We spend the majority of our waking hours at work. If you’re lucky, those hours are in a workplace that energizes and empowers you. But for many, it’s in a constraining environment, and work is a slog. This is sometimes a design problem, and by using data about what makes people effective at work, we can improve the workplace to dramatically boost performance, job satisfaction, and more.

Data, with an overemphasis on cost, often already influences workplace design. But instead of just looking to reduce costs, workplaces should find ways to create value by changing how people collaborate. One way we’re researching that is through the use of sensor-filled, wearable badges, which allow us to examine how interaction patterns and space usage relate to performance and job satisfaction. The badges measure who employees interact with and for how long, which can then be used to map workplace social networks.

Our research has revealed many things, but in one particularly striking example, we found that increasing the size of lunch tables in an online travel company’s cafeteria significantly improved performance by over 10%, as measured by peer reviews. We found that the most productive programmers were eating lunch in very large groups. The exact people in those groups changed day to day, but group size remained constant: around 12 people. The least productive people were eating in smaller groups of around four. Again, the exact people in the groups changed, but the size remained constant.

What was happening? A quick trip to the cafeteria revealed the answer. Tables by one set of doors had 4 seats while tables by the other set of doors had 12 seats. Even when controlling for demographic factors and other potential confounders, the table size effect remained. By eating in larger groups, these higher-performing lunch goers were building larger networks within the organization and research has shown that social network strength can influence performance. A simple decision to make all the lunch tables bigger led to performance improvements of over 10%.

We also gathered lots of insights around team “cohesiveness,” as measured by the amount of time teammates spent talking with one another. We found that more cohesive teams were generally less stressed than other teams and significantly less stressed after layoffs. We also found that giving out free beer on Fridays didn’t boost cohesion. We’re not against free beer, just don’t do it in the name of team cohesion.

Data not only validated this effect, it showed the way forward. Moving beyond lunch tables, we can apply such analytics to seating, office layout, and more. While workplace design is typically given short shrift by the C-suite, people analytics helps bring it to the forefront.

Ben Waber is President, CEO, and co-founder of Humanyze, a people analytics company that uses combine wearable sensors and digital data to deliver people analytics and insights to customers.