When Bank of America discovered a disparity in its call center productivity, it tried a novel problem-solving approach: outfitting employees with personal sensors. Developed by Ben Waber, MIT Media Lab alum and CEO of Sociometric Solutions, the wearable sensors are packed into an ID badge and measure social interactions in the workplace — such as how much employees move and converse throughout the day.
The data Bank of America gleaned from the study led to a surprising insight: The more employees interacted with one another, the better their productivity. It's just one example of how the nascent field of people analytics can allow managers to use facts — not gut feelings — to make decisions.
Here, Waber explains what companies can learn from social data, the future of personal sensors in the workplace — and why simple hacks to workplace culture can have a huge impact.
Data is the topic du jour in HR, but data-driven personnel decisions are still relatively uncharted territory. Why?
Businesses are so data-driven about customers. [But] when it comes to talent, they tend to forget about the data. The conversation becomes something along the lines of, “I read a Harvard Business Review case study about a company that did things a certain way. That feels right. Let’s do it that way!” It’s crazy.
Why are employee interactions worth measuring?
Until now, companies have used two methods to assess what’s going on with their employees: conducting surveys and having consultants follow employees around with a clipboard. Neither of those is very effective. Even if companies believe the metrics, it doesn’t tell them why — or how to change anything. We’re trying to answer the “why” questions. Companies can get some of that by hiring a consultant to follow people around, but it’s subjective and it doesn’t scale.
You call your wearable sensors "next-generation company ID badges." What can they tell businesses about employee happiness?
Today, ID badges have RFID tags, which are actually sensors that can tell a person’s location. But that doesn’t tell [you] how people are collaborating. We’ve added additional sensors to the ID badge to measure that: Two microphones do voice analysis (how loud or quickly somebody talks, the tone of their voice, how much they interrupt whomever is speaking) and an accelerometer measures how active people are in general. It turns out that’s a really good measure of how happy and productive people are.
That's a lot of personal data. Is that a concern?
The microphones don't record what people say. [And] we don’t give any of the individual data to companies. The company gets aggregated statistics to see what the happiest people do. The employee gets to see raw data about how they stack up against the average.
Where have you seen this social data have the biggest impact?
It’s really about making collaboration effective. If you sit next to somebody, you spend a lot of time talking to them, and if you put someone across town or on a different floor, they won’t talk very much. We should be able to show companies hard numbers about how that will affect outcomes like retention, performance and sales. Fortune 100 companies are using these metrics to, for example, design their headquarters or change how teams are structured.
The Bank of America example is fascinating. Can you tell us more about what you discovered?
Bank of America has call centers across the country. The organizational structures, training and employee demographics were similar, but the centers performed differently. The company had a sense that culture had something to do with those differences, but they needed to measure it.
We put badges on different teams in the call centers. At first, we tried to figure out what factors predict an employee's productivity. We thought it would probably be how employees talk on the phone — people who talk in a certain way would complete calls more quickly and be more effective. That wasn’t the case. We saw that people who had a cohesive network and talked to each other completed calls in half the time as those with the least cohesive network.
We then looked at when group interactions happened. It turned out that 80 percent of interactions happened when people’s lunch breaks would overlap by about 15 minutes. Bank of America used this to test a new break structure. For half of the teams, the company gave employees breaks at the same time. For the other half, they didn’t change anything. The groups where people took breaks at the same time were not only more cohesive, but their stress went down significantly. We measured it using surveys and the badges to detect changes in tone of voice. Both showed the same results. They completed calls 23 percent more quickly, and turnover went down by 28 percent. It was a small change that cost Bank of America nothing, but it had a massive impact.
What’s the takeaway for other businesses?
If you can find the social levers that people are responsive to — and if you can act on them in the right way — you’ll get good results. It’s not about looking for complicated new systems. It’s about figuring out ways for people to collaborate better.
It’s early days for people analytics. What's next?
I’m excited about two trends in particular. In about 10 years, every major company will have a people analytics division. Whether that division completely subsumes HR or not, I don’t know, but successful companies will have them. The companies that don’t will cease to exist.
The other trend is the proliferation of data about what people do. Smart watches and Google Glass are niche products, but companies like Rest Devices and Reebok are integrating electronics into clothing. That’s the future. People will put on a shirt and generate data. It will make it so much easier to get rich data about what people do while at work. But it’s critical that we establish regulations about what companies can — and can’t — do with that data.
Photo credits: Sociometric Solutions and Shutterstock