Dr. Shelley Correll thinks about gender equality for a living. As a Stanford professor of sociology and director of the VMware Women's Leadership Innovation Lab, she conducts on-the-ground academic research on how to make organizations more fair and inclusive. I was fortunate enough to work with Dr. Correll for several years at Stanford and caught up with her recently to get her take on how to create real, lasting change.
TT: Your work in the field of diversity has been rooted in the idea of small wins. Can you tell us a little bit about what that is and how it applies to diversity?
SC: I’ve been doing this for 20 years now, and when I started I was—and still am—really interested in understanding the barriers to gender equality in the workplace. Any profession you look at, the higher up you go, the fewer women. People in the corporate sector are doing all kinds of things to try to move the needle, but the pace of change is just so slow. This got me thinking that we needed a different approach to thinking about change.
Social scientists often think the change that is needed is so big that we can't possibly make progress. And people in the corporate sector think, 'We're going to do this huge program that has 10 arms to it, and all these people, and that will fix things.’ But when the people who are in charge of implementing the program get started, it's really hard. It’s as if you were told tomorrow you needed to solve world hunger. You would want to, but what would you actually do?
We know from looking across all kinds of change efforts, the one thing that helps people make progress is to take these large, big, hairy problems and break them into smaller steps. The beauty of small wins is, if you design a small win, people see that they have produced change. Anytime you create a goal for yourself and you achieve the goal, you start to believe I'm a person who can make a difference in my environment. That inspires further change.
TT: How do you create a structure for these changes?
SC: Our model is this: We go into organizations, we work with managers or leaders in those organizations to identify a change target. Then we design a tool or intervention to make progress. For example, we went into an organization where they noticed that the percentage of women in higher-level positions was lower than the percentage of the women a couple of levels down. They asked why and wondered if there might be bias against women in the performance assessment process. So we observed their calibration meetings [where managers come together to talk over people’s evaluations and make sure ratings are fair and match the target distribution.]
Like a lot of companies, at this company there's a limited number of people who can get the top rating. Because people with top ratings expect to get promoted, they get bigger bonuses, they get bigger raises. There's all kinds of reasons you can't just put everybody in the top box. What calibration therefore becomes is a negotiation about who gets knocked out of the top box.
One of the things we observed is that women were more likely to get downgraded from the top box. A manager might say, 'Well, I know Joe's not a top performer, but if we give him a lower score, he'll get mad and leave.’ We never heard them say that about a woman! We also noticed much more criticism of women's personalities. 'They're too assertive, they're off-putting to people.’ And the men were being described with more standout adjectives: 'This guy’s visionary or genius.’
When the organization saw that data, they decided this was a place where they could make change. We worked with managers to design a scorecard that they could use for evaluating each of their employees. This took a lot of time because we were going back to their company values and asking: How do you measure people against those values? What are the criteria that you're using? Are these criteria good criteria? Are these criteria being applied equally to all employees?
Six months later, the managers show up to the next calibration meeting with a scorecard filled out for every employee. And the discussions in the room were just qualitatively different. Employees were being talked about in terms of the criteria on the spreadsheet. If somebody would say something about an employee’s personality, one of the other people in the room said, ‘I don't see that as a criterion here,’ and they would drop it. The downgrading of women from the top box, that gender gap was eliminated as well. So this is an example of a small win.
Then some managers said, ‘I'm going to use this scorecard when we're hiring people.’ Then they started looking at the wording of their job ads and wondering whether maybe their job ads were lowering the number of women that were applying for their jobs. So one small win led them on their own to the next change target, which created another small win and so on.
This company, by the way, was GoDaddy. They were so happy with the success, they took away the non-disclosure agreement that we originally worked under. The end result with GoDaddy, which had been known for its sexist Super Bowl ads, is that it soon after became recognized as one of the best workplaces for women in tech.
If you're not fixing the culture, then you're not really fixing the problem.
TT: Whether it's from the public or from activist investors, there's a lot of pressure on companies right now to move fast. Also pressure internally, as people recognize that they need to make change. Does this pressure to move fast ever result in people misinterpreting or misusing some of the social science research?
SC: Yes, we see that. There's pressure to go fast and there's also pressure to do the newest thing. So [there is pressure] to not stick with something that you're working on that might eventually lead to change but just to latch onto the newest trend. For example, we've worked a lot with companies on the language of job ads, determining whether the wording of your job ad is more appealing to men rather than women, or white people rather than people of color. There are applications that will tell you This [language in a job ad] is masculine, this word is a problem, let's take this word out. But, by itself, removing one word from a job ad is not going to solve the problem, right?
When we identify a problem with wording in job ads, what that's really telling us is that there's a problem in the way you're thinking about who is a good employee here. That's going to affect not just the job ad language but also how you interview and how you treat people once they arrive. So you may not be using the phrase “ninja coder” in your ad anymore, but when they show up, you're expecting them to be a ninja coder. People pick up on that.
TT: Or they don't pick up on it, and they get hired, and then it's even worse.
SC: Right. Then they leave right away. The language in the job ad is reflecting something about the culture. If you're not fixing the culture, then you're not really fixing the problem.
TT: How does that fit with this theory of small wins because, on the one hand, a small win is a relatively small, discrete success. On the other hand, what we're really aiming at here is cultural change.
SC: Right. For example, the GoDaddy case, a big part of the small win was creating the scorecard where they went back to their core values and asked what criteria they could use to assess whether people were living up to those values. Once they listed the criteria, they had to say how they were going to measure those criteria. You can't answer these questions without thinking deeply about what you value at work.
There was one company where one of their values was that leaders needed to be responsive to the people that reported to them, which is great! But their measurement was how quickly people responded to email. What they found is that women weren't responding as quickly to email during the dinner hour. We asked them, ‘Is the email you receive from people who are responding during the dinner hour helpful to you as a manager?’ They investigated and said no. They realized they wanted their leaders to be responsive, but the [e-mail response time] wasn't a good measure.
Sometimes we create a scorecard at one company based on their values, and then I go talk to another company, and they ask, ‘Well, do you have an example of a scorecard we can just use?’ My response to them is this: Unless your values are exactly the same as that first company, the scorecard is not going to work for you. People are going to have a scorecard, but they're going to be confused by it because it's not reflecting the company values. When this happens, they will go back to whatever process they were using before.