This is a transcript of episode 113 of the Troubleshooting Agile podcast with Jeffrey Fredrick and Douglas Squirrel.
With a pandemic raging around the world, how can you and your agile team make sense of the world around you and adjust accordingly? The Cynefin framework gives you a way to categorise known and unknown information and figure out how to approach and change your work appropriately. We describe the framework (including how to pronounce Cynefin!) and explain how you might use it to help your agile team cope and even thrive in a challenging environment.
Squirrel: Welcome back to Troubleshooting Agile! Hi there Jeffrey.
Jeffrey: Hi Squirrel.
Squirrel: So I imagine that all our listeners are at home when I often imagine our listeners at home, but particularly imagine them at home because we’re all at home. And when I’ve been at home reflecting on how tired it’s making me, for example, just to be at home and not be going into London. Something came to mind and I thought it might be a very useful idea. I’m the man of my favorite things about it is it’s called a sense s e n s e like sensing something, sensing your five senses, a sense making framework. And if you’re like me, then you might feel like the world doesn’t make a whole lot of sense right at the moment. So a sense making framework would sure be useful. And so I thought we might talk about that today. How do you feel about that, Jeffrey?
Jeffrey: I think it sounds great. I knew we’d have to do something related to the virus situation this week. And I think this was a good one.
Squirrel: Excellent. So the first thing we have to do is I haven’t managed to tell you what the thing is called yet or what the what’s the method is. And the reason is that its name is Welsh. And here’s a hint. If you want to name something and have people use it, don’t use Welsh because it’s extra complicated and difficult. But actually it’s pretty easy to learn to pronounce this, although spelling it is something you’ll have to do by looking at the title of this podcast episode. So the way you pronounce it is the first you say the word ‘Kevin’, do you want. Do you want to do this with me, Jeffrey?
Jeffrey: Absolutely. So. ‘Kevin’.
Squirrel: Yeah. And then put a ‘nugh’ in between the ‘K’ and the ‘evin’. So it’s in the ‘evin’. So between the ‘K’ and the ‘evin’, you put a ‘nugh’. So it goes ‘K-nugh-evin’ go ahead an try that
Jeffrey: ‘Ku-Nevin’ How was that?.
Squirrel: Yep. Perfect. ‘K-ugh-evin’
Squirrel: So always just if you get stuck in it, it’s very easy with this word to get stuck. Just say ‘Kevin’ first and then put the ‘nugh’ in and you’ll be saying it like you were Welsh, I think.
Jeffrey: And I do think we should at least do spell it out for people, because if you go look it up, you’re using the show notes you might think. It’s not what we’re actually talking about because it’s spelling’s so unintuitive.
Squirrel: It is. Go ahead, Jeffrey.You spell it.
Jeffrey: It’s spelled C y n e f i n.
Jeffrey: So I did hear Dave Snowden talk about this once and he said this is in part in response to all of the people using Japanese phrases.
Squirrel: He wanted to use Welsh.
Jeffrey: Exactly. You’re going to pull in some slightly esoteric term from another language and and do that. Well, you know, guess what? We can. I can play that game, too. And so I thought. Fair enough.
Breaking Down the Cynefin Framework
Jeffrey: So now we know how to say the name, and I have pronounced Cynefin framework. But but what is it?
Squirrel: Yep. So it’s a visual framework. So it has a visual representation. That’s very helpful. And Jeffrey, you found a wonderful diagram that some I still have managed to look at all the bits of it because it’s so colorful and full of detail. So we’re gonna link to that. And it’s a complex it’s a complex example of the standard diagram. We will also link to the standard diagram. It’s imagine a square cut into four pieces in the way you’d expect four square shapes that make all the lines wiggly. So although the vertical and horizontal lines of that cut the the square into four equal squares are all kind of wiggly and wobbly. And then in the middle you’ve kind of carved out a piece in the middle that’s of a chaotic, not very ordered shape. And that is the domain of disorder. So from the lower right and going counterclockwise or anti-clockwise. So the lower right quadrant is simple. The upper right quadrant is complicated, the upper left quadrant is complex and the lower left is chaos. And then the middle is this weird shaped for disorder. So I always mix up what these all mean and how you use them and so on. And I even made a mistake in our book, which Jeffrey thankfully caught. Thank you. And then an early reader also caught it. So, Jeffrey, could you help us? What what do all these domains mean and when do you use them? How do you use them?
Jeffrey: Ok. So I’m going to talk about it. My understanding of it and I will also say that I am not an expert. So if there’s experts, Cynefin, who spot mistakes and we say please do write it and we’d love to hear back from you, though, we’d love to get the correction and, perhaps in the future we can get a actual Cynefin expert on on the podcast. However, what I’ve what I’ve seen it used as and how I’ve used it is you start by by trying. You have a problem. And then from that problem, you’re trying to say, well, how should I move forward? And the framework is there to help me make sense of the problem and how I might move ahead. And so if I don’t know where I am, then I’m in that disorder section. I’m in that sort of gap in the middle. So when you describe the wiggly lines, these aren’t just lines. These aren’t just borders between the different domains. There’s like gaps. You know, you fall into the gap where you’re not quite sure where you are in that big gap in the middle. So if I’m not sure where where my problem represents, that’s my starting point. And now what I do is get enough information to try to figure out what domain I’m actually in. And then that will give me a clue as to how I should next move. So that if we go back to sort of the non disorder ones, as we said, were simple, complicated, complex and chaos. And i’ve also seen simple, also described as obvious because that’s it. That’s kind of how the solutions relate to the problem. In the simple to mean is that there they’re pretty obvious.
Squirrel: So can we go through these and have a look at what the actions are in each one? I thought we might actually link it to current events. You might think about how in our response to this virus challenge, we’re all confronted with how might we respond? And for some problems are going to be simple and some will be complex and so on. How does that sound?
Jeffrey: I think it’s great. And for the listeners who were following along. This is something that Squirrel and I are doing in real time. We haven’t we haven’t mapped out all of our answers here. So we’re essentially going to be making sense between ourselves of if we agree or not about where these things fall.
Squirrel: Guess what? We might disagree. That would be good. Okay. Go ahead.
Jeffrey: Yeah, absolutely.
The Simple Domain
Jeffrey: So let’s let’s start with the simple domain. And each domain has what they called the ‘action mode’, how you proceed. So in the simple action mode, you sense the situation.
Squirrel: No, wait, wait. Jeffrey, so so what is simple first? I mean, I kind of have an idea what simple is, but how do I recognize when I’m in the simple domain so that I can apply the action mode?
Jeffrey: Ok. Well, it’s also called the obvious domain, and the idea here is that this is sort of well-known, well-understood. You know, it’s the domain of best practice. And so it’s a place where it’s the things are tightly constrained. When you look at it, everyone will agree that it’s simple. You don’t need to be an expert. This is this is a domain where everyone looks around and would say, yep, that’s that straightforward. Now, they may not always have knowledge of all the possible solutions, but they can understand that what we’re dealing with is a pretty straightforward, very mechanical, tightly coupled situation. You know, the input and output are tightly coupled and simply connected. So making a sandwich might solve as a simple problem. I have bread and I have peanut butter and I have jelly and I want to make a sandwich. It’s going to be pretty straightforward.
Squirrel: There you go. And to take an example from the the current situation, if I know the other people have a disease, staying away from them is pretty simple. I might not know how it’s transmitted. I might not know all the ins and outs of the biology. I might not know how to spell epidemiology. But like, don’t go near people who are coughing and sneezing when there’s a virus about. That’s pretty simple. I can apply that that social distancing rule without a lot of knowledge needed or a lot of preparation.
Jeffrey: Yeah, that’s right. And so if you if you kind of look at the three steps you have in simple, it’s it’s the number one sense, the situation to categorize the situation into a known bucket. Three respond with a well-known solution. And so one is that you’ll have seen in sort of public education is people you can see that the messaging coming out from the governments follow this category saying, you know, here’s what you should be doing. So, for example, if you’ve been out that since this situation, you returned home, you should go wash your hands. And when you wash your hands, you should take 20 seconds and you should do it in a certain way that covers all the parts of your hands. So these these are simple actions you can put in with a simple trigger and that the job of the government in communicating up to what people should do is to make simple instructions that the general public can follow.
Squirrel: But the problem is that not everything is simple. So how about how about another of these domains? About if we’ll have a look at that.
The Complicated Domain
Jeffrey: Yeah. Well, the next obvious one is complicated and it’s has some similarities with the simple or obvious domain, which is to say that there is tight coupling between items and there’s some constraints that are in place. But what’s different is that now in a sense there’s the coupling between things are less direct, they’re complicated, as the domain would say. And but they’re still there. And with with the complicated domain, that is kind of the domain of experts. So unlike the simple domain that you should expect everyone to be able to understand and follow and apply appropriately in the complicated domain. You really have experts who can come in and say, right. I’ve sensed the problem. I can now analyze the problem and provide roadmaps. And then three, respond with a plan. So this is a bit different than that sort of straightforward application, simple domain because of the extra complexity. It’s no longer something that a random person can look at and say, oh, I know what to do. Or actually someone may say that, but they’re more likely to be to be wrong.
Squirrel: Yes, exactly. So to apply it in our current situation, it might be that I’m a nurse in a hospital and I see someone come in and they have a high fever and they’re having difficulty breathing.
Squirrel: And I have sense that they might have the symptoms of the Corona virus. My roadmap is give them a test and then take them to the intensive care if they have more trouble breathing. And then I can follow those steps because I’m carefully trained. I know how to apply the test. I know how to understand the results of the test. And I know how to run a ventilator. So you and I, Jeffrey, probably don’t know how to do any of those things because we’re not experts. Conceivably, we could learn. That’s one of the characteristics I always think of. In complicated is like if I really tried, I could figure out how to drive a Formula One car or I could learn to run a ventilator. Maybe there’s some skill that is special that they’ll never develop. That would still be the domain of an expert, but it’s at least comprehensible. I can I can understand. Here’s the plan. We’re gonna do this, this and this. It’s it’s got some comprehensibility to it.
Jeffrey: Yes, and I like your example of Triage, because you can imagine as they are coming up with the different approaches, as the situations evolved, they’ve reanalyzed and reconsidered how they’re going to be triaging people, how they respond and changing the plans as more information becomes known. In the end, there’ll be parts of that of that plan when they put it into effect, that will be simple. You know, there will be steps where they they sit in this response. There are times where there are simple steps that will be helpful. But overall, the context, the complexity of the overall response plan is complicated and requires expertise then of everything else that’s going on the hospital and how we expect this to interact with our normal day to day operations are.
Squirrel: A relatively simple step. Could be kept count how many beds we have available and if there’s an empty bed, put this patient in that that I could probably manage to do. And that’s a small step that might be part of a complicated plan of treatment for that patient.
Jeffrey: Yes, especially when it’s things like, well, do. Yes, we have a bed. But are they in the right ward? Does that ward need to be isolated? How are people going to enter it? Do you need to be in containment suits when you do so, et cetera, et cetera? So there’s a lot of complexity comes up.
Jeffrey: But there are elements that are that are simple.
Squirrel: And that’s how things can move. This is something I’m very interested in, is how do things move from one domain to another. So something that we started with, it was simple. Count the beds, put the patient in the bed, suddenly becomes more complicated, moves into the complicated domain. If suddenly you have to wear special suits in order to do it, or you have to displace somebody and judge about judge which ward they should be on and so on.
Jeffrey: Yeah. Really? That’s a really good point because it is a case where as this knowledge gets more and more formalized, as our way of understanding the situation improves, our ability to handle a complex situation becomes better and it becomes closer and closer to it to a simple one, especially the more that people have a common vocabulary in it, especially in a subset of the population. You know, in your office, you develop a certain where you all become experts about some sort of things. And now suddenly a lot of things seem simple that before were more complicated.
The Complex Domain
Squirrel: Got it. So how about complex? So that’s where I start to lose it, so I this is where I made the mistake in our book. So I’d love to be corrected. I’m definitely not going to try to explain this one. Help us out with complex what’s going on in that domain.
Jeffrey: Ok. Well, I’ll tell you about complex. It’s actually it’s it’s a lot of sense. It’s it’s a for most people who are into the framework. This is people’s favorite quadrant. And oh, yeah. And maybe maybe it’s the one that people used too frequently because it just it’s way cooler than all the rest.
Jeffrey: This is this is the domain of chaos theory and emergent behavior and that is it. Now we have rather than having a tightly coupled situation where things are only loosely coupled and so there is still cause and effect. OK, we’re not we’re not in a world that’s random, but a lot of times the cause and effect, you can’t really predict the effect until after it’s done. You can really understand the link between cause and effect until after you’ve done it. So, I mean, there’s a property called retrospective coherence, meaning once the once the outcome has happened, I can look back and understand it. Oh, yes, I understand. Every step happened. However, it’s very difficult or even impossible to predict what the outcome will be ahead of time. My exposure to this, as this goes back to when I was in university and I got very excited about the book by James Gleick called Chaos. And it was talking about the the new science of this world of emergent behavior, of fractals and flocking in all kinds of fun stuff. And that sort of emergent discovered attributes are really what was behind the complex domain. So it’s not like complicated in particular in that they’re there. It’s not the domain of expertise anymore. It doesn’t matter how much you know, you can’t concretely narrowly predict what the next thing is going to be. So now we need to do rather than starting with sensing that we did in simple and complicated. Right in both of us have the first step of a sensing in here. Sensing comes later. You start by probing, right? You probe and then you sense. So you the steps here to probe the sense respond. So we will try experiments. We’ll try tests to see what happened to gather knowledge experimentally and then decide how to respond with the idea that at some point maybe we’ll be able to gain enough knowledge that we can actually move this at some point into the complicated domain, that it won’t stay complex and uncertain forever.
Squirrel: Got it. So in our current situation, we see we got.
Jeffrey: Also, you know, I had a good discussion about this a few weeks ago, I think. Off podcast, we were talking about this difference between complicated and complex and I was talking about the example of weather forecasting is that was a classically an example of the butterfly effect said, you know, yes, there are cause and effect, but the causes can be so far away from the results that you aren’t able to predict. You know, this is the butterfly effect being a butterfly flapping its wing and one part of the world creates a hurricane somewhere else. That kind of an instability, or sensitivity dependence on initial conditions, to use the phrase from from the study of chaos and fractals and whatnot, makes it so that the predictions aren’t very good. And so even today you’d say, well, and therefore we have limits on our ability to predict the weather accurately. And you had a really good response, you came back. We’re much better at it now.
Squirrel: Yes, exactly. We can be pretty sure that it’s going to rain in the next hour that we usually know it’s done two weeks from now. We have no clue because there’s too many. Yes, there’s too much. So some part of the weather prediction problem has moved into the complicated domain. I definitely can’t tell you if it’s going to rain in the next hour, but I know where to look. The meteorological office here in Britain has a Web site. I can look. It’ll tell me if it’s going to rain in the next hour because an expert and a computer have worked on that problem.
Jeffrey: Yeah. I think that’s that’s a that’s a great example of the difference between making something that is complex and therefore not predictable, adding it, doing enough experiments, adding enough data to a point where at least in some limited way, we’ve made it a complicated problem rather than a complex one.
Squirrel: There you go. So in our current situation, governments are trying to figure out what to ask their citizens to do. And we’ve seen publicly lots of probing. This is what actually first made me think of this. I thought, wow, this really feels like these governments are in a complex situation because we had some governments that were very restrictive and gave people bracelet’s and tracked where they were and mobile phone apps and that were really restrictive on their citizens behavior and some others were probing with other things. So here in Britain, we had a looser approach for longer. And we’ve seen the results of those. The problem is that, of course, no one. Going back in time. Back. Going back in time, we could now say retrospectively, this is the action that would help you the most to control the outbreak. But at the time, that was absolutely unclear. It was not at all clear whether the the probes that people were trying. We’re going to have the right result.
Squirrel: Some did and some didn’t. But it could easily have been something else. And no expert could tell you that, rather. Experts could tell you. But they all had different opinions. So different experts in different countries saying, let’s try this. But they’re really saying, let’s try. They’re not saying this is a well-known practice. This is something that’s either obvious to everyone or there’s a manual. There’s steps you can follow do this.
Jeffrey: Well, I think it’s interesting in that there I would guess that if you were looking at one dimension of it, you probably could have made it a complicated problem. I would guess sort of the pandemic response in China with lockdowns and things like that in other countries which are complicated require expertise, but they in a sense are not complex, because the reason is because it happens when you’ve made a decision about how to weigh different factors. I think when you’re early on with people going to say, well, what’s the right level of what’s the tradeoffs we want to make here between controlling the spread and the impact on the economy, say just to choose two elements? Yeah, that that that that interconnectedness, the play between those things and trying to figure out what the overall result would be, because that’s when we go ahead and we have these sort of unintended consequences, when we tell people, right. Stay out of all the public spaces. Don’t go to bars and restaurants and theaters. Well, what happens to people who are working in those bars and restaurants and theaters? Suddenly they’re there. They’re out of a job. And now they’re home. And I want to close the schools. But what about the health care workers who now who have children week? Oh, yeah, that’s a different problem. And I think this is one thing that happens a lot in the complex domain is we take one action and there’s not just the expected or intended outcome, but there’s all these knock on effects and things are interconnected with one another. And that’s where you get this sort of emergent result. That isn’t always what you predicted ahead of time.
Squirrel: And very frequently isn’t. That’s how you know, you’re in the complex mode.
Jeffrey: Yeah, that’s right. And so I love the phrase that the problems in this area are called wicked problems.
Jeffrey: There’s there’s not there’s not a good answer as you can pick up off the shelf. Instead, you’re sort of learning how to how to surf, you know, learning how to how to navigate through the problem rather than mastering it the way that you could master a complicated problem.
Squirrel: There you go.
The Chaos Domain
Squirrel: So how about the last one, or at least the last of the four standard ones? What about chaos? What characterizes chaos?
Jeffrey: Chaos really comes back to it. We come here and talk about cause and effect and I think that’s the way I think about the chaos domain, is what happens with cause and effect in the obvious domain that the cause and effect is simple, easy to understand, easy to predict. In complicated, It’s in a sense straightforward. But it may it may have or be as it says, be complicated. It might be hard to understand, but it’s still manageable in complex cause and effect is still, in a sense that straightforward that you can- But you may not be able to predict it ahead of time. You can, but you can look back afterwards and say, right. I understood how cause and effect work. In chaos there is no connection between cause and effect there’s at least not one that’s understandable even in in retrospect. So it’s not it’s not a place where analyzing is going to do a lot to help you. So instead of sensing or probing the way we did in simple, complicated, complex. Now, in chaos we’re going to act, we’re going to act to try to stabilize the situation.
Jeffrey: So we’ll act and then we’ll sense what happens. And then eventually we expect the situation to respond by moving into one of the other domains. So chaos is often described as when, you know, kind of hell breaks loose. Like all things are going on. There’s too many things going on for us to make sense of it. We need to do first is try to get the situation under control. And rather than trying to absorb and understand everything all at once, we’re going to take some some straightforward actions. And in the you want to try to get yourself out of immediate danger and get someplace where you can then begin to have a time and space to figure out where this goes. And the transition here is from chaos to disorder. And that’s the interim back at where we started. Once you’ve acted to stabilize, Now you’re ready to start asking a question that we we opened with, which is, OK, well, where are we? What kind of problem? How do we now get information identified the domain and then move ahead?
Squirrel: That makes sense. So I’ll take a personal example for for chaos, our current situation. A few weeks ago, which seems a lifetime ago now, I innocently thought to myself, you know, I go to London a lot and I see a lot of different clients. I probably should get some hand sanitizer. And I innocently went to Amazon and I noticed hand sanitizer was awful expensive. And I thought, I wonder how that happens. And I now have some greater understanding how that came to be. But I’ve been on a quest for hand sanitizer for for a while. And even when I’m here, here at home, I still would like to make sure I keep my hands clean and so on. But especially when I was going to London, it was quite a concern for me because I literally might see five clients in a day and I don’t want to be a vector. So I tried a number of different actions and the action that actually worked was poking around in my bag for something else. And it turned out my mother in law had put some hand sanitizer in there the last time I went home for Christmas just because she thought it was a good idea.
Squirrel: That was quite a chaotic result for actually getting some hand sanitizer and the way we moved it into another domain. Is that a neighbor and a friend made some hand sanitizer following a recipe. And she gave me some of that. So now I have kind of the official stuff and the the made stuff and the made stuff- I still don’t know how she did it. So I guess I’d call that complicated that we did, because I don’t know it. You require an expert. Not a lot of expertise, but some. And so we moved it moved the problem of Squirrel, get hand sanitizer into the complicated domain, ask neighbour. So but it was really just acting. I was just trying lots of different things. And every shop I went into, I’d say, I wonder whether there’s anything in the hand sanitizer isle. Nope, there isn’t. Let’s try another action.
Real Life Applications of the Cynefin Framework
Jeffrey: So I guess this is interesting because I think there is an element here where people may suddenly find themselves in chaos when simple things don’t work as expected. And I expect that sometimes as happens when people are treating complex situations as simple and what they what they receive, the opportunity, it feels like chaos. I would imagine for a lot of our listeners that the items that would feel most like cast would have been something like with an outage where there’s a lot of pressure, a lot of uncertainty about what to do. And because one of the outcomes of chaos is that you’re off script. You know, you’re run book. If you’re run book is working, then you’re not in chaos. You know, you’re either you know, you’re either in a simple one. You’re just following the written down steps or you’re in a complicated where you know, we have a blueprint that tells us how to do how to react here. It’s when it’s when suddenly it’s like, no, we can’t make sense of what’s happening, that you realize that you’re in chaos and you just need to try to act to stabilize the situation. What often happens is you’re inventing what we call novel practice. You’re trying something for the first time because, well, nothing else is working. So we’ve never tried this before, but let’s see if it works. And it’s kind of unfortunate when that one of the questions that people will be debating as to how much could we have been better prepared? How much could experts have prepared contingency plans, have blueprints to try and ready to go waiting rather than at least for some places being caught completely- seemingly completely unprepared and then reacting in a chaotic way where they’re acting because they are just trying to stabilize the situation, but they don’t really have a thought out plan.
Squirrel: Mm hmm.
Jeffrey: I think people can probably think of examples in the news where that will where that will come to mind as what seem to be happening.
Squirrel: And one thing that the Cynefin Framework helped me to do was to have at least a little more empathy for people in those situations, because in a lot of cases when I applied the kind of the complex lens, I was able to see that they were they were probing and there was some sensibility to their probing. It was pretty painful probing. I wish they’d had better luck in their probing to find better answers more quickly. But in many cases, I could have more empathy and more understanding that there was some reason to the probing. It wasn’t purely chaotic. So that was encouraging to me and I hope it might be to some of our listeners. And of course, that they can apply it in more normal times and more normal situations. And many of us are continuing with our work from home in situations where their agile team has an outage or delivers something that didn’t work or suddenly all everybody has to work from home. And you’re trying to figure out is working from home Chaotic, complex, complicated, I’m not sure. You’re here in the domain of disorder. And I hope this framework might help you to figure out new and different situations like those ones.
Jeffrey: One thing I really hope it helps you with is the idea that that these different. The first one is that the framework helps is that these different domains exist. And so if you’re in a situation where either with your work or you’re looking at what’s happening, you’re saying, I don’t understand why people don’t just X, Y, Z. You know, how come they haven’t figured out, you know, ABC? It could be that you’re asking for simpler or complicated solutions to problems that are actually complex. And when you try to respond to a complex situation as though it were simple or complicated, you almost inevitably end up falling into chaos. And so understanding that the domain you’re in is important to avoid bad outcomes. And I’m sure that’s something we’ll be talking about in the future in our future podcasts, because the way this appears in the book comes down to human interactions, which is, a lots of times people make the mistake of treating them as simple when actually human interactions are inherently complex.
Squirrel: All right. That makes perfect sense to me, and I’m sure we will be talking about it more because we love human interactions and figuring out how to improve conversations in order to to have better ones. Do you want to take us out, Jeffrey?
Jeffrey: Absolutely. As we said earlier, if the people who are familiar with Cynefin Framework, we’d really know what you think. Did you think that we mapped things correctly? Or do you think we we missed some important elements. We’d really love to hear from you. And if you’ve tried applying this framework, followed our links and looked at the diagrams and gotten maybe done some reading and start apply it to your situation, have you found it helpful? We would love to hear from you about that. And you can get back to us either on Twitter or you can go to ConversationalTransformation.com and go to the troubleshooting agile podcast link. And we have a button there you can hit to send us an e-mail. We’d really like to hear your stories.
Squirrel: Excellent. And of course, we also like it when people subscribe. So we’re we’re here every Wednesday and we’re going to continue. We do these from our home anyway. So it’s particularly easy to mosey over to the microphone and make one. So you’ll be hearing from us regularly. Despite the current difficulties. All right.
Squirrel: Thanks Jeffrey.
Jeffrey: Thanks Squirrel.