consider the larger context in which the employee works, Tignor said, and they miss key insights as a result.

Humu looks at the gap between an employee’s perceptions and the realities, and also at historical and social factors, Tignor noted. For example, “a new CEO joins; maybe their team size grows; maybe just as they missed a promotion, their social connection got promoted.” Humu is in the business of understanding the “complex, messy web” where everything is connected and interacting, Tignor said. “That’s where we get the most accurate predictions in our people data.”

Tignor shared some of the psychological intricacies captured by Humu which can affect individual behavior as well as corporate culture. For instance, if someone dislikes their manager, they are more likely to quit. That may seem like common sense, but the feelings of the people around them have an effect, too: If the rest of the team doesn’t like that manager, the dissatisfied individual is 12% more likely to stay than if they were the only person feeling that way.

You would think that having a universally abhorred manager would be the worst-case scenario, said Tignor, but actually, the person “feels more social support … than if they’re all on their own.”

In another study, Humu looked at how employees fared who disliked their manager and then changed to a different one. They found evidence that “negative baggage” from the original relationship carried over into the relationship with the new manager. “Again, that’s the power of combining the perceptual with the historical,” said Tignor. “If we were just looking at one point in time, we would miss this effect.”

In terms of creating behavior change, Humu uses “nudges,” which is messaging that Tignor describes as “a little shove to … get you to do the thing that’s most adaptive.” She asserted that these nudges help to change people’s behavior all over the organization at a “grassroots” level. An example would be a message encouraging an employee to speak up more often. Tignor noted that Humu also works to change the individual’s environment. “If we’re going to nudge one person to speak up, we also nudge others to be more effective listeners such that we’re lifting up that behavior and paving the way around them.”

“If 27% of the [firm] just looks at these nudges … that’s enough critical mass to start a cultural change throughout the organization.”–Stefanie Tignor

Humu offers its clients a diagnostic survey, and Tignor explained how nudges are used to boost the survey’s usefulness. She said that while it’s very important for managers to discuss the survey results with their team, they may be reluctant: “It can be a really scary experience, particularly if their results aren’t that great.” The software nudges the manager with suggestions about sharing the results, and makes it easier to follow through by providing a link with a pre-populated calendar invitation. In addition, a guide to holding an effective conversation is provided.

Then another nudge goes to employees, asking if the manager had the conversation with them. Employees of managers who received this nudge, said Tignor, are two and a half times more likely to have the conversation. “So these nudges really inspire action.” In addition, if the manager-directed nudges were personalized — expressing not just organizational results but specific team results — managers were 20% more likely to share the information with their people.

This single conversation about survey results, if it is held, has a big impact, Tignor said. Humu’s studies have shown that it caused employees to have 64% greater agreement that their manager was committed to change. These employees also viewed their organizations as more committed to change, by 27%.

Tignor distinguished Humu’s “nudge” approach from what she called conventional large-scale organizational change programs. With those, she said, “everyone really has to get on board and be engaged and follow through on the items. [But] with nudges, the efforts can start really small.”

How small? She identified a “tipping point” at about 27% of the organization. “If 27% of the [firm] just looks at these nudges … that’s enough critical mass to start a cultural change throughout the organization.”