You know the expression, “It’s all about the people”? One local startup has figured out a way to quantify that approach, using sensors and software in corporate settings.
The startup is called Humanyze—it was formerly known as Sociometric Solutions—and it spun out of the MIT Media Lab in 2011. Since then, the eight-person company has been heads-down developing technology to help businesses improve their performance by understanding how their employees behave on a daily basis.
The key? Gathering and analyzing data on how employees talk to customers, who talks to whom within companies, what times of day people send e-mail, make phone calls, go on break, and so on. If it all sounds a little Big Brother-ish, well, Humanyze has thought carefully about those privacy concerns (more on that below).
Now the Boston company has raised $1 million in venture funding led by Romulus Capital, with participation from Boston Seed Capital and Dunnhumby Ventures. That’s not a lot of money, especially in this day and age. But Humanyze CEO and co-founder Ben Waber says the company has “real customers and significant amounts of revenue,” so it sounds like the deal is mostly to get the right outside investors in place and make some hires. His top priority at the moment is to bring on a chief product officer.
The field of data-driven human resources is starting to see a lot of interest—and hype—whether you call it people analytics, reality mining, or “Moneyball” for business. The trend is being driven in part by mobile and wearable technologies, as well as the rise of big-data analytics. Yet, as Waber puts it, it’s surprising “how data-driven companies are about their business, but it never includes their people.”
Instead, he says, companies boil down their people-management approach to things like org charts, consultants with PowerPoint decks, and “some executive saying, ‘I read a case study in Harvard Business Review, so let’s do that.’”
“Just going with your gut when it comes to people is crazy,” Waber says.
Humanyze’s data-collection system consists of “sensor badges” (pictured above) for employees—these include a microphone to capture things like tone of voice and when and how quickly people speak, and a location/proximity sensor to tell where people are when they interact face-to-face. The company’s software can then mash that up with digital data from e-mail and phone communications to give a fuller picture of how people communicate with each other.
What’s more, the software can correlate these behaviors with key performance indicators so a company can see what a top performer is doing differently from the group, how often the best sales teams talk to engineers, and so forth—all in a neatly visualized dashboard for management or for individuals (pictured below).
At this point, many employees would worry—do I really want my boss tracking what I do all day and using the data to drive performance evaluations?
That’s not the point of the analysis, Waber says. “We don’t record what you say. Your boss doesn’t get to look at your data. And we don’t count how many times you go to the bathroom.”
Indeed, the company’s co-founders include Alex “Sandy” Pentland, an MIT Media Lab professor whose recent work includes thought leadership on privacy and data ownership, as well as research into human interactions in organizations. So Humanyze is acutely aware of issues around protecting individual employees’ privacy. The company says it only provides behavioral data in aggregate and/or anonymized form.
“Privacy is so central to what we do,” Waber says.
Even so, does the data-driven approach really work? Waber points to successful pilots with big companies such as Bank of America, Deloitte, and a top oil company. Fortune 500 companies, he says, want to “deploy our analytics at scale to answer big questions and problems around people. They’re looking to dramatically improve their people and workplace.”
Bank of America, for example, used the Humanyze system in its call centers to try to boost performance and lower its employee turnover rate. The bank noticed that some call-center teams were much more efficient and stable than others, but it couldn’t figure out why, Waber says. Humanyze looked at a few metrics, such as average length of calls, interactions on calls, and when employees went on breaks, and tied those into performance metrics.
It turned out the most important factor was how employees interacted with their co-workers, not customers. When team members had overlapping lunch breaks and talked to each other, their stress was lower (as measured by tone of voice), job turnover was lower, and they completed their calls faster.
So the bank made a management change and tested it over several months—it gave half the teams breaks at the same time and compared the results. It found the turnover rate fell from 40 percent to 12 percent, and the more cohesive teams completed their calls 23 percent more quickly—which is “worth tens of millions of dollars” to Bank of America, Waber says.
With Deloitte, meanwhile, the system was used to analyze how the geographic distribution of the company’s workforce—on a city, neighborhood, and even a floor-by-floor level—affects collaborative behaviors like talking face-to-face versus sending e-mails.
Proving that understanding these kinds of metrics can consistently yield concrete suggestions that boost companies’ bottom line will be key for Humanyze. For now, Waber says, it’s most important to get things right with the first batch of customers. “The next six months are critical,” he says.
“Everyone’s eventually going to have to do this,” Waber says. “Look at where Amazon was 20 years ago, doing A/B testing—now everyone has to do it. People analytics are at the same stage.” (See example of a manager dashboard below.)
Waber continues: “Our customers are investing in something that’s hard to do today. We are saying no to customers who are not willing to change their company. If you do change, you’re going to destroy your competitors. If you don’t, you won’t be around in 10 years, and we’ll work with your competitors.”
By posting a comment, you agree to our terms and conditions.