Host Intro (Gene Kim)

Okay. I am so honored and delighted about who is speaking next, Dr. Ron Westrum, Professor Emeritus of Sociology at Eastern Michigan University.

His name will be familiar to anyone who has read the State of DevOps Reports that I had the privilege of working on for six years with Dr. Nicole Forsgren and Jez Humble from 2013 to 2019. It is the cross-population study that spanned over 36,000 respondents that allowed us to better understand what high-performing technology organizations looked like. We looked at the architectural practices, technical practices, and cultural norms. And without a doubt, what those cultural norms might look like were made possible by the work of Dr. Ron Westrum.

He received his PhD in sociology from the University of Chicago and has spent decades studying complex organizations, including healthcare, aviation, and the nuclear industry. And one of the models he created was the famous Westrum Organizational Typology model that brilliantly categorized organizations into pathological, bureaucratic, and generative. It's featured prominently in the State of DevOps research, The DevOps Handbook, and Accelerate.

And when Dr. Forsgren CC'd me on some correspondence that she had with Dr. Westrum, I almost fell out of my chair. I was able to interview him for four hours on my podcast, The Idealcast, and I'm so delighted that he'll be teaching us today about information flow in organizations and providing a case study of an organization that you have probably heard of. Here's Dr. Westrum.

Dr. Ron Westrum

Today, we're going to talk about information flow cultures, and we're going to talk about the cultures of organizations.

So what the devil is organizational culture? Well, organizational culture is a complicated thing. For instance, it has the following characteristics. Organizational culture is practices, organizational culture is thoughts, organizational culture is feelings, and it is symbols. So while all these are important, we're going to use another index: the flow of information.

Why is information flow the right thing to do? The reason is basically information is the lifeblood of organizations. If the organization has a good flow of information, the organization will do well. If it has a bad flow, it's not going to do very well. Information is also a powerful index of how an organization functions. An information flow culture, in fact, reflects how managers shape values and behavior.

And we're going to describe three different information flow types. One of them is generative, where you have a high flow of information, the best. Then there's bureaucratic, which has a medium flow of information, and pathological, which has a low flow.

So let's look at pathological flow. In pathological organizations, you get low cooperation, very high conflict, an emphasis on taking care of the leaders, strict boundaries, messengers get shot, and you have low creativity. So you have a toxic environment.

In a bureaucratic situation, you get modest cooperation. The emphasis is on rules and regulation. You have problems with silos. Messengers are tolerated, not necessarily encouraged. Conflicts are tamped down, and creativity is allowed. And here is my slide, which I think reflects the flow of bureaucratic information, which is that it's slow.

Now, what we'd really like to have is a generative flow of information, where we have high cooperation, we have emphasis on the mission, we have a boundaryless organization where things move quickly over the boundaries. Speaking up is encouraged, and in fact, people have psychological safety and high creativity. So here is my example of how a highly creative organization is supposed to function. I think Star Trek is a perfect model.

Now, let me emphasize one of the features that goes with generative information flow. At Google, they had a project called Project Aristotle, and they studied what made for an effective team. The number one feature of an effective team was psychological safety, the ability to speak your mind without fear of punishment. When communication is easy, there is more of it, but it's also the right kind of communication.

I like to say that a high flow of communication has these three characteristics. Number one, it's timely. Number two, it's easy to understand and comes in a form that's easy to make sense of. Number three, it meets the receiver's needs.

Now, there's a classic example of this. During the famous Redstone rocket program, which was one of NASA's first, a prototype went off course and crashed. Wernher von Braun, head of the project, tried to figure out by many analyses what had happened. The analyses did not suggest a cause. Now they were going to have to start from scratch to redesign the missile.

But then an engineer came to von Braun and he said, "I think I did it." "But how?" von Braun wanted to know. Well, the engineer said, "I touched a part of the circuit with a screwdriver and got a spark. I checked, and the circuit seemed to be fine, but maybe that was the problem." Well, it turned out that was the problem. Okay, so the problem got solved, and then von Braun sent the engineer a bottle of champagne.

So take a moment to think about your organization. What would happen when an engineer admits to making such a big mistake? Does he get a bottle of champagne?

Generative cultures are often found in high-performance organizations. They are common in high-reliability systems that require greater cooperation for success. They're typical of elite military units whose cooperation is legendary, for instance, the Navy SEALs. And they are often seen in consumer and service industries when exceptional consumer satisfaction is the goal. And they are often led by technological maestros.

So what is a technological maestro? Well, this word was coined by Arthur Squires in his book The Tender Ship about leadership and technology in World War II, and it meant the top leaders had these characteristics. Number one, technical virtuosity. Number two, a high energy level. Number three, an ability to grasp the key questions. Number four, the ability to grasp the key details, high standards, and a hands-on attitude.

Now, here's another example of a maestro. In June 1978, an engineering student called an architect named William LeMessurier, who had designed key parts of the Citicorp building in downtown New York. The 57-floor building had an unusual footprint. The student wanted to know whether the building was stable or not. Was it going to be stable in a high wind? LeMessurier assured the student that it would be stable, and he personally had designed a special mass damper on the top floor to steady it.

But then he had a second thought, and that thought was that if the building was built according to specifications, there would be no problems. But had it actually been built that way? So LeMessurier called up the builder. Well, the builder said they had pretty much followed the plans that they'd been given, but there was one detail that was different. They had used rivets instead of welds to hold the building together. On a short building, this would not matter, but on a 57-story building, a quartering wind strong enough would bring down the building.

How often would such a wind show up? The answer was about every 16 years. So they had to fix it, and they did fix it. They told the newspapers about it but asked them to hold the story. So for several months after the secretaries had gone home at night, contractors pulled off the wall panels and welded the girders together. After they fixed the structural problem, then the newspapers published what had happened.

Oh, by the way, what is requisite imagination? It's the fine art of anticipating what might go wrong. So here is a prime example of requisite imagination. And remember, mastering the key details is one trait of a technological maestro.

So maestros build a generative information flow, and this creates the complex web that allows the organization to build things. For instance, this is how you build airliners. So we're going to look at how Boeing created airliners.

Building airliners is big business, and I have a law about this: the higher the stakes, the rougher the play. So when Boeing builds airliners, this is rough play. It involves very high stakes and high risk. Yet Boeing did it well for many decades. For instance, we have examples like the Stratoliner, the Stratocruiser, the 707, 727, the 747, and finally, the 777 airliner.

So how did Boeing do this? Well, Boeing had a lot of money, a lot of people, and a lot of machines. But Boeing also had a secret weapon, and that secret weapon was a culture that held all those assets together, a culture like a family, in spite of crises like business downturns and strikes and so forth. Culture is actually a form of capital. Any company that manufactures something as large and complicated as a jet airliner forms a complex web of knowledge.

So if we take the cultural capital and put it together with the technological maestro, we get planes like the Boeing 777, a marvel of precise engineering. Understand that this human web of knowledge and competence is fragile and may degrade under rough handling. So if you interfere with this culture of human competence, bad things can happen.

And at Boeing, this seems to be what happened. After Boeing merged with McDonnell Douglas, the merger caused damage that undercut the web of manufacturing know-how. And here is one of those pictures that is better than a thousand words. We have Phil Condit of Boeing listening to Harry Stonecipher of McDonnell Douglas, and you can tell this was not a happy marriage.

So as Boeing's culture went out the door, its aircraft maestro, Alan Mulally, went to Detroit, where, by the way, he took over Ford and did great. Harry Stonecipher of McDonnell Douglas soon became the new CEO of Boeing, and under him, the culture rapidly declined. Stonecipher wanted the new culture, what he described as going from family to teams. And this is a very important set of words because even though those things don't seem to mean a great deal of difference to the ordinary person, at Boeing, it made a huge difference.

One employee told Harry Stonecipher, "My God, Harry, don't you know you're changing the culture of Boeing?" Stonecipher leaped into the air and he said, "My God, that's what we want to do." That's what Stonecipher did. But was it a good idea to do it? What culture was being replaced, and what would take its place? Suppose that Boeing's great accomplishments had only been possible thanks to its culture. What was this culture?

Boeing's employees described it as being like a family, but this culture was actually a high-cooperation, generative culture. Yet Stonecipher was not happy with the Boeing culture for making planes. He wanted a culture focused on making money. So the generative culture got replaced by a bureaucratic culture. But the former culture, the generative culture, had been the key to Boeing's success.

So as the price of Boeing stock went up, the price of its technical product fell. So, the next airliner that came down the pipe, the Dreamliner, was beautifully designed, but messed up on batteries and other manufacturing issues, and I understand it's still messed up. Stonecipher, meanwhile, had left Boeing in 2005. Other CEOs followed, but success did not return.

Then Boeing made a more serious mistake. It put fatal flaws in a new airliner. The new 737 MAX had major defects. This airliner had to work to beat Airbus, but it didn't. The 737 MAX had new MCAS software installed that caused unexpected motions. This is a perfect example of a latent pathogen, using the term of Jim Reason.

Pilots should have been trained for the new software, but they were not. The full toolkit of the knowledge to operate this plane was not supplied. One U.S. pilot, after suffering from MCAS problems, said, "I am left to wonder what else don't I know? The flight manual is inadequate and almost criminally insufficient."

So if culture breaks down, things get missed. No maestro and a messed-up culture: you could be flying without a parachute. The flaws in the 737 MAX soon led to two crashes, killing a total of 345 passengers. A broken culture had led to a broken airliner project and a huge reputational loss.

So what are the lessons we learn from this story? The most obvious one is that if you have a working culture, don't mess with it, and if your culture is not working, you better find out how you can fix it. And finally, if you don't know whether your culture is working or not, shouldn't you find out?

I'm looking for dialogue, useful critique, case studies, potential consulting, and any questions about the concepts. I want to thank you for listening. Have a good day.

Q&A

01Gene Kim

Thank you so much, Dr. Westrum. It has been such a privilege to have interviewed you for four hours. And one of the big surprises to me in the interviews was that you had revealed that when you were introduced to the field of aviation safety, you were one of the few sociologists, that you were surrounded by primarily psychologists. Can you talk about why so many of the important insights came from sociology as opposed to psychology?

02Dr. Ron Westrum

Well, I remember the first time I went to one of these meetings, and basically the people who were the hosts were aviation psychologists. And I remember somebody came up to me and he said, "I don't understand why you're here."

Well, I found that was fairly threatening, okay? I had just joined the field, and here I am, and he's questioning why I'm even at the meeting. So, the next morning, I remember I got up, and it's one of those days where you're going to give a lecture, but you're sort of dead and you can't really get started because it's a different time.

So after my talk, I thought, "Oh my God, he's going to cream me now." He came up to me after the lecture and he said, "Thank God you're here." He said, "You've really got some stuff to give us." And I was really pleased by that. And I think the truth is that they really hadn't thought about the fact that basically an airplane flight deck is a group, okay? And not only that, but it's part of a larger organization, and you need to understand the organization to understand what the pressures on the flight deck are likely to be.

03Gene Kim

And what year was that? And could you even give us the briefest tutorial on what is cockpit resource management or crew resource management, and how did it impact the industry?

04Dr. Ron Westrum

Right. Well, back in the 1980s, basically, and I think this started out with United, NASA and the airlines got together and they developed something that would prevent some of the, what they call crew-caused accidents, okay? And that became known as cockpit resource management. I forget what the original name was that United had for this. But that basically was opening up the entire field of aviation to look at the group dynamics involved in keeping an airliner safe.

And I admit, I was sucked in like everybody else, and I thought, "Oh, this is really cool." And I cannot tell you how exciting working with that group of people was. It was really fantastic.

05Gene Kim

And for those of us who aren't familiar with CRM, could you describe what that is and what made it so different than how crews worked in a cockpit before?

06Dr. Ron Westrum

Yes. Well, basically, after World War II, there were a lot of pilots, and there was still a stream of them coming out of Vietnam and so forth, who were military pilots, who were basically fighter jocks. And their idea was, "I'm the boss. I'm going to do what I want to do, and the others can take a leap or something." And pilots would actually say things like that to their copilots, like, "Shut up, I'm in charge here." Okay? Well, this was obviously a very dangerous condition.

And so, a lot of what crew resource management was about was getting the pilots to listen to their copilots. Because in many cases, they actually had conversations where the pilot had said, "Shut up and don't tell me about what's going on with the altimeter," and so on. The next thing you hear is a plane hitting the ground because they augered in.

Well, so the whole thing about crew resource management is learning to use the intellectual resources you've got on the plane and take full advantage of what everybody has heard or seen or thought and so forth. And so we've come a long way. I knew Bob Helmreich, who was the great advocate of all CRM and so forth, and went all over the world basically selling his ideas. And that has made a huge difference to aviation safety.

07Gene Kim

Amazing. And so you introduced me to the term, the technical maestro, the technological maestro: high energy, high standards, great in the large, great in the small, and loves walking the floor. And you had also mentioned something else that also had a similar sort of searing aha moment for me, and that was Ravina's law, I think number 23. Can you describe what that is and why you think it's so important?

08Dr. Ron Westrum

Well, Ravina was an inventor. He had 230 U.S. patents and had basically seen and done everything in R&D. And he said, basically, he had a set of laws that came out of that, and one of them was number 23: if the boss is a dope, everybody under him is a dope or soon will be. Okay? If the boss is a dope, everybody under him is a dope or soon will be.

Well, I don't know why this was a revelation to you, but it was certainly a revelation to a lot of people, because that's the problem: lousy leaders tend to recruit other lousy leaders because they're not going to be threatened by people who are smarter than they are.

09Gene Kim

I love that. Because for me, it simultaneously describes those situations I've been in where everything was going great, fully enabled by that leader at the top, or things were going horribly, and fully enabled by that leader at the top. Am I capturing accurately what the implications of Ravina's law is?

10Dr. Ron Westrum

Yes. I would be in jeopardy here talking about my former employments and some of the experiences I've had, so I'm not going to tell you that. But all I can tell you is that that's so true.

11Gene Kim

You had also mentioned something that I found astounding. So, I think when we think about the greatest experiences that we have working, where we have fully unleashed human creativity, we have a sense of enormous satisfaction, we feel engaged in our work. But then you had described the opposite conditions in the Whitehall Study. Could you educate us all on what the Whitehall Study is and what it found?

12Dr. Ron Westrum

Well, the Whitehall Study was basically about the relationship between health and status in the organization. Whitehall is, of course, where the British government does its day-to-day work. And there was a big study, that's what it's called, the Whitehall Study, which basically showed that the higher up in status you were, the less likely you were to have a heart attack.

Everybody in Britain has the same health system, in principle, and yet the people on the top are less likely to have heart attacks than the people on the bottom. So, the fact is that status is a very important indicator of your mental health and your psychological, physiological health.

13Gene Kim

And you had mentioned that one of the implications is that in certain organizations, pathological ones, life is horrible for everyone except for the leaders. That the whole organization's actually optimized for the leader. Is that a correct interpretation?

14Dr. Ron Westrum

That's a correct interpretation. I think the thing that cued me into it was many years ago, I had a student whose father used to work at Fisher Body. And he said every day when his dad went to work, he got sick going to work.

And Fisher Body was apparently a place where there were a lot of problems like alcoholism and suicide and so forth, and it was a very punitive environment. This is probably where I originally got the idea of a pathological environment. But it wasn't the only one. I also taught a class, basically, which was right around the area where the Hydramatic plant was, and this was a plant that GM had that basically my students required counseling after working there. Okay. And so I began to realize that there's some kinds of environments that make people sick.

And I think by contrast, generative environments tend to make people more psychologically together and physically, health-wise as well.

15Gene Kim

So we talked about the pathological organizations. The origins of your study of generative organizations, I found genuinely surprising that you sought out specifically organizations like the Sidewinder project at the China Lake Naval Research Laboratory. Could you talk about what it is about an R&D organization that you thought was especially suitable for the study of generative organizations?

16Dr. Ron Westrum

Well, the reason that I studied China Lake is because the story of Sidewinder is basically the little engine that could, okay? So here is this small group of people, basically out in the middle of the Mojave Desert, literally. And they put together this missile that was better than everybody else's missile. Okay. Well, I found that a really appealing story, so I said, "What's the larger context of this?"

And I realized after a while that this entire laboratory, and it was big, actually, the laboratory itself, about 5,000 people, was a huge skunk works, okay? And I realized after talking to them that they routinely created miracles. And in fact, one guy said, "We always thought we could do one of anything over a weekend."

Can you imagine how threatening an organization like that would be to the ordinary R&D organization? They would take other people's systems and fix them. Like the Sparrow missile is a good example, which had a lot of problems, and China Lake fixed it. Well, do you think the others who got their missile fixed were happy about that? They were not. And so like Cinderella, the high performers tend to get punished.

17Gene Kim

And what was it about the way that the groups worked in the Sidewinder project that marked them as so unusual and so worthy of study?

18Dr. Ron Westrum

Well, first of all, they accomplished something that nobody else could do, and in fact, something that they had been told not to do. They were ordered not to create new missiles. And so they did it all on the sly, and they literally essentially rolled this thing out, basically at the end of this process, having disguised it as a targeting project and some of the other things that they used. And here was this thing that was better.

And so these guys would come out, the admiral would come out, and this thing would do amazing things. And they'd say, "We have to have this missile. In spite of the fact that we told you not to do it, we have to have this missile because it is so good."

So you have to ask yourself, okay, so what kind of culture creates that sense of engagement, that sense of excitement, that sense of the ability to do anything? And that's exactly what China Lake had.

19Gene Kim

That is awesome. So when I reread your information flow paper from 2013, it was actually startling to me just how powerful I found the last couple of sentences. You wrote, "In summing up, culture is no longer neglected. Information flow is, of course, only one issue among many in safety culture, but I feel it is a royal road to understanding much else."

Can you talk about what that royal road is and where you think that royal road takes us?

20Dr. Ron Westrum

Well, I've basically spent essentially my adult life developing this theory, or whatever you want to call it. And I think the thing is that at the end, when I realized how many different things the information flow was likely to predict, I said, "Oh my God, we have something here which is really quite amazing."

And how did I know that it was amazing? Well, because what I would do is I'd put this chart in front of other people, they'd say, "Oh my God, I know where we are on this chart." Okay? So it had a lot of surface credibility.

And as you know, in DevOps, you people did a big study, and that's what you found. So basically, it does correlate with creativity as you go from a pathological or bureaucratic to a generative organization. It is amazing. And did I know that it would do all that stuff? No, I didn't. It's just been delightful to see. In fact, as you do further studies, this seems stronger and stronger.

21Gene Kim

Awesome. In fact, let me end by just stating that every interaction that I've had with you, I've learned so much, and I cannot overstate just how much impact that you've had on the DevOps community, that initial finding we found where we found that the Westrum Organizational Typology Model was one of the top predictors of performance. So, from all of us, thank you so much, Dr. Westrum, and I look forward to catching you later in the conference.

22Dr. Ron Westrum

Absolutely.