Leading to chaos: A conversation with General Stanley McChrystal has been saved
Limited functionality available
As a former US military leader in Iraq and Afghanistan, General Stanley McChrystal knows firsthand what it’s like to operate under chaotic conditions. His wartime innovations in leadership and organization could also have broad applicability for businesses as they strive to deal with kaleidoscopic change and uncertainty.
In 2008, as a brand-new Marine Corps intelligence officer, I quickly found myself in Iraq trying to make sense of both a massive military bureaucracy and a distributed, adaptable insurgency. Very quickly, I became aware of an organization called the Joint Special Operations Command, or JSOC. Where my team would struggle for weeks to find one low-level bomb maker, JSOC seemed able to find even the highest-level targets nearly every night. It seemed to know everything about the country, be able to find anyone, and, more importantly, then be able to act on that information. It was, quite simply, one of the highest-performing organizations I have ever seen, and no small inspiration for why I eventually began working in business research. What I did not know at the time was that the high-performing machine that I saw in 2008 was the product of a massive reorganization and reorientation begun a few years earlier by General Stanley McChrystal.
McChrystal is a retired US Army four-star general who has led organizations in some of the most trying situations imaginable, from Special Forces in Iraq to all NATO troops in Afghanistan. In his recent book Team of Teams, he describes how exactly those difficult situations forced him to change as a leader and, in the process, find a new way of structuring and leading organizations in the chaos of the modern technological environment.1
Read Deloitte Review,
Even before his high-profile tours as a commander in Afghanistan and as director of JSOC, McChrystal was known as an innovator and a rising star in the military. In 2003, as the situation in Iraq began to deteriorate, he had stepped in to lead JSOC, home to elite forces such as Delta Force and SEAL Team VI.2 McChrystal discovered that, even for such well-trained, well-equipped forces, something more than just dazzling technology was needed to help defeat decentralized terrorists. Amid some of the darkest days of the Iraq war, he set about systematically changing how JSOC was led, organized, and operated.
I recently sat down with him to discuss how technology forced JSOC to change, some of the science behind what made it so effective, and how other organizations can benefit from the same lessons.
Joe Mariani: Could you start by briefly describing JSOC’s role in Iraq? I am sure we all have movie-fed visions of stealth helicopters and bearded men in camouflage, but can you tell us what you saw as the role of the organization?
Stanley McChrystal: We started that war doing raids, and that is what we were good at, but the ability to do raids becomes commoditized pretty quickly. Every organization could do a raid if it knew where the bad guy was and you told it to go there. It might not be as elegant, but they could easily get the job done. We realized pretty quickly that raids were not our product. Our product was our network, which knew more than anyone else in the region, not just about the enemy but about our own side. We influenced more people, influenced more organizations, and moved and coordinated faster. That became our value-add. People kept thinking it was the operators going through the door, but, no, it was knowing where the enemy was and constantly adapting.
JM: At first blush, it seems like the key to achieving that speed and knowledge is just to increase efficiency, to do things faster than anyone else. In your book, you talk about Frederick Winslow Taylor as the origin and exemplar of some older ideas of management that emerged from assembly lines and other late 19th-century advances. Specifically, Taylor’s goal was to reduce the number and complexity of tasks that workers needed to perform, so that they could increase efficiency. While efficiency is still a positive goal, you also describe how these ideas of simplifying tasks and centralizing authority may keep us from responding effectively to the rapidly changing environment that technology has created. Can you speak a little bit about that tension?
SM: I was always an enthusiast for technology. I am an enthusiast for doing things efficiently. So when I was a young captain in 1982, I was a company commander in a mechanized unit, and I went and bought a Radio Shack Model III computer. It cost me $4,800 for something with no hard drive. I took it to work, and then we took it to the field. We built a wooden box, mounted it in a tracked vehicle, and took it to a large exercise. All it could really do was word processing, spreadsheets, and a small database capability. But we started by word-processing all of our orders and formatting them, so that when you got details, you could crank them out really quickly. And we amazed the exercise evaluators.
“I was always an enthusiast for technology. I am an enthusiast for doing things efficiently.”
However, with technology like that, the thought was simply, “How are we going to make ourselves better able to control our organizations?” We didn’t really think about what was going to happen when everyone has that same technology. That was when things really started to change, with the proliferation of those communications technologies. First, it was only those who could afford it, which made its impact limited. But, suddenly, that technology became widespread, and instant communication became available to everyone. That is when I think everything changed for JSOC and for businesses.
JM: Did you find that the new communications technology made you better able to respond to this new environment, or was it a larger challenge than a help? You are leading virtually. You cannot touch people on the shoulder.
SM: It all depends on how you use it. When we got all this technology, common sense says, you have got to use it. And then, as you start using it, you start thinking, “OK, how are we going to use it?” The first thing you do when you have the ability to connect everyone is to say, “I am going to control every operation from the center, because now I can see and hear every operation.” That is what we were doing in late 2004. We were working as hard as we could, not sleeping, just going as hard as we could. Then you realize that the quantity of information [that comes from everyone being] connected overwhelms the center. There is no way that the headquarters is going to know everything and have granular knowledge at the speed things are happening. So the first attempt you make is to bring it all in, process it, and be this supercomputer center—and you realize you just can’t do it.
Then you say, “OK, what am I going to do?” Suddenly, you realize that communication goes both ways. This ability to communicate now, instead of just using small groups as tentacles or sensors pumping information to the brain, allows you to put the brain out there. Let people operate. It is too fast, too complex to process it all here in the center, so we process it across the entire organization without really controlling that process. We are letting everyone think, letting everyone have the information, then they can act locally. That started to work really well. That was probably in late 2004, and partners from across government started playing into it, and that, of course, just reinforces what works.
JM: How did seemingly simple communications technology change the environment that you were facing with JSOC? Was it just the pace of things? Was it an increase in uncertainty? A combination of all?
SM: Yes, it’s a combination. The first thing you tend to do with an organization such as JSOC—which had a lot of money and resources, so we could be at the cutting edge—is you build this structure to make the organization more efficient. Then you are in an environment where that proliferation of technology is out there, but you find that the enemy you are fighting isn’t traveling the same path. Al-Qaeda in Iraq was using the same communications technology we had, and it allowed them to be dispersed but coordinated. All of the cells operated independently, but they were doing so toward the same goal. That produces a very coordinated effort, but one that is impossible to predict because even they cannot predict it.
That, in turn, made our task uncertain by definition. In addition, the enemy was affected by so many input variables, and we became one of those input variables. Every time that we struck the enemy network, we changed it. So every time we acted, we changed our problem, and we took a step back in understanding it.
That speed at which things are happening in the external environment was a problem for JSOC then and remains a problem for businesses now. The problem with rapid change in the environment is the impact it has on the rate of internal coordination or synchronization: how often you have to get your organization coordinated. [See figure 1.]
When an organization is operating in a simple environment, things are moving kind of slowly. You don’t need to coordinate that much, because things don’t happen that fast, and it’s pretty straightforward. Cause and effect are clear to everyone. When the environment becomes complicated, suddenly it takes expertise and study to figure out the relationship between cause and effect, so you have to coordinate your organization more often. Complex is different because the relationship between cause and effect is only apparent after the fact. So, as a reality, you have to coordinate your organization much more often because you have to even figure out what is going on at any particular moment. Of course, in chaotic, there is no apparent relationship between cause and effect, requiring near-constant coordination.
What I would also throw out here is a third axis: the percentage of the organization that has to be coordinated. So before, this might have been just a small group or small percentage, maybe a few key leaders. Now, when things speed up, not only do you have to coordinate more often, you also have to coordinate more people. Suddenly, you are in this place where everyone basically has to be coordinating all the time.
The interesting thing about this figure is there is a dotted gray line here, below which things are fundamentally predictable—not always easily, but they are fundamentally predictable. Above the line, prediction is impossible. What we found in JSOC, and what every organization finds, is that it periodically gets pulled up above the gray line by a natural disaster or marketplace thing. So what do you do? The financial crisis happens; everyone goes into a conference room for three or four days. They buy pizza, they work it out, and they survive it. Then, as soon as it is over, they go right back to complicated mode, with defined teams and the line-and-block organizational chart, because these are the structures we know and love.
At JSOC, we found that we got pulled so often that we just had to live in complex mode. You kept wanting to go back—everything about you wanted to get back to complicated. In fact, the US Army is classic for that. It would really want to get back to simple: Our organizational structures go here, our office space goes here. But what I would argue is that, just as JSOC got pulled, there are more things pulling organizations more often up to complex, so you have to coordinate a lot of your organization fairly often. That is a whole new world. People look at this thing and say, “That is true, so how do I do that?” Well, it is hard.
JM: It seems doubly hard because in complex or chaotic environments, we cannot be sure at any moment in time of everything that is occurring in the environment, so we do not know how to respond. For example, take something called a “design structure matrix.” This matrix is how you organize and sequence all the various tasks that go into making any complicated device, such as a cell phone.3 But a design structure matrix only works if you can map all the interdependencies between tasks and parts. If you can’t map the interdependencies, you have to operate a different way. That seems to be the great thing: that you stumbled onto this new way of operating.
SM: You know, if you consider Alfred Sloan’s famous construct for General Motors, it worked very well for quite a long time. Nowadays, if you tried to design that, things move too fast. Also now, each person in every position is a human being who not only is different from other human beings but is different from themselves the day before, and they will be different tomorrow. So you are working with these infinite variables.
This is the core of “shared consciousness.” The theory is that you are dealing with rational, smart people, and that, given the right information, they will get pretty close to the right answer—if they have all the information, and if their interests are aligned. So you have to align people’s interests. If there are competing or diverging interests, you have a problem from the beginning. We work hard with companies to align those interests, and it is hard. You can find that you have a great strategy but are just not aligned on it. But then you use these cross-functional teams to create the conversation to force this interaction that won’t organically occur, because our physical structures do not enable it and certainly don’t encourage it.
JM: So how do you create the physical and organizational structures that encourage the type of open, rapid communication needed for “shared consciousness”?
SM: Typically, strategic leaders operate at a longer wavelength [figure 2]. They think deep thoughts that really do not change very often. At the operational level, they do more things that matter, so you have to revisit the tasks more often. Down at the tactical level, it is everyday things—in the rifle squad or at the point of sale for a company.
The challenge for any organization is that the tasks at each level need to be synced together. Traditionally, there was a very regimented way to do this. In the military, the top commander formulates a plan and issues an operations order that has the tasks for the next echelon, and they issue and order the tasks for the lower echelon. In a company, you have a similar process with the annual strategy. But over time, because of different wavelengths and inputs, you get out of sync, and things drift apart. So the organization needs to resynchronize. Traditionally, if you think about it, the frequency of doing this is yearly, because companies do an annual budget and an annual strategy. They may have a five-year plan, but the majority of the decisions are taken and revisited in an annual cycle. In JSOC, this cycle was 24 hours—not by choice, but in response to the war we were in—and that was completely different. So we had to coordinate a lot of the force, most of the force really, every 24 hours. It was like the no-huddle offense in football. That is a fundamental difference for people who were brought up in a more comfortable mode. Leaders who are resynchronizing every 24 hours cannot have the type of control they used to have. It’s just not possible. They are human.
JM: Do you think there is another change there as well? Because, as you describe the different wavelengths and how they break down, you can almost see the neat top-down breakdown. You have one big strategy, which breaks into three operational tasks, which break into a hundred tactical tasks. But it sounds as if when you are doing the 24-hour communication, you also are doing the communication from the bottom up. So you adjusted the strategy in a way every 24 hours?
“Leaders who are resynchronizing every 24 hours cannot have the type of control they used to have. It’s just not possible. They are human.”
SM: It wasn’t a top-down thing; it was a conversation. People ask me, “How did you change the strategy and issue the orders?” Well, we sort of didn’t. There were some times when I made big decisions, like we are going to go out to the western Euphrates River Valley and fighting our main effort there. But that was about the extent of decisions I made. Everything else is this daily conversation. In it, everyone listened to what we were doing, what the situation was, then acted autonomously. Interestingly, because they had all been in that conversation, they didn’t need to be told what to do. There were not a lot of questions like, “Boss, should I do this—or this?” Everyone looked at it and said, “OK, I got it.” Then every 24 hours, because the situation kept changing, they would just adjust.
It even went to things such as allocation of resources. We started the war with us allocating resources from the headquarters. Then it went to this give-and-take system between the task forces. They would cut the deals between themselves. They would pass helicopters and Predators, because they all saw the big picture. If someone has a priority, everyone can figure that out: “Give them to Joe, because he has the intel right now,” but then give them back to the next priority. So it became this organic thing that you could not control from headquarters without messing it up. That is the “eyes on, hands off” part. You are making sure it is happening, but you are not messing with it.
JM: That final portion seems to be key: Once you have achieved “shared consciousness,” that alone is not enough; it’s important to empower the employees to act quickly and independently on that information. You talk about going from 18 raids a month to that many in a single night, and I myself can remember sitting and watching a feed and being amazed. Sometimes leaders, especially in business, struggle with that, not because they do not want the same successes but because there is a loss of control involved. Did you have those same fears coming up as a junior officer, and how did you overcome those as the task force commander?
SM: I was absolutely raised in a more mechanical environment. I was a control freak as a young officer. When I was a mechanized company commander, I put everyone on the same frequency so that all the tracked vehicles turned when I said “turn” and that sort of thing. You could kind of get away with it at that scale. When I got to JSOC, I still thought we were going to be able to use technology to control this thing. It wasn’t until our opponents proved to be just so radically different than before that we had to do something different.
JM: That fear in managers of losing control seems rooted in the fact that employees can make some serious errors. So how do you avoid that individual failure? Was it just all down to having good talent or elite trained special operators?
SM: What I would say is that we did have some really good talent, but the thing that really made JSOC different was not the experience or the brains, but the culture in the organization, which was really interesting. It was different in every part of the force, but the healthiest or most valuable part to me, at least, were Delta Force and SEAL Team VI. The reason they were was that they had been raised in an environment that said, “The rules don’t apply.” So when they got a mission, as long as it wasn’t illegal or immoral, they were expected to do it. If you think about it, if you put a whole bunch of rules on someone, it gives them a whole bunch of excuses for not getting something done. What we did was that we had a slightly different part of that culture in the organization that said, “We are going to get this done. There is not an excuse; don’t fall back on doctrine, don’t fall back on anything.” When I first took over, we had risk aversion in the force. It wasn’t risk to their personal safety; it was risk to mission failure. They were raised that you don’t fail a mission, so they didn’t want to go ahead unless they had perfect intelligence, unless they had backups to backups. So I said that we don’t have time for that in this type of war; there is going to be a lot of failure.
In many organizations, shaping the culture is the biggest part. Something that I think companies can do is to take a very different view of how they are going to shape expectations and culture for their people. But when you do that, you have to be willing to underwrite some failure. You have to do many different things. You have to reward differently, and that can run counter to what has been done.
JM: As much as companies like these ideas of transparency, communication, and accepting failure, they often seem to have trouble actualizing them. From research, we can see that what you did with JSOC’s Situational Awareness Room wasn’t just an accidental success. You have all of these drivers that are key to organizational performance:
Those are, we found, huge drivers of team performance. So how did you stumble onto this? Was it a momentary flash of genius, or was it built up over time from your experiences?
“We didn’t come up with the concept of ‘shared consciousness’ first, then try to get it. We started doing it first, then we found we had it, and it was just magic.”
SM: We started with a typical operations center. Then I just started working and staying in there, because things were happening so fast—and we found that that just worked so well. We didn’t come up with the concept of “shared consciousness” first, then try to get it. We started doing it first, then we found we had it, and it was just magic. At first, we kind of strained against it. You try to go back to the processes and structures you are familiar with. Then you sort of just let go.
I think an apt analogy is a planned economy. Every time people create a planned economy, it is just not adaptable enough. A market economy is not perfectly efficient, but it responds. Nobody knows how it works, as we proved with the stock market and financial crisis. We have theories of how it works, but ultimately it is people. And like our conflict with al-Qaeda, every time there is an intervention, it changes the basic nature of the system. So no one really knows, but they have a general concept.
That is kind of what happened with shared consciousness at JSOC. If you asked someone to perfectly draw out what was happening in the force and how it was working, I could not have done it. I don’t think anyone could have. I think that if we tried to regularize it, we would have screwed it up. That is tough for a CEO, to stand in front of his shareholders and his board and say, “I don’t really know how this beast works, but it is working.”
JM: When you look at what you created at JSOC and now with other organizations, using the same principles, you see very diverse cross-functional teams set up in open plans. One of the traditional arguments against such setups is that you are not allowing each function to train or support its unique, specialized skills. How did you keep your intel analysts sharp and your operators sharp even in that mixed environment that is so diverse that you described it as “the Star Wars bar”?
SM: Well, there is a balance, because they have to do a certain amount of just their focused stuff. The Army has gone through this cycle where you create units with infantry and all the supporting arms together because it creates better teams. Then, 10 years later, someone says, “artillery skills are degrading,” and we have to put all the artillery together so that they can train each other. So, in the end, it has to be a hybrid.
The simple fact is, every organization is by definition a matrixed organization. You are both functional, and you are task-driven. The open environment and the democratization of information are key to making everything work in a matrixed environment.
JM: Another difficulty with the change to “shared consciousness” and empowering employees is that managers and leaders typically see their value to the team a certain way: often charting the strategic direction and being the decision maker. So that change to empowering employees is really a change of your role as a leader. How did you chart that course?
SM: If you think about the role of most leaders, they are commanders, which is the word we use in the military, of course. You have the responsibility to make decisions, and, in the minds of most people, the commander is supposed to be the person with the strategy. The commander is supposed to make the big decisions as well as provide inspiration and the other things that come with it. So if you don’t do that, you start to think, “Well, maybe I am not doing my job.” Now, certainly it is true that a commander is responsible for everything that a unit does or fails to do. What that does in many people is to reinforce the idea that, if I am responsible, I better control it. But responsibility can have a broad sense.
“We believe you need to have this intersection between what you teach people to do, what the processes in the organization are, leadership, behaviors, and so on. All of these factors need to support each other, or each one withers.”
JM: Is that why you describe the role of a leader as being closer to that of a gardener? Is it much less making those decisions and being directly responsible, or more making a culture change?
SM: That is absolutely right. The gardener’s primary responsibility is to create the garden. It is not to grow things. If the senior leaders are focused on the growing of each thing, they are not going to spend their time making sure the garden is protected and watered. That is what I find in organizations now: Those leaders who step back, create that environment, and make very few specific decisions get a much better outcome.
JM: At Deloitte, we have talked to a number of organizations in industries ranging from government to financial services to oil and gas about similar culture change. One recurring challenge seems to be the constant struggle to keep the culture real and regular, and not just confined to a manual on the wall. How did you manage that? How did you put the cultural artifacts in place? The shared values?
SM: I would be lying if I said I am sure. I will tell you what I think we did, and that was the constant conversation. When we synchronized every 24 hours in the daily video teleconference, we didn’t just synchronize on operations—we synchronized philosophically every 24 hours, 90 minutes long, across the whole force. A certain percentage of it was just passing information, but we had all the information on our portal, so you didn’t go through all the details, because it was there already. We spent the time talking about, what does that mean? What is each team trying to do? Philosophically, I am hitting key points on a constant basis. It was—and I don’t mean this to be a negative thing—almost like a communist indoctrination that happened every day: This is what we do, this is what we are, this is what we aren’t. You are hearing it not just from me but from across the force. Everyone is chanting the mantra every day.
So, as we became more powerful as a network, the key attribute was information sharing. The information you share actually has power, because someone is going to do something about it or with it. That became a responsibility, and that was different. Before, in very few organizations did you ever get in trouble for not sharing information; you got in trouble for sharing information with people who shouldn’t have it. We tried to make it the culture where, if you don’t share information, you can be held accountable for that. If somebody didn’t know something they needed to know, and you had that information, then they shouldn’t have to ask you the question: If you know they need it, you need to make sure they get the information. Now, it is hard to do that because you don’t know who needs to know every bit of information. But it leans you in the direction of, wow, we need to get this information out—a different mind-set.
JM: A thing that often scares companies about that is you want to regularize that transfer of information. You don’t want to leave it up to the organizational culture or individuals, because they may fail. You want to write a rule that says, “You send this information here.” Especially when you are scaling to organizations of 2–3 million employees, like the largest corporations, how do you have the confidence to just let that work, and let the information flow?
SM: That gets back to the idea that you have to constantly synchronize a huge percentage of the organization, because you are never going to get it all right once a year or once a month. Too much is happening. I just don’t think you can regularize it as much as you want to.
JM: Research also backs up the fact that you may not have to regularize it, even for the largest organizations. If you take social network theory, around 150 is your maximum for individual connections.7 The good thing is that, with that 150 number, you only need three degrees of separation before you have covered the 2.1–3.2 million employees of the largest companies. So all you need is for the information to go to a friend of a friend of a friend to have covered the largest organizations on earth.
SM: Right now, they are trying to introduce these principles in graduate schools, but if you go to an organization that doesn’t practice it, you may know the theory from school, but you will not do it yourself. We believe you need to have this intersection between what you teach people to do, what the processes in the organization are, leadership, behaviors, and so on. All of these factors need to support each other, or each one withers. DR
For more information about Stanley McChrystal and his book, see www.mcchrystalgroup.com.