Illustration by Anja Sušanj

Because women continue to bear most of the burdens of caregiving, many have benefitted from the flexibility of hybrid and remote work, which has allowed them to remain part of the full-time workforce while spending more time at home. In theory, this is good news, because full-time employees tend to be promoted more often and let go less often — but in practice that’s not what’s happening. Why? Because when women take advantage of hybrid and remote work options today, they become subject to biases that creep into the performance-review process and cause them to be judged unfairly.

I witnessed this problem firsthand during a consulting project I did recently for a global company with hundreds of thousands of employees worldwide. During the project, which took place as the company (which I’ll call HAPI  Consulting, to preserve its anonymity) was in the midst of a large reorganization, I was reporting to the CEO, his executive, and the board. With their support, I gathered and analyzed data on performance reviews given in a hybrid work environment (pre- and post-pandemic), and in that context I identified three specific kinds of bias that were hurting women and their career prospects. I’ll explore them in turn below.

Experience bias

The first bias I observed was what we call experience bias, one aspect of which is overvaluing tasks that are easy to define.

This form of bias, I found, was prevalent in performance reviews at HAPI Consulting — where, because of the reorganization, many employees had been forced to put their routine work on hold. During that period, reviewers had judged employees instead on how they had chosen to refocus their energies. Men, on the whole, had chosen work that was much easier to recognize. They spent more time on highly visible external tasks, such as speaking at industry conferences and giving interviews to journalists, bloggers, and podcasters. These tasks were straightforward to review when it came to short-term measures of success. Meanwhile, women spent more time on much less visible internal tasks, such as boosting team cohesion to help get projects back on track and providing psychologically safe spaces for team members to ask questions and process elements of the reorganization that were taking place at the company. These tasks had longer timelines and were harder to evaluate in the short term.

These choices lent themselves to experience bias. It’s easy to understand, for example, how somebody who has presented at a major conference (probably a man) might have immediately provided a measurable benefit to the company. In fact, men were three times more likely than women to speak at outside events. But how easy is it to measure and evaluate the work that somebody (probably a woman) has begun to do in a long-term effort to repair team dynamics or prevent deep-rooted problems that might ultimately affect hundreds or thousands of employees? Needless to say, given the nature of experience bias, reviewers at HAPI Consulting praised the men more regularly in their performance reviews for having succeeded in their work. More specifically, men who reviewed other men assessed them 12% higher than women, on average, as compared to when women reviewed both men and other women.

With the support of leadership, I tested several interventions designed to help reviewers at the company correct this bias. While some interventions failed, two had a statistically significant chance (with a 95% confidence interval or above) of reducing biases at this company.

The first was for managers to help employees set individualized goals with tailored metrics and then ask them to track what they had done every week — and to do it on a virtual whiteboard that the rest of the team and the directors up the chain could see. In this way, people who had contributed to less-easily-defined outcomes, such as team cohesion, could still point to a record of their efforts over time for all to see. This whiteboard could record important work even if it wasn’t easily described in terms of traditional goals and metrics. This was helpful come performance review time. Even if employees had failed to reach the goals laid out for them in their last performance review, there was still a record of all their efforts and achievements, which, according to the post-survey results, helped reviewers better grasp the value of the more ambiguous, internal work that (mostly) women were performing.

The second was designed to help employees become better at describing their own efforts. At the beginning of the week, for each project they were working on, employees were asked to provide two different estimates on their virtual whiteboard: how long each task would take to accomplish and its level of difficulty (high, medium, or low). For example, an employee might identify the three projects they’re working on for the week and allocate the amount of time they plan to spend on each, based on the type of task and estimated difficulty. At the end of each week, employees were asked to revisit their estimates, note how their time was actually allocated and how difficult the task ended up being, and reflect on the reasons for the differences between their forecasts and “actuals.” This practice allowed employees to look back on their recorded predictions, reflect on what the record showed in comparison to their perceptions, and fine-tune their ability to paint an accurate picture of their efforts.

Initially, the data showed that women were more accurate in their estimations of the time a task would take, although they often overestimated the difficulty of tasks. Meanwhile, men underestimated both the time a task would take and its difficulty. Following the intervention, the likelihood of women accurately identifying the difficulty of the task increased by a factor of 28, while men were 12.5 times more likely to do so. Additionally, men became 7 times more likely to accurately estimate the time required for a task, whereas there was no significant difference in this aspect for women.

Proximity bias

An other form of bias I observed was proximity bias, or the tendency to think that people who are in your physical orbit do the most important work. A hybrid work environment exacerbates this bias and was penalizing women at HAPI Consulting, where men were more likely than women to come into the office. Reviewers tended to express skepticism, or at least confusion, about what women did when they worked remotely, whereas they tended to give men the benefit of the doubt and assume that they were more focused and productive.

Even when women came into the office as often as men, reviewers assigned them achievement levels that were a statistically significant 3% lower than those they assigned to the men. And when women came into the office more than men, reviewers seemed to believe that they were doing so only because they were looking out for themselves and wanted to “have a seat at the table” to defend their position in the reorganization of the company. Conversely, when men came into the office an above-average amount, they were seen as strong leaders.

To address this problem of proximity bias, I administered a survey to employees to gauge their preferences among several choices: more online gatherings, an “anchor day” when everybody was expected to be in the office, or more offsite team-building opportunities. The employees’ preference was that the company create an anchor day, which it did. This intervention led to a statistically significant drop in proximity bias that was observable in the next performance-review cycle. In addition to having employees post their current projects on a virtual whiteboard each week, teams were also asked to hold weekly meetings during which they would discuss three achievements from the previous week, their goals for the current week, and three challenges they were facing in trying to achieve them. By drawing more equal attention to the work that everybody was doing, no matter where they did it, these meetings helped the company resist proximity bias.

In-group/out-group bias

The third form of bias that was creeping into remote-work performance evaluations was in-group/out-group bias, or giving preferential treatment to people you feel belong to a group you identify with. Not surprisingly, employees who were working remotely, and particularly those who had moved away from the office and across state lines, were increasingly seen as members of the out-group — and they had annual performance review scores that were lower by an average of 20%. Benefits would accrue to those in the in-group, including vertical promotions, horizontal promotions (providing employees with new opportunities to learn new skills and take on more prized tasks that would position them well for vertical promotions), and bonuses, such as equity shares in the company. Both men and women were subject to this form of bias, but because men tend to be seen more readily as leaders, they were more easily able to join the in-group than women.

To increase a sense of belonging, the company launched a campaign called “I am HAPI Consulting, We are HAPI Consulting.” One element of the campaign asked employees to answer simple, sometimes silly questions so that teams could come to know the commonalities among them. For example, when providing updates at meetings, employees might also have to tack on the answer to the question, “Dog or cat?” or “What is on your bucket list this year?”

Questions like these provided some memorable moments that everyone could share, no matter whether they worked in the office or remotely, and they allowed more people to experience in-group moments. Strikingly, this practice led to a 14% increase in feelings of belonging to the in-group, and a 17% increase in employees’ perceptions of which other employees belong to the in-group. At performance-review time, reviewers were also asked to recall one thing they had in common with the person they were about to review. We found that this simple intervention, which meant that reviews began from a point of commonality, decreased the gender discrimination in the company’s performance reviews by several percentage points.

Men and women often experience the workplace differently, whether in-person or virtually. To include and support the best talent for your company in a dynamic environment, it’s important to take a careful look at the data, be alert to the three kinds of biases discussed in this article, and identify practices that are leading you to overvalue the efforts of certain groups. The insights you gain from doing so will not only help you treat women more fairly but also will lead to fewer predictable surprises, more productive work hours, and increased efficiency in smoothly leading teams — and your company — to success.