Going from fragile to agile

| Podcast

Rapidly changing business conditions are demanding more flexible organizations, say McKinsey principals Aaron De Smet, who coleads our organizational-design efforts, and Chris Gagnon, who oversees our Organizational Health Index (OHI) research. In this episode of the McKinsey Podcast, they talk with McKinsey Publishing executive editor Luke Collins about their work, explaining why agile companies outperform others, how organizations are reacting, and what leaders should do. An edited transcript of the conversation follows.

Want to subscribe to The McKinsey Podcast?

Podcast transcript

Luke Collins: Hi, I’m Luke Collins, an executive editor with McKinsey Publishing. I’m in Miami with Aaron De Smet, a principal in our Houston office and the coleader of our Organization Design Practice globally. With Aaron is Chris Gagnon, a principal in our New Jersey office. Today we’re discussing one of their favorite topics—agility. More specifically, we’ll look behind the issue of organizational agility, which has been the subject of several recent McKinsey Quarterly articles. Aaron, Chris, thanks for your time today.

Chris Gagnon: Nice to be here, Luke.

Aaron De Smet: Thank you.

Luke Collins: So why is this a topic now? Companies have obviously always had to adapt to the environment in which they operate and change over time. But it seems like, suddenly, the need to be agile is more important than ever. What’s happened?

Aaron De Smet: The pace of change has gotten a lot faster. It’s just a lot more intense.

Chris Gagnon: Look at the statistics on how long companies last in the Fortune 500. They’ve dropped so precipitately that all of our clients are talking to us about the pace of change. Doesn’t matter industry, doesn’t matter geography.

Aaron De Smet: CEO tenure is way down.

Chris Gagnon: Way down.

Luke Collins: And is that a result of . . . I mean, we can run through the gamut of reasons—digitization, increasing globalization, companies looking well beyond their traditional borders, short-term thinking. Is it just a confluence of factors?

Chris Gagnon: Well, I’ll tell you, I’ve tried to lay out the arguments for why that is, in a number of talks, and I get short-circuited every time because as soon as you say it, everybody nods their heads. I mean, everybody knows and feels what’s going on—and it’s all the things you described.

Luke Collins: So against that backdrop, McKinsey’s been doing a large amount of research, utilizing a big tool that we have—the Organizational Health Index. Tell us about some of the research that’s been undertaken and what that seems to be finding about how companies should be operating.

Chris Gagnon: I lead the team that runs the OHI globally, which is a fascinating database. I would argue it’s the largest database in the world on how companies run themselves. People suggest things, all the time, that we might want to measure or add to it, and most of them don’t actually add much to our predictive ability or our ability to explain things.

The true answer here is, we’re looking at this more because clients suggested it. And when we created a simple index for speed, clients said, “boy, we need to look at speed. We need to pay more attention to it.” At the same time, or shortly thereafter, we said, “well, if we’re measuring speed, we ought to measure stability as well.” We created an index there.

What did surprise us is that we kind of assumed that speed and stability would be antagonistic. And there were, sure, a set of companies that are so fast and crazy and chaotic, you know, that you think of them as typical start-ups. And there are companies that are bureaucratic and so stuck in their ways that not much changes. But there was this really neat group of companies which had both.

Luke Collins: And in fact, that neat group of companies seem to be arguably the best performers.

Aaron De Smet: Oh, by far. I mean, the thing Chris said that’s really noteworthy is that we can test things all the time. And if you just want to find something that correlates with something else, that’s easy. If you want to say, “this actually is predictive and explanatory of what we find, above and beyond what we already know,” almost nothing turns out. People say, “oh, test this.” Nope. “Test this other idea.” Nope. “Well, does it not correlate?” Well, it correlates, but it looks like it’s just correlating because good companies do lots of good stuff.

But this one [speed and stability] is different. This one looks like it’s actually driving completely different behavior and experience and performance levels. If you emphasize speed—and we have some very specific questions on that—and if you at the same time emphasize stability—and we also have some very specific questions on that. . . . If you can do both, you have hit a sweet spot that will lead you to far and away outperform anybody who’s just doing one or the other.

Luke Collins: And of course, at this point a lot of people will say, “well, hang on. How can I have both? I can’t have my cake and eat it too.” So talk to me about the definition that you place on organizational agility. Because McKinsey seems to be defining it in a very particular way, at least in terms of how organizations should think about it.

Aaron De Smet: Well, take speed as a proxy for how responsive, nimble, and dynamic the organization is. And speed is the best proxy we have to measure it. Just go fast, in lots of ways and on lots of dimensions. Go fast. Going fast does not mean that you have to be completely chaotic and unstable.

Chris Gagnon: And, in fact, what we saw was that the best companies create what we’ve started to refer to as a “stable backbone.” I don’t want to seem like we have all the answers here. We’re working this live with our clients and looking at some fascinating companies and what they do. But if you were to draw commonalities, you’d say that what that stable backbone seems to do, most of the time, is take care of people. Right? It answers the “what’s in it for me” questions, like “how am I going to get evaluated? How am I going to get paid? How am I going to get developed?” And the best organizations seem to say, “don’t worry about that. We’ve got that taken care of. You have a home, and that home doesn’t change. Our strategies may change all the time. The things we do to be fast and responsive to the market may change. But your home’s not going to change. Relax.”

Aaron De Smet: What you see in some really agile companies is that, first, they have a matrix structure, which is kind of important. There are a lot of companies which really don’t like the idea of a matrix, and so they pretend they don’t have one. But almost all big, complex global companies have a matrix, whether they want to admit it or not. Which means, if I live in this company, I have a sense that I have more than one boss. I may only have one real boss who hires and fires me. But there’s a sense that I answer to, and am accountable to, multiple people on multiple dimensions. And in many cases, we’ve seen that agile companies give people a permanent home, say, in a function. But then they’ll have some other dimension be the one that’s giving them most of their day-to-day marching orders.

There’s one company we worked with, and we totally redesigned everything. They had a very complex matrix; we made it very simple. It’s a functional organization but also has product lines. The product lines are the most important, primary, dimension, and they “own” virtually no people—over 99 percent of the people report into other functions. The product lines, however, have most of the decision authority. They get to set pricing. They get to say what people work on. They get to make capital-allocation decisions. But people’s homes are in the functions, although what they work on is things important to the product lines.

Chris Gagnon: It leads you to an idea that I’m scared to death to utter on a podcast, to be recorded for all history. We all grew up learning from [Alfred] Sloan that structure followed strategy.1 And, of course, that’s still got to be true. Your structure is designed to deliver on your strategy. But I think it’s true in a different way today. It used to be if you said, “we want to get really good geographically,” you’d go to a geographic organization. If you wanted to get really good at product, you’d go to a product organization. Today, strategy changes too fast for that. And so I think what these organizations have in common is that they try to provide a stable organization that can take hold in a flexible, dynamic, ever-changing world. So you don’t have to reorganize, you know, every time you think your market strategy needs to change.

Luke Collins: And the way that you’ve been describing that, Aaron, is with a smartphone analogy. The idea that the company fundamentally has an operating system that evolves slowly over time, incrementally. And then on top of that, you’re plugging apps into it. You’re sort of doing more dynamic things on top of that operating system.

Aaron De Smet: Most of the things you do with your phone are through apps, and those change all the time. The old Sloan way, just as Chris was saying, was figure out the strategy, then design that and hardwire it in [to the organizational structure]. [But] if you try to hardwire everything you want your company to do—or your organization to be able to execute—into what is, effectively, the hardware and operating system, you’re going to find that you’ve just designed a phone, or an organization in this case, that’s obsolete by the time you’ve implemented it.

What you need is just like a smartphone. I’ve got a smartphone, and it can do new things that I hadn’t even dreamed of—the app didn’t even exist, the idea for it didn’t even exist, when I was designing the phone. But now I have a phone, somebody has an idea. They can design an app in a week or two, and I can suddenly do something on the phone that I never thought I needed to.

Chris Gagnon: These apps are really the flora and fauna of modern corporate life: initiatives, teams, special projects. I think, in some ways, they used to be thought of as in the purview of, you know, upper-middle management. Now you’ll find bellmen in hotels on teams trying to improve the customer check-in experience. This is how you keep up with a fast-changing world. The false goal of alignment—this idea that you’d set a magical strategy, and you would design an organization, and design jobs and compensation systems, and everything would fall in line in this very engineered way—just isn’t an idea for the modern world.

Luke Collins: Is it confronting some large organizations, somewhat tradition-bound et cetera? Managers and executives working for those organizations?

Aaron De Smet: Oh yeah, it’s hugely problematic for them. You probably need something that’s stable, that you can rely on. But if you want predictability and you want things not to change, forget it.

Chris Gagnon: You were born in the wrong century.

Aaron De Smet: Yeah, you were born in the wrong century. I remember doing a research report on what’s really changed as a result of the Internet. This was still in the Internet boom in the late ’90s. And what I came back with is, “look, I know there’s lots of Internet companies. But in some ways it’s kind of silly. Yes, everyone’s going to need to use the Internet, for sure. But everyone uses the telephone now, and there are only a few telephone companies.”

The only thing the Internet’s really done, besides [foster] a few companies that really will be Internet companies, is dramatically lower transaction costs and interaction costs. You can now interact in different ways, and share data and information globally, in a way that you couldn’t before. That’s different. And that has increased not only the speed but the nature of interactions, to create this emergent complexity that is actually looping on itself and actually fueling more complexity and more emergent properties that you could’ve never predicted. More fast-cycle turbulence. So if you want that kind of stability that you’re talking about, which is things aren’t going to change, forget it.

Chris Gagnon: We went to one of my favorite companies, that we can’t name, and said, “how do you innovate so quickly?” And they said, “well, how do you measure our innovation?” And we said, “well, look—this, and this, and this.” It was, like, four things. And they said, “oh, well, you missed the 32 we started that didn’t work.” The tolerance for inaction in these places is just nonexistent.

Luke Collins: And we hear a lot about “fail fast,” and it seems like these companies that are in that sweet spot. . . .

Aaron De Smet: By the way—one thing, though. It’s “fail fast and learn from it.”

Luke Collins: It’s not just “fail from fail,” right?

Aaron De Smet: Fail fast and learn from it. And sometimes fail multiple times before you get it right. And make smart decisions about when to cut it and say, “this just isn’t working.” Or “we should try again.” A lot of companies, when they say, “we need to embrace failure”—well, I’m not sure it’s embracing failure so much as you’re willing to try things that don’t work the first time. And that’s different.

Think about an athletic team—any athletic team. What if they never practiced? You realize we hire people and we tell them, “you just have to show up and perform.” You never get to try; you never get to just try it out in practice.

Chris Gagnon: It’s game day, every day.

Aaron De Smet: It’s always game day. And you get people who inherently are like, “well, I’m not going to try a new play on game day.” Right? “The brand-new play that might be brilliant, I would have to try it in practice first. Make sure we knew how to execute it.” So you have companies that really have every day as game day, don’t try the new play that they [could] scale up immediately, and then fail. That’s a massive failure. You have to fail in small, contained ways.

Chris Gagnon: Like you, I’m deeply uncomfortable with this language about embracing failure. I never experience failure as anything but at least a little bit unpleasant. But there is something about creating a culture in which there are experiments in a scientific mind-set. You’re learning something. You’re trying to drive through. But you’re telling the truth about what happened. And I think the most agile organizations really value the ability to tell the truth. In the bureaucratic ones, you know, if you get assigned to a project, you better make it work, because you and the project are the same.

Aaron De Smet: I’ll give you a great example of a simple concept that can either play out really well or really poorly, depending on if you’re agile or not agile. Chris just gave one about embracing failure. But I’m going to give you one on one of Chris’s favorite topics, which is role clarity.

The OHI measures 37 behaviors. And whenever we look at these, one of the behaviors that always matters a lot is role clarity. Now, a bureaucratic organization, when they hear “role clarity,” what comes to mind is a job description. A detailed role description, where I’ve written down everything I want you to do and be accountable for.

Chris Gagnon: Which is therefore wrong the minute you wrote it down.

Aaron De Smet: It’s the completely wrong way to think about role clarity. The role clarity we measure is behavior. It’s you talking to someone else about your role and their role and how you work together. It could be a subordinate. It could be a peer. It could be a supervisor. It’s behavior. Role clarity is clarifying “what are you accountable for? Where should you spend your time? How should you do your job? How do we work together? What do you get to decide on your own? What do we have to decide together?” That’s behavior. In the OHI, the 37 practices are measured on frequency. We’re saying, “you have to clarify roles a lot.” And the reason you have to clarify them a lot is they’re changing all the time.

Luke Collins: So let’s run through some of those. You mentioned there are 37 categories that the OHI examines. Of those, there were ten, a dozen, that seemed to really stand out for organizations that performed well on organizational agility. A couple of them were role clarity, yes. Top-down innovation was another that seemed to be big. The ability to capture external ideas.

Chris Gagnon: Let’s talk about those two because I think they really travel together. One common issue we see in companies that pursue this is they only focus on bottom-up innovation. Management’s [saying], you know, “I’ve got to let my hands off the steering wheel a little bit. We’re going to let everything bubble up.” That can pretty easily lead to chaos and a bad situation for the people you’re asking to innovate, because you haven’t told them what you’d like them to innovate around. Telling people, “we’d love to reduce cycle time, we’d love to improve our customer experience,” is different than saying, “and here’s exactly how you will do it.” In the absence of that top down–driven guidance, innovation tends not to make as much performance impact or strategic sense.

External sensing really plays into that. We started all this by saying, “the world’s changing fast. Customers are changing fast. Competitors are changing fast.” If senior management doesn’t have a good external view, the direction they’re likely to give to innovation is likely not to be very good. It’s likely to be too internally focused.

Aaron De Smet: And when we say “top-down innovation,” it includes just basic improvements, too. It’s not always these breakthrough innovations. It’s any kind of improvement. But the top-down part means the decision makers who matter have agreed what our priorities are and how we’re going to resource this. Without that, you often get initiative overload—everybody working on everything. And it gets to be chaos, quickly.

Now, if you really are a start-up with a really small organization, it’s all very easy to manage because everybody knows everybody, and they know exactly what’s going on. But when you get these big companies, people are so willing to launch a new initiative without ever actually having figured out what it’s going to take or how many resources it needs. And then you’ve got everybody kind of working on everything. I have clients where people in really critical jobs—where the job they have actually is a true full-time job—have a second full-time job doing initiatives. And they end up doing it all badly.

Luke Collins: You’re not doing any one thing well. But this is slightly counterintuitive, too, because there seems to be a popular notion that innovation is “let’s brainstorm. Let’s crowdsource. Let’s just throw everything against the wall.”

Aaron De Smet: That’s fine—I think that’s a great thing to do. But then somebody’s got to pick. On some level, strategy is about choices. And if you are not making choices about what you’re going to really do, and really commit to, and really resource, and really spend leadership time on, then you’re not really making choices. And top-down innovation is as much about what you’re not going to do.

Chris Gagnon: Yeah. When we look at the highest-performing companies, there’s four clusters—we call them “recipes”—that really do manage in very different ways and feel very different. One of my favorite things to do is to go into [a company] and suggest something that doesn’t fit with its model. Mediocre companies always say, “boy, that sounds great!” And the best companies say, “nah, that’s not what we do.” You know? That notion of making choices, knowing who you are.

Aaron De Smet: Well, I mean, we say there’s these four cultural recipes that really work. There’s actually thousands of recipes—it’s just that all but these four don’t really work.

Luke Collins: Well, this is encouraging, too, on a couple of levels. One is that as a big organization, a big institution, there’s a bunch of inherent strengths to running an organization like that, which are completely applicable today. More so than ever. And they’re not incompatible with being agile in any way, shape, or form.

And then, at the other end of the spectrum, I guess you have the start-ups that are envied by a lot of major companies because of their degree of agility, because of how fast they’re running. But they have the challenge of implementing the foundation that we talked about. The operating systems. At some point, you’ve got to stop running at a hundred miles an hour and put some processes in. You have to put in some of the foundational elements that large organizations have.

Chris Gagnon: There’s a great TED Talk—and I wish I remembered who gave it because they deserve the credit—about if start-ups are so great, why do most of them fail? Some of that is market risk. But some of it is that you can really predict how well, and how long, that management model’s going to work. And we have clients right now that have gone through hypergrowth through that start-up phase and they’re hitting very, very predictable problems in scaling their management approach. So you can get to these problems either way.

We’re always looking at new experiments because we find them fascinating. You know, “holacracy” is a fascinating concept that’s getting talked about a lot. And there’s a lot in there to really like. You hear about it—it sounds too good to be true. Well, when you actually look at it, you realize that to run a holacratic company requires a whole different set of rules and structures that maybe don’t look like traditional structures, but they’re rules and structures. There’s no free way to organize tens of thousands of people.

Aaron De Smet: For those who don’t know it, there’s lots of new ways of organizing that aren’t traditional. And one of them is this idea of holacracy. Holacracy completely throws out the org chart and bosses and this idea of hierarchy. What it does do is say, “everybody’s part of some team.” But the teams, depending on how they work, look a lot like committees.

What they’ve really done is take on one structural element which most companies overindex on, which is the management hierarchy. The lines and boxes. Who reports to whom. Who is whose boss. They’ve completely tossed that out the window, which people love. “Oh, yes. Toss the hierarchy out the window!” What they’ve replaced it with is, essentially, a massive overlapping committee structure.

Now, [holacracy] has democratized the workplace a bit because the people on those teams have a vote and have a say in the decision. But if it doesn’t work exactly right, and it gets too big and complicated, it starts looking an awful lot like Congress. It looks like these slow deliberative bodies that are sort of death by committee. And if you look at holacracy, most of them, when they get too big and too complex, they have—in a different way—the same bureaucracy problems that a management hierarchy has.

Chris Gagnon: You know, when I hear you say that, I realize we probably skipped over one belief that’s so fundamental to us that we ought to lay it out here. This is not about style. It’s not about culture. It’s not about people acting a certain way and everybody’s got to look and talk like that. It’s about how you run the place. It’s about management process. And it’s all measurable.

So while there are no easy answers, I think there’s no other area of business organization where senior leaders have been trying to manage something, without measuring it, for so long. And you no longer have to. You can measure all of it, and you can decide what’s working and what’s not working.

One of my favorite management maxims is “do more of what’s working and less of what’s not.” You can learn from others, as opposed to feeling like every situation’s truly unique. Every situation’s unique, but not that unique.

Aaron De Smet: The useful way to think about culture is “how we do things around here.” Now, there’s a bunch of really important things that are much deeper—at a much deeper, fundamental level—when you look at culture from values and thought systems and beliefs. But those are hard to access and they’re hard to act upon. The “how we run the place” version of culture is immediately actionable. You know what to do and you can measure it.

Chris Gagnon: It’s not a question of right or wrong, it’s a question of useful. I was asked to address a group of executives, and I’d heard, coming in, that their mind-set was that culture was equal to values. So I cheated a little bit, and I put up a slide with eight values on it. And I said, “I don’t want to embarrass anybody. Would anybody admit whose company values these were?” Two-thirds of the room put up their hands. And, of course, we had taken all their values statements and mushed them together, because they’re just not that different. And I’m not arguing that they’re not important—I think they’re critically important. But it’s hard to distill what’s different about them in a way that makes them actionable.

Aaron De Smet: Because [those values are] excellence, integrity, respect for people, collaboration.

Chris Gagnon: Customers.

Aaron De Smet: Innovation, customers. And you’re, like, “well, all my clients have those values.”

Luke Collins: Arguably common sense, to some extent. So then it comes back to the question. . . .

Chris Gagnon: Why don’t people include common sense, right?

Luke Collins: It goes right to the question of how we do things around here. Which can be very unique by company.

Chris Gagnon: Well, a story. I was doing a discussion about this in front of a company, and I had been pulled in with very little prep time. We were talking about this notion of these recipes, and I used three big, iconic companies. I think Coke, P&G, and GE were examples of parts of it. Not perfect, but you know. . . . And this CEO gets kind of a funny look on his face, and he raises his hand and says, “I worked at those three companies.” So now I’m fairly exposed as the facilitator, because he’s got all the authority in the world. And he said, “you just explained what I’ve never been able to tell people—why good management at one [company] wasn’t the same as good management at the other. And if I acted the same way at one place as I acted in another, I never would have made it.”

Aaron De Smet: The other thing this brings up is that we often treat, and our clients often do too, organizational structure and culture as different things. What we’re saying is, actually, no—this is all part of the holistic architecture of a healthy company. You have to treat them as different design levers, but they’re all trying to build a high-performing company.

One of the worst things, when we think about organizational design, and when our clients do, too, is almost immediately to go to structure. Then they not only fail to design the other elements, like decision processes and linkage mechanisms and culture—culture in the behavioral sense of “how do we behave, and how do we run the place?”—they also do things like cascade targets and metrics and objectives down the hierarchy. Then we’re shocked when, at lower levels, nobody’s collaborating, because they all have completely different metrics of what they’re optimizing for. That should not be a surprise. That will happen every time, if you do it that way. And most of our clients do. There’s lots of conceptually easy things to do differently that most of our clients don’t do. They’re just not thinking about this in the way that will help them be more agile.

Chris Gagnon: Well, you know, we’re here to talk about agility. And earlier we brought up project-based organizations. I find it hard to get away from that notion. There may be a better way to articulate the model for big companies than project-based [ways] over time, but for now, I think that’s a placeholder—that you use projects to develop new products, to bring things to market, to approach customers in different ways, to reduce costs. Everybody’s life feels like a series of projects. And I think one thing that a well-designed organization does, in the broadest sense, is enable people to react to that really naturally.

If we go back to look at these agile companies and approach them and say, “hey, we’re going to change this element of our strategy,” in an agile company people can say, “great, I’m in. How do we start?” They know the backbone’s there to take care of them. They’ve seen this movie before. They know they’re going to see it again. It’s part of their job expectation, as opposed to “whoa, wait a minute.”

I think it gets at this notion we’ve been telling each other forever—change is hard. And there’s some truth to that, I get it. But change that’s good isn’t as hard as change that’s bad. I think a really important thing is that people enjoy being in these agile organizations.

Luke Collins: I think that’s a great note to end on. Thank you so much for your time today, Aaron and Chris. [Readers can] go to mckinsey.com to read our extensive coverage of this issue, as well as our broader examination of how leaders should be thinking about organizing for the future.

Explore a career with us