Microsoft had discovered some unsettling news. An entire group of its top-performing employees was miserable.

Compared with other groups at Microsoft, the 700 employees working in its Surface and Xbox device development division were scoring disastrously low on work-life balance surveys. To avoid losing hard-to-replace talent, Microsoft needed to figure out why. What made this team more miserable than others at the company?

In The New York Times' Upshot section, writer Neil Irwin explained how Microsoft went about it. The company's organizational analytics team is dedicated to studying data about how employees work. They're all about finding data-backed ways to make employees more engaged and productive.

Learning how employees spent their time

Instead of relying on self-reported surveys, Microsoft looked at how employees spent their time for clues. The source of this data was employees' own emails and calendars.

Microsoft compared metadata collected from their inboxes and calendars. They compared it to the same type of data gathered from other groups at Microsoft who had scored higher on the work-life balance surveys. This way, the analytics team could see if happier employees spent their time any differently than the disgruntled ones did.

Ruling out likely culprits

Were these particular employees more miserable because they were working longer hours? Was it because they had too many requests flying at them during after-work hours?

In short, no. Microsoft's data crunchers weren't able to draw either conclusion. It was true that the employees worked hard and long hours. But happier employees from other teams worked just as long and sent emails at all times of day.

"Gut instincts about overwork just weren't supported by the numbers," wrote Neil Irwin for The New York Times. "People who had taken jobs requiring that sort of commitment seemed to accept these things as part of the deal."

An aha! moment in meeting invites

Microsoft's data analysts kept digging. They uncovered that employees were spending a high percentage of their time in meetings each week -- an average of 27 hours. But it was about the same amount that other groups did.

There was one key difference, though. Meetings in the Surface and Xbox division tended to be huge, with 10 to 20 employees per meeting. The meetings were too big. This left little time for employees to do focused or creative work. That's why they were logging extra late nights and weekends. It was the only time they had left to do the real work they needed to do.

"The issue was that their managers were clogging their schedules with overcrowded meetings, reducing available hours for tasks that rewarded more focused concentration -- thinking deeply about trying to solve a problem," wrote Irwin.

Protecting employees' time for more focused work

To address the issue, Microsoft encouraged managers in its Surface and Xbox division to do a few things differently:

  • Be thoughtful about how many large meetings they scheduled and require their direct reports only attend the most essential ones

  • Encourage their employees to block out time on their calendars for concentrated work

  • Check in weekly to see how much time they had spent sending emails during off-hours or the weekend

Over time, the work-life balance numbers improved. Employee retention numbers were stable. ​Workers were able to spend more time getting work done instead of squeezing in work between meetings.