The business unit was doing well, but the employees were sad. Could data offer a clue?
Back in 2015, inside Microsoft’s human resources division, a former actuary named Dawn Klinghoffer was taking on a difficult task. She was trying to figure out if the company could use data about its employees — which ones thrived, which ones quit, and the differences between those groups — to operate better.
In one case, she and her team had found that when people moved to another unit within the company, they tended to become more engaged and ultimately to become more valuable employees. Yet surveys showed that workers felt it was easier to quit Microsoft than to transfer internally.
At the time, company rules inadvertently discouraged transfers. You had to log at least 18 months in your current position before you were eligible; and to interview for a new role, you had to get permission from your manager, which had a predictably chilling effect. When Ms. Klinghoffer’s team showed Microsoft the evidence, it relaxed the rules, and transfer rates soared.
Ms. Klinghoffer was frustrated that this and other insights came mostly from looking through survey results. She was convinced she could take the analytical approach further. After all, Microsoft was one of the biggest makers of email and calendar software — programs that produce a “digital exhaust” of metadata about how employees use their time. In September 2015, she advised Microsoft on the acquisition of a Seattle start-up that could help it identify and act on the patterns in that vapor.
Called VoloMetrix, it had been founded a few years earlier by Ryan Fuller, a former management consultant. The start-up had gotten good at analyzing vast amounts of metadata from office productivity software. One of its foundational data sets, for example, was private emails sent by top Enron executives before the company’s 2001 collapse — a rich look at how an organization’s elite behave when they don’t think anyone is watching.
Together at Microsoft, Mr. Fuller and Ms. Klinghoffer set a goal of figuring out what behaviors — especially those that individual employees had control over — tended to predict and contribute to success within the company. They called their work “organizational analytics.”
They wanted to know things like: Is there an optimally productive length of the workday? Should salespeople focus on deep contact with a few clients or shallow relationships with lots of them? Ms. Klinghoffer and Mr. Fuller came up with some answers that amount to a data-driven guide to being a successful employee — not just at Microsoft, but at nearly any ambitious corporation.
One of their findings was that people who worked extremely long work weeks were not necessarily more effective than those who put in a more normal 40 to 50 hours. In particular, when managers put in lots of evening and weekend hours, their employees started matching the behavior and became less engaged in their jobs, according to surveys. Another finding was that one of the strongest predictors of success for middle managers was that they held frequent one-on-one meetings with the people who reported directly to them. Third: People who made lots of contacts across departments tended to have longer, better careers within the company. There was even an element of contagion, in that managers with broad networks passed their habits on to their employees.
Some of these nuggets are unsurprising: Communicate well, take care of your people. But their work would turn out to have a surprising usefulness in solving Mr. Ostrum’s puzzle. To understand why, it helps to know a little bit about something completely different — the modern history of baseball.
Click here to continue reading Neil Irwin’s article.