Humans and Technology

Technology for workplaces that work: Humanyze’s Ben Waber

What can new kinds of sensor data tell us about the merits of open offices and remote work?

Jan 24, 2019
Sponsored Content

Produced in association with Citrix

Do open offices foster more collaboration, or just more frustration? Should managers encourage employees to telecommute, or is a scattered workforce less cohesive? The conventional wisdom on these issues swings like a pendulum, and for managers the only constant seems to be anxiety that they’re not getting it right. But new technology may offer some real answers. Ben Waber, a former MIT Media Lab doctoral student, is president and CEO of Humanyze, a Boston startup making software and sensors that give companies a better picture of how people actually work. He says the data the company gathers can predict employee performance and fuel a new form of “people analytics.”

This episode is sponsored by Citrix, the company powering the digital transformation inside organizations of all sizes. In the second half of the show, Citrix's chief security strategist Kurt Roemer says technology can help sustain work environments and policies that serve workers of all backgrounds and needs.

Business Lab is hosted by Elizabeth Bramson-Boudreau, the CEO and publisher of MIT Technology Review. The show is produced by Wade Roush, with editorial help from Mindy Blodgett. Music by Merlean, from Epidemic Sound.


Show notes and links

MIT Media Lab Human Dynamics Group


Citrix Workspace


Full transcript

Elizabeth Bramson-Boudreau: From MIT Technology Review, I'm Elizabeth Bramson-Boudreau, and this is Business Lab, the show that helps business leaders make sense of new technologies coming out of the lab and into the marketplace. This episode is brought to you by Citrix, the company powering the digital transformation inside organizations of all sizes. I'll be talking later in the show with Citrix's chief security strategist, Kurt Roemer. He's going to help us with the big question we're tackling in this episode which is, how can we build workplaces that actually work?

I mean, when you're designing an office space, is it better to give everyone private space to think, or is there more collaboration and communication when everyone shares a big open plan space? If you're trying to build strong teams, is it better to make everyone show up at the office every day from 9:00 to 5:00, or is it OK to let people have flexible hours and work remotely to suit personal or family needs?

Companies are trying it all different ways. And for managers, the only constant seems to be anxiety that we're not getting it right. I know we're still trying to figure that out here in our offices at Technology Review. Well, it turns out that new technology may offer some real answers. Our first guest today is Ben Waber. Ben is a former MIT Media Lab doctoral student who's now president and CEO of a Boston startup called Humanyze.

Humanyze makes software and sensors that give companies a better idea of how people actually work. Not just how workers move around an office, but how communication happens between teams and how all of that predicts performance. It's all giving rise to a new science Humayze calls "people analytics." And Ben is here to explain it to us.

Ben, it's wonderful to have you here. Thanks for coming.

Ben Waber: Thanks for having me.

Elizabeth: So, Humanyze is a spin-off company from the MIT Media Lab. I know that the work you were doing here at MIT sort of inspired the spin-off of Humanyze. But can you tell us a little more. Can you paint a picture of the origin story?

Ben: Sure. So my co-founders and I, we'd been using sensors in the lab to try to study things like salary negotiation. And the question there was, based on how people spoke, not words but just the way that people talk, could you predict, for example, who's going to win a negotiation. And you could, really accurately, with about 85 percent accuracy.

Elizabeth: So you're talking about things like tone of voice? Body language?

Ben: Tone of voice, interruptions. The body language as well really does matter. And these are things where, especially as social ecologists who had studied these things for decades... Typically what you do to study those kind of features is, you'd have some poor grad student record a video of an interaction and then they'd have to go frame by frame the video, over weeks for a single conversation, to figure out what was going on. But what we could do now is with sensors and with obviously more and more powerful algorithms, we could actually process that in a matter of seconds.

And so originally as we were starting our PhDs, we were doing this still in the lab. But then one day we had a professor from Sloan, from the business school here, say, well what you're doing in the lab is really interesting but actually you know I've been collecting things like e-mail data and surveys in really large companies, and it seems like face-to-face interaction data from sensors would be a really good way to understand holistically how people work. Because you have to understand that you can go to pretty much any company in the world and ask really basic questions about what goes on internally that people can't answer. Like, How much does management talk to the engineering team? Nobody knows. Even, How many hours do people work? And you think about how simple those questions are, how critical they are, and the reason people can't ask them is they don't have data, right. They use surveys, use human observers, but we all know how that's limited.

So we go to this major German bank. We just thought that was a cool idea, first of all. We'd never done anything like that before. We'd gotten these sensors in particular to work for something like two hours straight. And now we were being asked to get them to work for months at a time without any of our intervention. And we're collecting all this e-mail data survey data at the same time. And we even got to Germany to start.

Elizabeth: So you did this on behalf of a German company. They said "Try it out here." Is that what happened?

Ben: I mean it was our research, right. But this was, our collaborators over at Sloan had already been working in this bank. And so now we were just able to insert these additional sensors into it. And we collected this pretty massive amount of data about how much people worked. I mean on average at that time we were collecting about four gigabytes per person per day. And just to be clear about the data itself, it's not audio content. We don't even collect data with names or e-mail addresses or anything. It's essentially tone of voice, volume, how quickly you speak, location within the office. From email as well, again, it's not content, it's who communicates with who, when, how often. And we had all this data. And we also performance data and turnover data from this company. And no one had ever looked at any data like this before. The question was, Well could we maybe predict performance? Long story short we could, really really accurately. About six times better than all the other data put together. We wrote a number of papers on this, but after we wrote the first paper, of course we sent it to the leadership team at this bank, because they had been nice enough to let us do this. And when they saw all the results they said "Well, this is great. We're going to do a reorg based on this analysis." This was a paper a couple of grad students wrote right? And this multi-billion dollar company was going to completely reengineer the way they work based on our analysis.

Elizabeth: Okay so let's go back and talk about the way that you're gathering this data, collecting this data. You mentioned sensors. What's that mean?

Ben: There's two broad ways we collect data. One is digital data. So it's data that actually every company in the world already has. Email, chat, meeting data. Again, not content but really looking at the metadata from those platforms. Who communicates with who, when, how often, that sort of thing.

Elizabeth: You're looking at Slack?

Ben: Slack, yeah. We work closely with really big companies, which none of them really use Slack. But we can hook into slack. We use Slack. So for most people it's G Suite, Office 365, Exchange, that sort of thing. The other side is sensors and there's still two classes of sensors. So there's some that really are an evolution of what we developed at MIT, which are essentially a next generation company I.D. badge. And if you use a regular badge to tap into a building, it's typically got RFID in it which is a sensor. It's a radio. If I put little RFID readers in the ceiling I can forget where everybody is. Besides being a little creepy that doesn't tell you how people are collaborating, how they're interacting.

Elizabeth: It tells you that they are in the same space.

Ben: Exactly. Which is still useful first of all. But really at MIT what we what we saw pretty quickly was that the fact that for example you and I are talking right now is actually really important, rather than just sitting next to each other. And we need some way to determine that. So in these next generation badges we have two microphones, which again don't record audio but in real time figure out first of all that you're talking. Your tone of voice, your volume, how quickly you're speaking. We have an accelerometer that looks at posture so we can figure out how much we're mirroring each other, which gets very interesting. We can of course also do location within the office. I have one here, which I apologize—people who are listening are going to have to imagine it. Basically have this I.D. badge it looks like a regular I.D.

Elizabeth: Yeah, it's a little thicker. It's kind of a souped-up I.D.

Ben: That's right.

Elizabeth: It's the same size and he's got it on a kind of a key fob, like probably most people listening have access to.

Ben: That's right. And so what that has in it is a Bluetooth sensor. It's got a couple of other environmental sensors. The battery lasts for two years. It knows where I am within my office. It again can also then transmit in real time some other very simple information. The point is, two million people have that. From that it's still hard to recognize interactions. Again we would need other sensors in the space to do that. But that's coming.

Ben: And so typically what we're doing is, across all or a large section of an employee population, we're using all that data that already exists, so typically from digital communication and existing sensors, and then for really important groups where you've got a really big problem, that's where for a limited amount of time, you'll probably roll out these these souped-up badges and then rotate them to different groups.

Elizabeth: OK. So, I would really like to dig in a little bit more into the kinds of findings that you've been able to develop through these tools. So, tell us what can you see about the way work actually gets done in organizations that's been traditionally hard to see, and what kinds of new insights are you surfacing through this work?

Ben: There's a lot, because we have a lot of data, so I'll try to name just some of the top things that I find just personally interesting. One the big ones and something I see just a lot of companies looking at more closely is around the physical location of work. And I mean that pretty broadly. So I want to open a new office. Where should I do it? I want to build a new headquarters. Which division should I put there? Even then within a single building, within a single office, which group should I put next to each other. Traditionally those things are very subjective and political, in that I'm an executive and I read some article maybe in the Tech Review, about what Google does. And I say, “Google's cool, let me do the same thing as them.“ And that's crazy. That's not the way you should run your business. You are not Google. Beyond, that for a lot of these things, Google is, just like a lot of these other companies, doing things very subjectively.

The question is, how would a particular decision change the way your people work. And so there's this basic question, “Well, in our company, when we put two groups next to each other, how will that change how they collaborate?” Because the vast majority of companies are making these workplace decisions or workforce decisions to try to change how people work. But if you don't measure that it's very hard for you to understand if what you're doing is actually effective or not.

And so one of the things that's been most fascinating is, and again this is now over millions of people, I can see that no matter what communication channel you look at, email, chat, phone calls, face-to-face communication, the likelihood that you will communicate with somebody is directly proportional to how close you are, from a desk perspective. If we sit next to each other we tend to e-mail each other more. We tend to chat more. As you get pretty much onto different floors you might as well be in another city. You still have the meetings. But again if you're remote you're still going to have the meetings too. And I think that's the challenge with a lot of these tools and with these decisions, is that people assume that OK if I put everyone in one building it doesn't matter how big the building is, everyone's still going to communicate. Which isn't really true. But then beyond, that you think, well, I can communicate with anyone. I have the tools to do that. So I don't have to do any work to make sure that people collaborate effectively. And that's also not true.

Elizabeth: Wow. Ok this is quite a conclusion given how much energy and sort of thought is going into remote working and people working at home. And there's been a lot of talk about whether that's a good idea or a bad idea, and certain companies have said no no no, we're not going to allow it anymore. What does this mean for the folks who want to have a little bit more of a flexible or more balanced attitude towards work and home. Say more about what the what the implications of that could be.

Ben: Just to be clear, the data is pretty clear that if one day a week or something you're working from home it doesn't significantly change how people collaborate. But there's a big difference between that and literally never being in the same physical place as your co-workers. And again I think a lot of times those distinctions get lost. And I think that a lot of times those decisions are made just on a cost basis. You know people are a cost. People are cheaper here. So I am going to pick the cheaper thing. But what we actually care about is performance. If people are 30 percent more expensive here, but it makes us more than 30 percent more effective, then I'm doing that. The issue is it's been really hard to show those kind of effects in the past. And what you're now seeing is companies can start to make those calculations.

Elizabeth: I love the data approach to something that's kind of inherently woolly or hard to pin down. And I find it a little creepy. So let's talk about the privacy element of it. The sense that kind of we're watching employees. How do you mitigate the concerns that I'm sure you hear about a lot from both employees and the folks who are bringing you in.

Ben: And those are legitimate concerns. There's certain things we do, which are that, again, if we're deploying new sensors, we deployed on an opt-in basis. We don't collect any names or e-mail addresses. Beyond that, everything, all the metrics we provide to our customers is aggregated. We never show any data points with less than five people unless there's noise added and then it's statistically completely anonymous. That's important because then you can't see what is person A doing at 2:30 on Tuesday, because who cares. There's not really a good business use case for that, and we can all think about how that can be abused. Where you do need to know is, all right, maybe I have top performers. What do they do differently than everybody else? Not a single person because that could be a total outlier. I want to see the distribution. In a similar way these teams need to work a lot with each other. Is that happening? Or again for our customers for example in Japan, I really care about overwork. I really want to make sure that we are not you know having people work a ton. What percentage of the workforce works more than 80 hours a week? If I roll out a program to reduce that does it change that. But it's these big macro level things.

Elizabeth: Let's talk a little bit about physical space. We've heard a lot about the trend toward open offices, now away from open offices, where no one has any private space and people put their noise-canceling headphones on. Do you have a philosophy or a point of view about what is the best way to design an office?

Ben: That's a good question. Because a lot of times people ask "What is the best office?" Which, of course, there is no answer to that. What I would say is, you really want to understand what kind of behaviors you want to create. You of course also need to consider things like the public image of the company, so branding also does matter. And then you want to understand well what are the tradeoffs as you weigh both this public image as well as the internal actual function of how this space will change behaviors and different spaces for different companies with different cultures do different things. We for example, at our headquarters here in Boston we do have an open office. At the same time, we have lots of big movable whiteboards. And so for example right now the engineering team, there's a big wall of whiteboards on one side of them so that they're essentially separated from the commercial side of the business which is actually on another wing. But there's a visual separation. It also separates noise as well. And I can quantify what those things do they can also move them around depending if they want to have even smaller sections. They can do that.

But again I can tell you with some or other customers, even in those environments you would actually dramatically reduce the amount of interaction within teams. Probably actually increase it between teams, but what I would say is that given the culture of a company the impact of space is relatively consistent relatively. So that if in one part of the company we have an office that looks like this, I can predict fairly confidently that if I put that same type of an office in another division, it would probably have a similar impact.

To your point, you get these big trends that fall in and out of favor. So you know now open office is falling out of favor, maybe, a little bit. But then, do you move to cubicles? It just goes, it cycles. And rather than saying "This kind of office is cool and that's what we should do because it's cool," you can say, Well, that does matter to a certain extent but given what behaviors we want to create that's why we picked this thing. And what we can also admit is we're not sure if that's going to work. I'm not sure if I let people work from home two days a week is that going to be good. If we have an open office here but then private offices here, is that going to be good. These are hypotheses.

Elizabeth: So it's more of an iterative approach.

Ben: Yeah. And it needs to be even because, even if today I find an office design across my whole business which is going to be a mix of different spaces that is effective, your company changes. Business changes. And so the behaviors that are effective today, in couple months, especially years, are not going to be effective. And so there is actually a need to's not just finding one thing and then doing that forever. It's let's keep changing. But the best way to keep changing isn't waiting for some external market signal to say hey your offices look old. It's, OK, we're now not aligned with the behaviors that we need to create to be successful. We need to try something else.

Elizabeth: So is there a workplace engineer role that might emerge? Someone whose job it is to think about what the algorithms are saying, what the plan is for the day, what the goals of the company are, what maybe some weak spots or frailties might be?

Ben: This is actually what we're seeing in our customers. One of our longest running customers who's been using our technology for over four years, they now have a team, an internal organizational analytics team of think 17 people now. Literally their only job is to use our analytics to manage the business. And that's everything from workplace decisions to H.R. decisions to I.T. tools to operation. You name it. And actually all of their divisional leaders in the C suite every quarter meet for half a day and all they do is review these metrics and they say well you know is this what we want. Here are the initiatives. We're going to plan to change these specific behaviors and then we're going to check back on them next month. Now they're not the ones who are looking at this sort of stuff on a daily basis. But I think the fact that we're seeing you know Fortune 100 companies do that is really important and I think it's going to be a model that others follow.

Elizabeth: Well this is great. Thank you very much, Ben for sharing this with us.

Ben: Thanks so much for taking the time.



Elizabeth: This is the second in a three-part series of episodes on the future of knowledge work sponsored by Citrix. Citrix has put a lot of thought into how to make workplaces more worker friendly. So we wanted to round out this episode with a conversation with Kurt Roemer. Kurt is chief security strategist at Citrix. That's an important role at a time when knowledge workers are accessing company data and applications from all different locations on all sorts of devices. So I wanted to ask Kurt what Citrix is doing to help people work productively and securely no matter where they are.

Kurt Roemer: You do have people who are primarily working from nontraditional locations these days. So instead of going into an office every day and sitting in an office or cubicle, they're working from airplanes, they're working from home, they're working from hotel rooms, and especially public areas as well. And the challenge there is to work on things that are risk appropriate. I sat next to a tax attorney on an airplane had no idea that he was a tax attorney until I looked over and he's working on his corporate tax returns. And that's some very sensitive data subject to government and S.E.C. regulations. He should not have been doing that within the purview of a lot of other people on the airplane. And you also hear walking through airports, salespeople who are trying to close deals and are talking about some very sensitive data very loud over their cell phones. And you know part of this is helping to make sure that you're encouraging people to do the right things giving them training. But increasingly the technology can come into play and help us to work in risk appropriate ways.

Elizabeth: In this episode of the podcast, we've been digging into how to organize workplaces and workforces. In this particular conversation, it's about how to design an office to encourage greater communication and productivity within the office. We've also talked about how to achieve the right balance between on-site, face-to-face work and telecommuting and remote work. What do you think that business leaders should be thinking about when they want to take a decision about changing the layout of the office to make it more open, or indeed maybe less open. And then I'm going to ask you the same question about telecommuting.

Kurt: When I take a look at this, office layout, the workspace should really balance the benefits of collaboration along with individuality and privacy. Everyone works in slightly different ways. And there are some generational work biases and practices that need to be understood and accommodated. In other words you might have some people who if they're in an open area with a whole bunch of other people and maybe they're working in a call center they're working collaboratively on a project, they're the most effective. You might have some other people that for them to think they need it quiet or they need to be able to have their own space. They should have a room to be able to go in and have that space. And likewise if you're a manager having a sensitive conversation with an employee you need to be able to do that in more of a private room type of environment. So you need to have both available and it's really helping people to migrate to the type of work environment and office space that makes the most sense for them for what they're doing at that point in time and oftentimes getting people together and encouraging collaboration really opens up some of the benefits of people being in the same space and really building off of each other's ideas so that can be can be very rewarding. But it can be overwhelming to some people who just don't like the constant background noise.

Elizabeth: And how about telecommuting?

Kurt: Telecommuting is kind of a special challenge. We had talked before about working from airports and hotel rooms and others. People are working from a lot of nontraditional locations, and so telecommuting is not just working from home. It's really how can you be productive regardless of where you're working from. Part of it is understanding the physical needs. You have some place that's quiet, some place that can appropriately protect your privacy when it's needed and sensitive information. And you also have a culture that helps to encourage people to work together regardless of where they are you know. In other words you've got somebody who really doesn't come into the office very often, how are they still seen as part of the team and how do individuals and the rest of the team communicate and collaborate with them and bring them in, even when something would have been a last minute type of watercooler conversation for example.

Elizabeth: What are your Citrix offices look like, and how common is telecommuting for Citrix?

Kurt: From a Citrix perspective, we have open collaborative offices, we've got private space that people can use. We've got space that even as contractors or customers come in they're able to go in and utilize a private area. And they've got a system that they can go in and get to their web and SaaS and Workspace remotely without even having to think twice about it. Be able to log directly in. And we also encourage people to work in offices throughout the world as well as work from home when it makes sense.

So you might have someone who has young children and the best thing for them is to be able to be there with those young children during certain times of the day. Why burden them with having to come into an office when they're going to be much more productive and much more satisfied if you give them a little bit of flexibility a few times a week. And that's been a real boon for us at Citrix. It's really knowing that people are working more, knowing that they're working from home, oftentimes right when they get up, but giving them flexibility to be able to work from anywhere and be able to use that to help fit their lifestyle and help benefit the company and our customers.

Elizabeth: Do you think that your physical offices and the way that you organize your workforce remotely, versus in-office—have you been able to kind of test this new ideas and new products within Citrix organization as sort of a petri dish or a guinea pig?

Kurt: Yeah, we bring our customers into our offices fairly often, especially for our executive insights program, where we bring in executives, talk to them about our vision and future direction, and then also show them our offices and how people are collaborating. And it's a big point of interest, because many other organizations are still struggling with how do we move from cube farms and how do we move from more of a traditional environment to a more collaborative environment and we're able to show them how people are working, how they have information displayed, and where they can go if they're working on something more private.

Elizabeth: And what about how all this impacts the experience of being an employee? Retention, or even recruitment. Do you see a difference in companies' ability to retain employees when they have sorted through workspace and workforce management?

Kurt: Yeah definitely. If you can show an employee that they're able to help manage their own employee experience and if there is a major snowstorm, instead of having to spend an extra six hours on the road, they can work from home and be able to utilize that six hours to be productive as opposed to being frustrated. It's a simple example but it plays out quite often throughout the winter and it really helps people to be much more satisfied with their work. And then from a Workspace perspective, instead of having to remember to open certain applications to go in and do work instead of having to go through and search for information that you need or might be relevant for you today, including news, if your Workspace presents all of that to you, you have much less action that you have to take on your behalf, to think about what you've got to do to be prepared. Your workspace is helping you be prepared to be the most productive at that particular point in time. And as analytics and evolve it will definitely help you prioritize work as well so that you're working on the most meaningful things at that point.

Elizabeth: Well Kurt, thank you so much for talking to me about this.

Kurt: Thank you, Elizabeth.

Elizabeth: That's it for this episode of Business Lab. I'm your host Elizabeth Bramson-Boudreau. I'm the CEO and publisher of MIT Technology Review. We were founded in 1899 at the Massachusetts Institute of Technology. You can find us in print on the web, at dozens of live events each year, and now in audio form. For more information about us please check out our Web site at

This show is available wherever you get your podcasts. If you enjoyed this episode we hope you'll take a moment to rate and review us at Apple Podcasts.

Business Lab is a production of MIT Technology Review. The producer for this episode is Wade Roush with editorial help from Mindy Blodgett. Thank you to our sponsors Citrix, the company creating people-centric solutions for a better way to work. Thank you for listening. We'll be back soon with our next episode.