A Data-Driven Approach to Workforce Decisions

 MIT Professional Education, which provides professional education courses and lifelong learning opportunities for science, engineering, and technology professionals at all levels from around the world, is an MTLC member and a sponsor of MTLC’s Technology & Innovation community.

Among the many courses offered through MIT Professional Education is a three-day program, “Workplace Analytics, AI and Ethics”.   The lead instructor Ben Waber is recognized  worldwide as one of the leading thinkers at the intersection of management, data, workplace, and people. 

We had the opportunity to chat with Mr. Waber recently about how business leaders should use data to determine optimal work styles for employees and teams rather than applying one-size-fits-all remote, in-person or hybrid policies. Our conversation follows.

For the past several years, business leaders have been trying to figure out whether employees are most productive in a remote, in-person, or hybrid work model. But that’s the wrong question, according to Ben Waber, a visiting scientist at MIT’s Media Lab.

Rather, business leaders should be looking closely at their internal data to determine what working styles are the best fit for particular employees, teams, and projects, says Waber, who teaches an MIT Professional Education course on workplace analytics. Here, Waber discusses which workforce numbers organizations should be measuring, which metrics can be misleading, and why blanket workforce policies are often a bad idea.

Q: It’s fairly common for organizations to issue company-wide return-to-office mandates to their employees, but you’ve said this isn’t the best approach. Why not?

Waber: Choosing one policy for an entire company doesn’t make any sense, because you’re essentially saying that every team needs to work in the same way. That alone is a ridiculous idea, but even if you do find an optimum today, is that way of working going to be the optimum six months from now, or a year? No, of course not. People are hung up on deciding which model is better, but that framing is so wrong. There’s no reason we have to ask the question that way.

Q: If the current approach to return-to-office is misguided, what’s the alternative? 

Waber: A more data-driven process is usually going to yield better results. First of all, data lets you admit that your decisions are essentially tests of your hypotheses. We don’t know for sure what’s going to work, and that’s fine. If we say something is going to make the company more productive, that’s a hypothesis. When you admit that, then you can test it. Pre-pandemic, there was more of a belief that executives know what they’re talking about when they say that something will make the company more productive. And now it’s clear that often, the emperor has no clothes.

Q: Are there clear trends in workforce data that can help business leaders make these decisions?   

Waber: When people work remotely, they spend much more time communicating with their strong ties. These are the people you spend an hour or more with in one-on-one conversations each week—the people on your team. On the other hand, we see a very significant reduction in weak ties. These are people you spend the equivalent of five to 15 minutes with each week. Those contacts decline by more than 20 percent.

For short-term project delivery, those weak ties may not be very useful. However, when it comes to things like milestone attainment on large projects, and when it comes to innovation, those weak ties are by far the most predictive factor, because people are being exposed to different kinds of information.

Q: What about company-specific data? What are some examples of internal data points that leaders can use to make better workforce decisions? 

Waber: The data coming out of things like email and chat—not at the individual level, but the patterns of communication and collaboration between people—those are the biggest predictor of organizational success at the macro level. And everyone has that data. There are also sensor systems that can help you track collaboration and communication in the office.

Leaders want to know whether workforce decisions are going to make the company more productive. But you can’t answer that without contextual metrics about performance. You need key performance indicators for that person, role, or team. Most companies haven’t developed those. Instead, they look at employee reviews, but often reviews come down to how much the boss likes a person. That’s not performance.

Q: Are there data points that are less helpful?  

Waber: If you’re trying to make sure employees are working by bringing them into the office, that’s largely pointless. Data shows that people work at least as much when they’re working remotely. Even their outputs often play out over much longer timescales. So, what we want to measure are their behaviors.

Q: How can leaders develop those key performance indicators, when they’re not sure which data points will lead to the results they’re seeking? 

Waber: Distinguishing ‘okay’ from ‘excellent’ is basically impossible without those metrics. However, distinguishing ‘awful’ from everything else is pretty easy. For example, if over a whole month a manager is spending less than 2 percent of their time with their team, that’s a problem. If there’s an obviously dangerous level of weak ties, or a dangerous level of siloing, that’s a problem. If a large portion of your company is working more than 60 hours a week, that’s a problem. If people don’t have blocks of uninterrupted time to do their creative work, that’s a problem.

If your workforce interventions don’t improve the ‘awful’ areas, then they’re highly unlikely to have the desired effect over time.

Source: MassTLC