Ben Waber, Ph.D., is the President and co-founder of Humanyze. He is a visiting scientist at the MIT Media Lab, previously worked as a senior researcher at Harvard Business School, and received his Ph.D. from MIT for his work with Alex “Sandy” Pentland’s Human Dynamics group. Waber’s work has been featured in major media outlets such as Wired, The Economist, and NPR. He has consulted for industry leaders such as LG, McKinsey & Company, and Gartner on technology trends, social networks, and organizational design. His book, People Analytics, was published by the Financial Times Press in 2013.
At Humanyze, we believe that a company’s greatest assets are its people and how they collaborate (and have the data to prove it!).
Born in MIT’s Media Lab, we are redefining the future of work through science-backed analytics and data-driven insights that help companies make continuous workplace improvements that benefit both employees and the business.
With a global presence spanning the Americas, Europe, and Asia, our mission is to help leading companies around the world unlock their organization’s full potential with workplace analytics.
Rebel HR is a podcast for HR professionals and leaders of people who are ready to make some disruption in the world of work.
We'll be discussing topics that are disruptive to the world of work and talk about new and different ways to approach solving those problems.
Follow Rebel HR Podcast at:
The issue is that most companies for people decisions are essentially when it boils down to it, most of those decisions are just this executive thinks this is the right decision, because they feel like it's the right decision. That's it. And then your only way to convince people that's a good decision is to tell a better story than that. That's it. Right. And I think that that is a fundamentally limiting factor on how effective we can be as people managers within organizations.Kyle Roed:
This is the rebel HR Podcast, the podcast, where we talk to HR innovators about all things people leadership. If you're looking for places to find about new ways to think about the world of work, this is the podcast for you. Please subscribe, favorite podcast listening platform today. And leave us a review. Rebel on HR rebels. Rebel HR listeners, thanks for joining us this week really excited for the conversation. We have a special guest, Ben WebVR. PhD is the president and co founder of humanize. He's a visiting scientist at the MIT Media Lab, previously worked as a senior researcher at Harvard Business School and received his PhD from MIT for his work with Alex Sandy Pentland Human Dynamics group. Welcome to the show. Ben, thanks so much for having me. Really excited for the conversation and you know, your background is is is going to be really interesting to dig into. I'd like to start by understanding a little bit about the organization that you co founded humanize.Ben Waber:
Yeah. And humanize really spun off the PhD research that my co founders and I were doing back at MIT, I'm really motivated by this idea. I guess I had always assumed that when particularly big companies made big people decisions, you know, do a reorg, build a new headquarters? Well, of course, they must collect a lot of data and run tests. And then based on those tests, make people decisions is a very quickly disabused of that.Kyle Roed:
I'm laughing.Ben Waber:
And so again, it comes down to this basic fact that you could ask pretty much any company in the world really simple questions about what goes on internally about how work happens? And they just can't answer it. You know, how much does management taught me and guaranteed? Nobody knows, right? I mean, even how many hours people work. And you think about how simple those questions are, but also how critical they are, in the reason that people can't answer them as they don't have data. Again, we might use surveys, we might use human observation. I mean, of course, that's very useful. But I think we understand the limitations of that, right? I mean, if it's a sunny day out, like it is here in Boston, then you'll answer differently, that it's cloudy, at a basic level. But of course, beyond that, we just as a course of our normal work today, generate so much information about how we're collaborating about how it working. We have email, chat, meaning data, even sensor data about the real world. And essentially, the metadata from those different systems tell you well, how are different teams collaborating with each other? How does that change over time? How do people spend time? Really, our focus had been at MIT and actually continues to humanize not so much on the individual level, right? We're really looking at a team in higher level metrics. But the idea is that at very large scales, these network effects, you know, how siloed is the organization? You know, for the most successful managers, what do they do differently than other managers? That's, that's actually not about individual level questions. It's about these these networks, distributions and the organization. And so at MIT, we were doing studies in a lot of real companies, and then figuring out which metrics matter for predicting outcomes, we all care about performance, retention, what have you, after we started the company, then we've been able to do that at larger and larger scales. And essentially, what we do is we hook into this existing data, again, the companies already have, and then provide analytics on the health of the organization, to these companies. And it actually was pretty exciting. In late 2019, we got to this point, where, because we're deployed across every single boy and a number of fortune 100 companies globally, we actually have four information workers, a globally representative dataset on how work happens, which is very exciting. And we could talk more about how we got to the point of validating that, but then, of course, it gets to this thing in March 2020, where these metrics move, you know, more in a month than you expect in a 10 year period. And, you know, when I talk about these metrics, it's things like how long are workdays but also, how many weak ties do people have? And we could talk about how we define that. But there's a whole variety of metrics where I actually know what that distribution looks like globally, which is very, very interesting, and enables you to do some other much more powerful things as we get more and more data. So really, what Companies use our technology for, again, obviously today, a lot of it is workplace strategy. Frankly, it's which teams based on how they collaborate, coming to the office, how often but also with which other teams, you know, who are they collaborating with what they need. But really to get a handle on what's going on internally, what's driving the actual KPIs we care about in terms of when people side, right? And then also, as we make management decisions, what is the impact of them, and not waiting five years to figure out this reorg work, but being able to validate very, very quickly, or even test, ideally, very rapidly, different changes to, again, the way we work.Kyle Roed:
Fascinating. Yeah, you know, you ask those questions, and I, and I, I kind of laughed at it initially. Because, you know, a lot of times, you know, especially big organizations, you assume that somebody has this all figured out, right? They've they've some super smart person sat in a room and crunched the data and just figured out, okay, this is the right path forward as to how we structure work. And then, and then, you know, and then like you said, 2020 happen. And I actually think that was a great example of proof that nobody knew what the hell was going on, in their workplaces. And, and were surprised when they figured out Oh, we actually can do some of this virtually, I guess we didn't actually need all of this office space to in order to effectively do our jobs, because because there was a fundamental misunderstanding about how people actually performed work.Ben Waber:
I mean, also, I think, on the change management side, all of us would have assumed that we would need a six month change management process to effectively move to something like remote work. And you basically given a good reason, and two weeks of time, people were able to do it pretty effectively. Right now, it's not to say there aren't issues with remote work, and we can talk about that I you know, date on challenges, but in terms of, you know, are people working? And are they doing, you know, the best they can? Is it like pretty good? Like the answer is yes. Like definitively, right? Yeah, I'm in a you think this is the pandemic is sort of exposed this dirty secret, not just within HR, but I would say management in general, that I think a lot of folks have this assumption, again, that because people are executives at a large company, they must know what they're doing. Right? And no one knows what they're doing. Right? It's actually it's very, like an example that I like to use to think about this, like I fundamentally believe that some of the challenges we see in people management in particular, a lot of it boils down to this fact that there just aren't people aren't data driven and don't have a good reference points for what are actually effective management processes. And so what I think this leads to is a lot of people just copying what some successful company does, because they assume that Google's people decisions or Netflix people decisions, well, that must be why they're successful. Of course, they could have been successful in spite of that, right? And in a way to think about this, for people who are sports fans, I encourage you to look at the graph of three point shots in the NBA per game. Okay. It's, it's really interesting, because so think about think about basketball, pro basketball, that is a workplace now, of course, is very different than where probably most of us work. But it is a workplace with a key difference that we have very easy to understand metrics about success in that environment. Right, we have points, and we have wins and losses. And we also have the advantage that each team plays, you know, at least 72 games a year. So you're sort of beaten over the head with what essentially are effective, quote, unquote, management practices. Now, of course, again, we're talking about basketball, but still think of it as a workplace. And it's interesting because the NBA, if you're familiar with it, you know, back in the late 70s, they introduced the three point line. So again, before then every shot was worth two points. And you can write down a math equation, showing that if you take more three point shots, you will score more points and way more games just straight up. Again, that's subject to certain assumptions about accuracy. But did you just leave it at that? Okay. You could maybe make the argument that maybe it would take pro basketball players a couple years to get used to taking longer shots. And so you would what you would expect is if you looked at a graph of how many three point shots two teams take per game, you may be expect, you know, an initial, you know, uptick and then after a couple years, you would assume it would plateau, which would be like an option. to my level, and maybe even if you could use a different argument, because I know it's a generational thing people need, you know, the new players need their entire careers to be that fine. So you wait 20 some years, and you'd expect to see this. That's not what you see, when you look at that metric, what you see is a slow, steady increase in the number of three point shots per game to today, it's still going up, still going up today. And this is the thing about this environment. If this work environment were optimum people were rational about their workplace decisions, then that shouldn't happen. What's fascinating about basketball, again, is there are objective easy to understand metrics, right? There are point like, literally, every night you are beaten over the head with the fact that taking more three point shots is better. He's better, it's objectively better, I can read an equation down and do that. And yet, it's still taking them 40 years to get to the point they are today. But now we think about other organizations think about other types of work where you don't have that kind of objective metric, right? If you're building software products, for example, right? How should you work today? How does that impact the actual sales of that software product? It's actually unclear. Right now, you can, again, overtime show correlations, but it takes a lot more time to show that effect. So there's, you know, and also, the rules aren't a static, the rules of society of what the market demands changes all the time. And so And beyond that, we're not we don't have 72 games a year, right? Maybe if you have over your entire career, maybe you have a couple dozen large projects that you do over your entire career, not every single year. And so I think all these effects compound, just the misunderstandings that people have about people management in particular. And this is why sort of a lot of folks just resort to copying, rather than what I would argue, looking at data, reading, you know, academic meta analyses, I think that's the right way to go. But I think that this is, like a really big issue for HR. And I think the pandemic has exposed this, because so many decisions that companies are making are falling flat on their face, and just exposing to employees that people don't know what they're doing, which is going to make any change. Very, very hard to implement moving forward. So sorry, sorry, for that. I had to get that out. Because I think this is, this is like the critical topic, in my view for you know, the next I mean, for a long period of time, honestly.Kyle Roed:
Absolutely. There's so much. There's so much to unpack there. But I think the three points story was was a perfect way to explain it. And I, you know, I will call this out. I mean, when I was early in my career and kind of learning about the world of work and trying to figure out where I was going, you know, you It's comforting to think somebody has this all figured out, you know, somebody up here, they've defined a best practice, all I have to do is execute on this best practice, great. Like, okay, I can sleep at night, all I have to do is what this manual tells me to do. But the longer you're in your career, and the more exposure you have to it, the more clear it is that, you know, it's just like, everybody talks about adulting, right? It's like, at one point, you look around and you realize, oh, I guess I'm the adult in the room. And I'm supposed to make this decision. And I don't know what decision to make. What do you think, Bob? What do you think, Sheila? I don't know. Yeah, let's just let's just go this way and see what happens. Right? I mean, like, that is literally what happens a lot of times.Ben Waber:
But I think that this is this sort of thing. Like it's okay to say you don't know, right. But I think that companies are so afraid to do that. And this is what's an interesting with a data driven perspective, right? So we sort of look at how are people collaborating? How does that change over time? And even with that data, it's still a hypothesis, whatever your people decision is? Or you could say, Well, I think based on how they work, this team should be in three days a week. I don't know if that's actually like, my hypothesis is that that will mean that it is a hypothesis. And so by calling it that, you give yourself the permission to be wrong. And then if you're wrong, saying, hey, you know, what, a couple weeks later, that didn't work. We're gonna try something else. But I think the default approach that most companies take is to say, Well, we know what the right decision is. And I think in the past, what that enable people to do is fall into line and do this thing more easily. What I think you see happening now is people pushing back and say, Well, I just disagree with you. An easy example to think about, again, are these workplace strategy decisions company saying we we should come in three days a week, because that makes us more productive. And then you have employee saying I disagree. And there's nowhere to go from there. If you don't have data, right? It's not you can't say, if this team comes in three days a week, there's this specific behavior that we believe will be impacted by it. And so we're going to look at that metric, and then we're going to report it back. Okay, is that so you're gonna make everybody happy with that decision, but at least people can agree on a shared reality. The issue is that most companies for people decisions are essentially when it boils down to it. Most of those decisions are just this executive thinks this is the right decision. Because they feel like it's the right decision. That's it. And then your only way to convince people that's a good decision is to tell a better story than them. That's it. Right. And I think that that is a fundamentally limiting factor on how effective we can be as people, managers within organizations.Kyle Roed:
That's a really interesting insight. And I think, you know, going back to your, to your your NBA analogy, one of the terms you use, I think, is a really big assumption. And that's, you know, the assumption that people are rational. There's a really big question mark there. And I think every HR person, you know, every HR person listening this right now is like, I know the answer to that. Yeah. It but but yeah, and, and I do think, you know, the, the the power dynamic is shifting in the workplace as well, like, you know, it's 20 3040 years ago, people didn't have things like the Internet to go Google and have the collective knowledge of the universe available at their fingertips in order to find data or to find something that's, you know, that supported their their hypothesis that, hey, no, three days a week isn't good for me. I want zero days a week, and maybe I'll come in once a quarter for, you know, you know, random meeting here and there if I absolutely have to.Ben Waber:
Exactly. Yeah, I think that that's the point, it's, people aren't rational, which is we just have to deal with that. And that, you know, wouldn't pretend that having, you know, having data Hana have having analytics fundamentally changes that. But what it does enable people to, again, is to be able to level set, because we don't have this objective perspective on work. And not to say that data is fully objective, like it's not, but that is, at least, you know, a more uniform view on how work is happening. And I think enables us, again, to just ground our conversations in a shared understanding. And I think that is just so helpful for not just making better initial decisions, but moving much more quickly, once it's clear that that decision isn't working, because it takes companies so long now to change. Things about work that clearly are not effective, like clearly ineffective. But that to your point, well, this is a best practice that we've just done for 30 years, or that company X does this and they're successful. But that neglects this idea that I mean, you can take, take any large successful, for example, tech company, right? I mean, you know, all the big guys. I mean, if we're being honest, for the next couple years, they could lock their employees in broom closets. And just by notion of their overwhelmingly powerful market position, they would still be pretty successful. Right? But that doesn't mean they were making the right people decisions during that time. And so just blindly copying them won't give you the sort of results you're looking for. But if instead, we were able to say, well, here's qualitatively, you know, how what behaviors relate to performance in these organizations or in our organization, and this specific decision led to this change in behavior. Again, I think that's the way that I at least like to think about it, because then it's a lot clearer what the effect the cause and effects are, and not 100% clear, but much more than just, you know, oh, well, Company X implemented holacracy and they're really successful now. So it must have been because of that, versus it could have been totally irrelevant. totally irrelevant, or it could have even detracted from their success. And it just unclear.Kyle Roed:
Yeah, I think that's a that's a really interesting point. And, you know, I'm just thinking through, you know, the workplace change initiatives and the challenges, you know, a lot of times, you know, HR people, a lot of times we're not in this because we're really good at data analytics, we would probably be doing something different, we probably would have gone to school for something different. You know, if we were really in the you know, all that all that a lot of times we're more you know, empathetic and we're about emotion and energy and, and, and culture and all these all these like cool buzzwords, you know, and we're trying to, you know, trying to help increase collaboration and all that stuff. But what, what where we can get paralyzed is if we have enough assumption, or and there's an emotion behind it, you can just get frozen, because you're afraid of what the reaction might be from employees or you're afraid that, you know, this might degrade the culture what? So if you don't have any sort of any sort of foundational data, you know, even if it's not quantitative, you know, even if it's quality, just just general feedback survey, whatever, yeah, like, you still need to have some sort of a base case, like some sort of a thesis, and then you can actually, then you can make decisions. Yes. Objectively, right? Because if, I mean, if we, if, if all of us empaths are running around making all these decisions, we just won't go anywhere. It you know, it'll be a drama every five minutes.Ben Waber:
Just to be clear, right? Is that, like, to your point? It's not to say that, you know, subjective feedback surveys, just interviews, like that is so useful data, like just to be totally clear, right? It's a different kind of data. And I think you need that as well as behavioral data to have this holistic view on my opinion. I think the other point, though, is that it's I don't want to say that everyone in HR should be data scientists. And that's how that people side of business works. It's not right, these are I do think that these are new competencies that needed to be added to the mix. That doesn't mean that those data scientists should, you know, maybe in some cases, lead those teams. But that doesn't mean like always right, it means that I do think, to be competitive, long term, and to frankly, just be successful, even in the near term. Like you need some folks like that on the team, right? Just to crunch over data that again, you already have just to be totally clear, right? Like, everybody already has this kind of data. What we've tried to do it, humanize at least is make it much easier to get started. But again, I don't want to pretend that like we do something that it's impossible for other organizations to do not write. And that I do think that the HR the future, it's not just the way we've always done things, obviously. But it's also not just adding data scientists to be clear, it's, you know, and I do think I see a little bit more of this. It's having people with real estate background 10 people's IP background, right. So you think about all of the things that you again, we talk about these fuzzy, fuzzy concepts like culture? Well, how would we define that? But of course, you know, and so then data scientists can talk about who here some actual metrics that we could calculate. But of course, it's not also just the HR processes we have, it's okay, well, what tools do we use to collaborate and those tools, it could be, you know, could be zoom, it could be our physical offices, can we make those decisions holistically, rather than having those siloed in different parts of the organization, I mean, I would also think having people with advanced degrees and things like organizational behavior, io, Psych is very, very useful, not just to, again, I think, try to bring in some of the external research on the topic that is continually evolving, but also for planning experiments, I think that is incredibly critical. For HR being successful long term is, again, the ability first of all, to message this, which again, those people can't message, right, we come back to, you know, folks HR much better at messaging those things. But that, alright, well, which parts the organization, you know, what data we should collect? Again, I think all these folks working together starts to get very, very exciting, and I think also become a much more central part of the organization. And I do feel like this happening a little bit in different pockets of industry, but I think there's really opportunity for this to happen, you know, much more broadly, and for this to just be the way that HR functions for the long term. Yeah, it's,Kyle Roed:
you know, it's really interesting and encouraging for me. So, you know, when I started in the, in the field, it was totally by accident, I didn't go to school for HR, you know, as many HR people that just kind of fall into it, right. And so I was kind of coming from a different perspective came from an operational background, and, and kind of just stumbled my way into it and learned along the way, but it's really encouraging to see some of the new some of the new hires that I've brought into my team, you know, out of college, or recently out of college, and some of those topics and, and things that they're teaching at school, in IR Psych and, and in, you know, you know, analysis and organizational design and comp design and that sort of thing. And it's, you know, the other thing that I love is, and this is maybe generalizing a little bit, but they're also they also are not afraid to challenge and say, hey, you know, what, why don't we do this? We should do this. You know, and a lot of times, I'm like, Well, I don't know what that is, but that sounds exciting. So yeah, let's discuss what this what is this thing, but it's, it's really, you know, what we've been talking about, it's like, it's just utilizing the scientific method, yes to to do something. That's it, you know, and that's and if you don't have some sort of a method like that, then you're you are just going to be throwing stuff against the wall to see what sticks and, and and you could approach really assume that something that you did actually caused something. But it was just a corollary effect. And it didn't, it didn't have anything to do with it.Ben Waber:
It's very, very that up because I still, I still teach the people analytics executive course back at MIT. And one of the slides I have on there is actually the scientific method. It's like, let's just review, you probably remember this from like middle school science. But actually, you could apply this in your job. And let's just review it. And it's not crazy complicated. It's just, it's a slightly different way to think about making decisions about how we work. But that I, again, I believe that pre pandemic, because I've been teaching that course now for years, and I, pre pandemic, it was maybe slightly a harder sell, not that people didn't agree that oh, yeah, that's probably the right way to do things. But I think there was just so much inertia in the way that we've always done things in that it's so hard to overcome that. And I think if there's a positive to come from the pandemic is that it's broken those things, it said, you know, everything is in flux. And that actually makes it easier to change. It also, I think, very nicely makes it harder to copy other people, which again, I think, is this huge, really unproductive crutch that I would just say management has in general, because no one can gradually claim to know what they're doing today. I mean, if in the past, you could say, oh, well, this successful company, they must know what they're doing. Well, no one's done this transition successfully. So no one can claim Oh, I actually know how to manage in a hybrid workplace. No, you don't, no one knows. So please, Like, and that that I think, is very, very exciting, because there's going to be so many more models that people can can try. And then again, this idea of, oh, how should we evaluate these different models? Oh, here's the thing we've used for hundreds of years in science, it's pretty, pretty damn successful. But at that, it's been much easier to have those conversations.Kyle Roed:
Absolutely. And, you know, I think about it too, especially in in when you're dealing with people, you know, you are never going to have the answer. Yes. You know, the ever I mean, if you think that this is the best practice, and you know, what Jack Welch did this, it worked, you know, we're gonna, let's just use an example, we're gonna force rank our employees, because it worked, you know, 40 years ago. And we'll put everybody and we'll make it, you know, we'll make it transparent. And it'll, it'll just make sure that people are driven to excel and the bottom 10% will either self select or get removed from the organization. Perfect. That worked. It worked back then it should work. Now. If you do something like that right now in this environment. Good luck. Because, because that, you know, that that what may have worked at that company, could be a complete demoralize or in your organization, it could work. But if you think that we're just going to plow through this, and you know, we'll know if it works in five years. If when we look at our turnover levels, you know, good luck to you. Yeah, yeah. But yeah, and I do think it's also, you know, the other topic that I, you know, I think is really fascinating. And I sorry, I told you, I was gonna nerd out on this a little bit. But, you know, there's a reason that a theory is always a theory. You know, can you ever really prove anything? 100%? The answer is? No, not really, especially when you're dealing with people.Ben Waber:
Yes, I do think this is actually an important distinction between industry and academia, when it comes to experiments and analyses, because academia does default to this statistical significance, which why is that point 05? Unclear. So like, that's, that's also arbitrary. So whatever. But the reason that's important is because when it comes down to it, we have to make decisions. And so we're never going to have complete information. And in some cases, so ideally, of course, we're able to run experiments and have statistical significance. But let's say you run the experiment, you could even do that internally do everything, right. And you don't have a significant result. With that that happens, you still ultimately have to make a decision. Right? And so there are ways to evaluate, you know, experiments, not just using statistical significance. But I do think that this is actually a real challenge that I've started to see this. Actually, it's interesting, because marketing has sort of gone through this transition. I do think that the marketing model is sort of the model that HR could follow. Because marketing, right, 30 years ago, you asked someone what is a good click through rate, people had no idea what that was. And now it's an incredibly data driven part of the organization. Right? There are still creative folks there but they're running tests. They're, they're using data. And a lot of the things I think that HR is for to bump up against over the next, you know, couple decades, like that's what they dealt with, right? That sort of cultural transition? How do you build up these teams? How do you make this part of the core competency of the field? I really think that that's the model. But that one thing that you can see folks in marketing struggling with now, again, is when they bring in, you know, a lot of these, you know, pretty, pretty sophisticated data scientists, but they have this, you know, rigorous academic background, which says that, you know, here's statistical significance, if it's below that you basically ignore it. But that actually, that doesn't apply as much in business, it's still as important information, but that I do think this, again, that's why I really think that these multidisciplinary teams are going to be very, very effective, because you might have some, some academic, you know, people with a strong academic background, who really hesitate to make a decision. But then again, someone with, you know, more of that day to day business experience is gonna be like, listen, I totally appreciate that, um, that objection, but like, I've got to choose a or b, right? And so ultimately, all I care about is do we think a or b is likely to be better, and we might not have high confidence. And we can say that, right? And that that's actually useful is saying that as well, like, it's because then when you are really confident, right? Not say again, you're 100% confident, but then you can communicate that, like we're gonna do this. And we're pretty confidence going to work because of x y&z Next time, we're going to do B, but I don't have the same degree of confidence, actually, they have. And again, it just allows for much more nuanced messaging to the organization, which I think is very, very helpful for the long term success, not just for organizations, but people management in particular, as well.Kyle Roed:
Absolutely. You know, and, you know, I had a great mentor earlier in my career. And she, she, at one point, we were struggling with the decision, and it was my decision to make and I didn't have it, I didn't have a breadth of experience to draw upon, you know, I was relatively young. And she, she looked at me, she said, Listen, it's, it's worse to make no decision at all than to make the wrong decision, you just have to, you have to decide and move forward and adjust if it doesn't work. And she was 100%. Right. And now I think about it more in terms of like, it's just making a probabilistic bet. You know, it's like, Okay, what's the standard deviation? That I am going to totally screw this. And, and if that if, you know, if I've got enough of a sample size, I can make that bet. And you know, what, if I'm, if I'm about 70%, I'm good, dry, you know, but but if I'm, if I'm about 40%, you know, maybe maybe I go back to the drawing board. And I, I do a little more analysis, but but I think that the more the more that we think in those terms, the better it is for the organization, because so often we get, we get subjective because of a really small sample size. And the example that I'll use is like, you know, you know, that whatever decision you make, probably 10% of your workforce is going to get pissed off. That's just I can get I don't have to, but I bet you probably do that there's, there's like 10% of workforce is going to hate whatever you do, because they just don't want to change 10% of the workforce is probably going to be like, great, I love this, regardless of what you do, even if it's a terrible, and then you've got the you've kind of got the the middle. It, that's it. But that's so often we focus on the 10% that are going to be negative about it and at the detriment of everybody else, you've got to think bigger, right? That was a totally non quantitative gut feel, you know, Empath describing that. SoBen Waber:
I'm gonna repeat that. Again, I think that that's, this is sort of the approach. And again, I think that you know, even for the non data scientist, folks, right, like that approach, like will serve statistically will serve us very well, I think is the important thing. And I think that this, again, thinking probabilistically is so important, because we could have all the confidence in the world, then when it comes down to that ultimately, the outcomes of many, many things, particularly regards to work are probabilistic. And you can do all the right things, you could have all the right data and even show you all the right things and that there's still a chance that things won't go like you wanted them to. And that that's okay. Right. It's I think embracing that is actually extremely helpful, not just when we're making decisions when it comes to you know, how we organize a company, these larger strategic decisions, but even when it comes to individual decision, you know, smaller things that we make all the time that sometimes you're just in the right place at the right time. And that that that if you accept that That might be why a particular program was successful, then you're going to constantly question, are we making the right calls? And not just think that oh, well, listen, I've been really successful. So I must know what I'm doing. Right? Like, no, you actually, like probably do a little bit like you probably did some hard work, but also you probably got lucky, like you were being honest. And if you have that, that perspective, then, right, you're never gonna drink your own Kool Aid too much. And you're always gonna say, like, Alright, let me try to go check this again. And it's, you know, you know, a lot of times, you'll still go with what you you thought, but that hopefully over time, you won't keep repeating things that actually are ineffective. And you just, you never, you never figured them out? Because you weren't interrogating them enough? Yeah,Kyle Roed:
absolutely. Absolutely. One, so I think this has been just a fascinating conversation, but one of the areas that I maybe want to dig into a little bit is, so you know, I, I love the theory of love the, you know, kind of the approach. How does humanize, think about taking all of this input, and then giving it to a company that is actually actionable? Yeah, that actually gives them you know, some some action items or some things to, to think about or work on, you know, how do you how do you take the big data and make it make it something you can do somethingBen Waber:
with? No, and this, this was really, obviously getting coming out of a research, a lot of it was, well, what metrics actually matter for predicting outcomes? And that's sort of like the first step. The second step is what are the metrics that how can I give this data back, to focus on where they are today, so that they can understand what these metrics are, and they can take action on them to your point, right. And that has resulted honestly, in us, like, we don't know, all metrics that we can calculate, or even that are predictive, because some of them are just, they're still hard for folks to get their heads around. Like, as some of our customers get more sophisticated, we do show more and more. But, you know, essentially, what we did is universe, we have dozens of dozens and dozens of peer reviewed articles that we wrote about how here's how it behavior acts relates to an outcome we care about, again, retention, performance, decision making speedy, what have you. And so then we did is we built this model, we called the organizational health score. And frankly, the important point about it from my end is not the score itself. It's this. There's actionable behavioral metrics that drive that. And that's actually what people pay, like you should pay attention to. Right. But so what we do is we take all this data, and we're able to combine it to understand how is work happening? How are people collaborating? And we show you all right. So, you know, high level, you know, for example, this part of the organization has a performance problem, say, okay, like, what is that? And you dig into it? Oh, it turns out, actually, it's manager visibility, it's how much time managers are spending with teams, and it's specifically in this region. Right? Like, that's where it's the biggest problem, right? And so essentially, we're able to do is automatically flag like, here are things that because we have such a massive data from across many companies, like the ideal to be clear, is you feed in KPIs in your business, and then so you can say, a 10% change in this outcome leads to why you know, in this behavior leads to y percent change in this outcome, like, that's the ideal, but we've tried to do is even without that, because, you know, probably a lot of us know, most organizations are pretty poor KPIs for across the organization, right? Like, what's the performance of Team X, like, unless it's sales, honestly, like, you don't have great KPIs. And so what we're able to do, though, is be able to say again, automatically, here's where you should focus your attention. And, you know, probably a third of the time, there's some contextual explanation for why something really concerning is going on. And I'd say that's what out of the box are analytics are really good for is saying, what, what things are almost certainly bad, right? At a team level. And so you focus that in because we can have data from 100,000 people. And I can tell you, you know, here are the couple, you know, a couple dozen teams, where they're really big at certain here, the specific behaviors that are concerning. And so then that's your cue to dig in more, right? That's not to say, again, that our algorithms are like the be all end all. That's where, okay, we see that Team X again, maybe managers have spending more time a lot of time with their teams. Alright, well, it turns out that the products on fire and we're having, like lots of meetings, trying to figure out strategically what to do. Great, okay, like, that's, that's fine. Now you understand what's going on there. But again, but of course, it could also be like, here's an actual, like, you know, leadership problem, we need to run people to training we need to, again, there's all these things you can think of. But the issue is that though, what we're really trying to do is expose these things are normally very opaque at macro scales. Like if you're on that team, you know what's going on. But if you're running HR for a large organization and trying to figure out what actually are the most critical issues, the things that you normally rely on are, you know, essentially self report, it's, you know, people say these things, but importantly, some of these things are stuff that is very, very challenging or impossible for individuals to be aware of, like so good. One is weak ties is an interesting example. So this is these are People tell you how we define it. Although the numbers are robust to definition, people, you spend on average five to 15 minutes with in one on one communication, the equivalent of that a week, right? And a good way to think about these people imagine someone who you're in a four person one hour meeting with once a week, right? So they're not on your team, but you like know each other, and you like, sort of like, we're a little bit on something together, but you're not working very closely with each other. Right? So one thing we've seen with the rapid shift to remote work during the pandemic, is those ties have fallen off a cliff, like there is, you know, 20 plus percent decrease in the number of those ties, you know, again, across the board. The thing is, is it people don't feel that at all, if you ask people in surveys, which we've done, like, how do you feel like your communication networks? Oh, I think it's the same. Because you think about, like, what percentage of your day are those weak ties, and it's very, very small, right? Because again, by definition, these are not like the people you work with on your team. And so that falling off by 20%, that's like changing how like 1% of your week, like one to 2% of your week is spent, you don't feel that you don't feel that change. But that if you look at Milestone attainment, large projects, if you look at innovation, those ties are disproportionately important for those outcomes. And so really, what we're able to do is surface things that yes, like some of these things on surveys, if you explicitly targeted, you can figure them out. But there's a lot of these things that just individuals, we just don't, because we're we don't have that macro perspective, but that this kind of data can make that clear. And so we've tried to do is use terminology that people understand, again, have lots of actual, you know, citations that people, you know, understand, we say this important, here's, you know, all these different cases, and actually academic articles about why it's important. The idea, of course, is to get KPIs. But that, again, even in the absence of that, we're able to do some pretty, pretty powerful things.Kyle Roed:
That's really fascinating, you know, and that weak ties example, you know, I just think about, you know, that you've just quantified what we've all been trying to describe as to why the workplace is important, why the physical workplace was important, you know, or, or the accidental, you know, interruptions and, you know, water cooler talk. So it's that, that weak ties, that was kind of a lightbulb moment for me to think through that. Yeah, and when you don't get that in your workplace, it makes a lot of things more challenging. So you're interesting. All right. Well, this just a fascinating conversation. So you know, I'd encourage anybody who wants to learn more, go, go check out humanize it's human yz. And, you know, get connected and learn. I'm definitely gonna take it from here. But I do want to shift gears and go into the rebel HR flash round, I'm fascinated to hear these these responses. So questions ready? Yeah. All right. Question number one, what is your favorite people book,Ben Waber:
my take a little bit of liberty with this one. My favorite people book, I'd argue to people book, it's called a small worlds by Duncan Watts. And testing out the books I was 20 years old. But it's a really fascinating look at networks and the different aspects of them, and how small changes at the individual level have these really systemic impacts on the network as a whole. And a lot of it is focused on people. But there are things like analysis of internet networks and other kinds of networks as well. But I think the focus is on people. So I will call it a people book. But this is one of the few books obviously I read it a while ago, but it really dramatically changed my perspective on sort of what what actually matters when it comes to systems. And I think, again, especially for organizations, I think it's a very easy, I think it's a very, you know, general consumption book, but um, it, it gives some very well researched perspectives on on people's, I highly recommendKyle Roed:
that. Very cool. Very cool. Question number two, who should we be listening to?Ben Waber:
A couple people I have actually over the past, let's say, like six months, gotten more and more interested in the whole area of AI ethics, obviously, because it's related to our human eyes. But I also just think it's this really emerging area that I think is incredibly important for all organizations that deal with data in any capacity. Right in an argue that people data is some of the most sensitive data out there. A lot of the folks in this space. And there's two people I want to call it in particular, Kathleen Creel who's over at Stanford, and also you film a Judwaa, who's at UNC. Looking at Two different aspects of AI ethics, but they are focused on people data. within organizations, there's folks like JoyBell belong Winnie who's out of MIT who's done a lot with just general Seidel data looking at like things like facial recognition. And that's, that is very important, but it's not as relevant as some of the work that Kathleen and Ifeoma have been doing, where Kathleen has really looked at a framework around how organizations use algorithms to make people decisions, and looking a lot at algorithmic hiring in particular, which obviously, is a very hot topic right now. Um, but I think her research, it's this very rigorous framework for how organizations to think about, like, how should we apply ethically, these kinds of technologies? What should we be looking at in terms of evaluating not just their efficacy, but also their pitfalls? And how should we constantly question these things, and it's just really fascinating work there. Whereas a FOMO, June, we're over at UNC has been doing fascinating work on people data within organizations. And what are some, again, obviously, this is very directly relevant to humanize. But she's focused on a lot of things like Fitbit data, you know, like when companies give out Fitbit for wellness programs, like what, what happens with that data? And it's, it's sort of concerning, but fascinating, as well. And I think the implications for folks, you know, frankly, like humanize, but also other companies doing this sort of thing is not just what are your intentions, but really trying to think through what are the implications of how you're collecting this kind of data. And so I would, again, highly recommend, you know, folks, like, you know, there's, there's, there's more talks than ever before, from academic departments available for consumption. So, you know, I highly recommend checking out some of these folks. Again, Stanford, Stanford High has a lot of these talks, so does, again, if home is program over at UNC, I think it's called decision. AI and Decision Sciences, I think, are two off the top of my head, but I tweet about lots of those talks. So if people like talks, but that's what I gravitate towards, honestly, is the sortKyle Roed:
of most fascinating and I do think, you know, what, whether whether you like it or not, it's it's going to be a part of hrs job, you know, in in an RD is in many, many respects. So yeah, it we've got to be mindful that entering the workplace. Last question, how can our listeners connect with you?Ben Waber:
Yeah, I mean, I think it's so as you pointed out, there's a humanized website where you can can reach out there, I'm, I'm, I'm also on Twitter. So it's just at the WebVR. And yet, feel free to reach out to me there and I will probably respond.Kyle Roed:
Yeah, absolutely, I think probably worth worth of follow. And I'm definitely going to be be following along, you know, really appreciate the work, the approach. And I think that my challenge to everybody who's listening right now is if you're not thinking in these in these concepts, or or thinking about how these sorts of things can help you make decisions in your organization, you owe it to your teams to start to get educated and learn. So, Ben, thank you for sharing your your wisdom with us today. And thank you for the work.Ben Waber:
Thanks so much.Kyle Roed:
Take care. All right. That does it for the rebel HR podcast. Big thank you to our guests. Follow us on Facebook at rebel HR podcast, Twitter, at rebel HR guy, or see our website at rebel human resources.com. The views and opinions expressed by rebel HR podcast are those of the authors and do not necessarily reflect the official policy or position of any of the organizations that we represent. No animals were harmed during the filming of this podcast. Maybe