 Let's get this show started and our first speaker then is Dave Snowden. We are so fantastic proud to have you here. Creator of the Kinevin framework. Welcome up. Dave, you're the founder and chief scientific officer at Cognitive Edge and your work is international in nature and covers government in industry looking at complex issues relating to strategy and organizations drawing out on anthropology, neuroscience and complex adaptive systems theory. Very popular, passionate speaker, complex speaker, maybe. On a range of subjects and you are well known for your pragmatic cynicism and iconic iconoclastic style should be. Right? Take it away. Firstly apologies, I won't be here tomorrow but that gives me the great delight of saying but don't worry Sonja will explain it in the morning. All right, so you'll see a lot of that coming through. Sonja and I worked together for about 15 years and I picked the dog theme deliberately but we'll come back to that one. Second thing is I'm going to be around for there's a panel session I think this afternoon. I'll be around on the Agile 42 stand this afternoon but I won't be here this evening. I've got to go on to another event tomorrow so apologies for that but I'll make myself as available as I can. What I want to do is to start off by talking a little bit about where Agile is at the moment. A lot of us are really concerned that what's as Agile starts to move out of the IT development space into HR and marketing is taking the worst aspects of the last 10 years rather than the original inspiration and I'll talk a bit about that as we go through that. I then want to go on to what for me is key which is basically taking a natural science based approach to social systems. That's what I've worked on for the past 15 or 20 years and I'll explain the basis of that. And I'm going to look at a few things. I'm going to look at the role of failure in human cognition. I'm going to look at cognitive bias and I'm going to look at complexity and I'll wrap that up. So that's the idea to do some positioning. I want to start over this that's a wolf. It's a gray wolf. Anybody who grew up in the jungle book will recognize the gray wolf and we'll finish up with just those stories so I'm getting symmetry around Kipling here. If you actually look at the canine species in the wild there are about five or six types. There's very little variation and they're highly resilient. If you look at what happens with domestication within seven generations of selective breeding we get that sort of nonsense. It's highly unresilient. We get individual species which require major care just to allow them to breathe properly. My father was a veterinary surgeon and he had to put a lot of these down but he really felt that the breeders and owners should have been executed first. Some of my attitudes come from my father in that respect. The problem with domestication is it destroys a lot of the resilience in the gene pool because it produces a high level of specialization and that's where our job is at the moment. I'm still trying to understand the debates between Scrum and Kanban because both of them make the fundamental error of thinking it's all about clearing cards pinned to a wall rather than delivering value. How you can see any difference between the two I just don't understand. Just giving you a flavor how this is going to go on. Let's look at some of the issues. This is a cartoon from Gape and Void. I'll talk a bit about Gape and Void later. We just completed a major project with them to identify 49 archetypal cartoons about agile culture which we're about to use in agile readiness surveys but I'll talk about that at the end. This actually came out this morning last night so it was highly appropriate. I don't buy the natural law thing but I've taken the law to science. Basically the problem we got is we do know things about nature which are true. We know things about cognition. Most of the time what we do is socially construct meaning around what we want to believe. That's a real problem. It's impacted agile in a major way. Let's look at a couple of these. First of all if you look at the original adoption of agile if you go back to the agile manifesto XP dominates the meeting in Snowbird far more than Scrum. In fact Scrum is an afterthought. The thing which drives agile to scalability and this is good and bad. Agile would never have grown if it had stayed with XP. XP people are brilliant but they can't talk to ordinary human beings. Therefore there's a problem on scaling. When Ken puts it together as Scrum suddenly it's got structure. We got two week sprints. We got project roles. All of a sudden we can now scale because people can see something. The trouble is that then becomes doctrine and we get into another set of problems that go with that. Scrum was really important without that agile would not have taken off and it's still highly useful. But we get the certification scams. The idea that you can call yourself a master of anything having done a two day course and filled out a multi-choice questionnaire in an open book exam over the following eight weeks with other adults present is a travesty. The pyramid planning scheme which got safe started. Do my four day course and pay me a fee and you can deliver the three day course provided you pay me another fee. These are actually a disgrace to the actual original principles of agile and we're still carrying on with it. I'm now seeing people certifying people as HR coaches in agile on the basis of a two to three day course. The pattern is continuing. Nothing wrong with structure to get things started. Everything wrong with structure when you make it an ideological imposition here and you start to create certification. We're actually launching our own certification program. He says not with hypocrisy. But it's actually a three year process in which your participation in wiki and your colleagues opinions of whether you contribute or not a part of the certification. Yeah, you can't grant somebody competence in less than two to three years. So bit of cognitive neuroscience. Number one in Australia. They don't let kids drive cars with other than elderly relatives for the first two years after they passed their test. And there's a good reason for this. It takes two to three years for the brain and the body to co evolve to the point where you can drive without paying attention. Is actually a very good safety move. The same is true of coding coding is not just something you do in your brain. It's something you do in your fingers. It takes you two or three years before you really acquire skills to the point where you can use them in a masterly or professional way. So I'm throwing that I went out because we need to start to think about this and a certification scan is a major problem. More money is made out of training and certification in Agile than is made out of producing quality code for users. Yeah, that's kind of like the emphasis next one. We got too many recipe book users and not enough chefs. Yeah, there's nothing wrong with recipe books. You know, they how you get started, but if you haven't got the right ingredients and you haven't got the right equipment, you're lost. A chef can produce a wonderful meal with whatever they happen to find in your kitchen. Yeah, and we develop in a factor. I've seen that, for example, the last time I looked at the safe certification. You basically have people reading tax with slides and say in their training. We just completed our latest train the trainer. We don't train people in any of our slides. We train them in the principles which underpin the slides because a trainer should know far more than the material they give in to the users. You need to go deeper in that sense. Next one, yeah, starting at the end of someone else's journey. If I hear one more person say they're adopting the Spotify model. Then I feel that not only have they got the same design features as the Vasa, but they deserve the same end, right? The reality is that you can't copy the end point of a journey Spotify themselves said. Yeah, individual companies have different journeys. Companies actually often do innovative stuff on self organization because they're starting from scratch. It doesn't mean you can replicate what they did in a company which is mature in a different way. Yeah, we need to start to look at underlying principles rather than following recipes that same point again. But basically, starting journeys allow you to discover novelty and adapt to context. Copy and what somebody else did over a 15 year process based on what they think they remember they did is actually rather dubious. Confusing correlation with causation. I went past Nobel Prize Nobel Prize Museum yesterday. If Sweden wants to increase the number of Nobel Prizes it wins or it has to do is increase the amount of dark chocolate consumed. If you don't know it, dark chocolate consumption for head of population directly correlates with Nobel Prizes per head of population over the last 20 years. All right, so there's an obvious causal impact. Isn't there go eat more chocolate? If you don't know it, drownings by suicide coincided in their peaks with the release of Nicholas Cage movies. But I can see a causal connection there. 99% of all know management books confuse correlation with causation. They grab five or six cases. They believe what people tell them about those cases and from that they create a recipe and say do this it will follow. It's a confusion of correlation with causation. If you look at the two worst books in the agile movement, one is lean startup. Anybody read that one? Yeah, it goes and studies friends of his who succeeded and believes what they tell them and writes a recipe book based on that. He doesn't bother to study the people who failed. If he actually did he discover what we discovered in IBM with Dorothy Leonard of Harvard Business School. All the companies who failed do exactly the same things as the company succeed. There is no difference. The point is you've got a market. There are so many players. Some are bound to succeed. But of course it's sold and paddled as a recipe. Do this you will be successful. I could equally say because there are more failures than successes. Do this and you're doomed to fail. The other one is Lacroix's book reinventing the organization. Anybody read that bunch of twaddle? Okay, that's what we actually call an ideology seeking cases to justify itself. One of its main cases Zappos has only succeeded in imposing a so-called self-organising system by a draconian CEO imposing it and firing. Does that unlike that follow? He's got an ideological position and he finds aspects of cases which back that ideological position. He doesn't even investigate his cases properly, but people want to believe it. Therefore they listen. The chairman started off by introducing the political crisis. We're in worldwide at the moment. People are listening to what they want to believe not investigating. But I'm sorry to say this, the liberal elite in the agile community are as guilty in that respect. It's a human tendency and we need to get more disciplined about this. And the final line, there's a deference there to agile 42 by the way. It is actually mostly harmless and if you don't realise that's an insult you haven't read Hitchhiker's Guide to the Galaxy properly. Okay, this is the natural science concepts. Let's start to hit them. First I use this a lot of confidence because it makes the point. Radiologists actually have four or five years of training unlike two to three day courses. They on average have 10 to 15 years experience of actually reading and looking at x-rays rather than coaching other people to do things rather than doing it themselves. So this is a highly disciplined profession. You give them a batch of x-rays and on ask them to look for anomalies. On the final x-ray you put a picture of a gorilla which is 48 times the size of an average character nodule. 83% don't see it even though the rise physically scan it. And the 17% who do see it come to believe they were wrong when they talked to the 83% who did. This is called inattentional blindness. We do not see what we do not expect to see. And the only people who are exempt from this are people who are fully autistic and of course they can't operate. The reality is there's a spectrum between the one end full autism where you see everything therefore you're overwhelmed. And a lot of IT people at IT people are towards that end of the spectrum. They like structured material. At the other end you've got what's lack hand called pervers visance. The like of Trump where everything that happens has to fit his perspective on himself. Either filtering is no exo extreme. It doesn't permit a variation. All of us, this is Andy Clark's work, sits somewhere on that spectrum and we move around in different contexts. In a way we're all mentally ill under the conventional definition. Because we have a filtering mechanism. You only take into account about 5% of what's available to you. That's the most you scan if you're really focused. 10% of you are Chinese. There are different cognitive developments based around language types. But for most people here it's 5%. But most of the time it's 2 or 3%. You then match that partial data scan against patterns stored in your brain, your body, your social interactions and your tools. Consciousness is a function of all of those. It's not just the brain. And you do a first fit pattern match, not a best fit pattern match. So you ace in a partial data scan, you use the first pattern which fits. And in evolutionary terms you can see why this happens. Think about the first hominoids on the savannahs of Africa. Something large and yellow with very sharp teeth runs towards you at high speed. Do you want to artistically scan all available data? Look up a catalogue of the flora and fauna of the African belt. And then reference having identified a lion. Reference your certification program notes on how to avoid lions. By that time the only document of any use to you will be the book of Jonah from the Old Testament. Which is the only example I found of how to escape from the digestive tract of a large carnivore written by a survivor. We evolved to make decisions very, very quickly based on a partial data scan, privilege and our most recent experiences. That's how you all make decisions. So anybody who thinks they're a professional systems analyst is severely deluded. You will actually only hear the things that match your expectations when you go into the meeting. If you conduct more than two interviews, your brain forms a subconscious hypothesis and you literally only hear things that match that hypothesis thereafter. You see why we have problems? Collectively we're actually very good. Yeah, if you have cognitive by behavioral and neurological diversity within a team and you prepare to listen to it, we're extremely effective as a species. But the focus on the individual in individual change is bad biology as well as bad sociology. In terms of the way we work. I say there are implications for this and you can work on that image. Most of our job is actually a sheep in wolves clothing, right rather than the other way around. Think about that one. It's a good one. So there's some clear implications here. First of all, we don't see what we don't expect to see right. The solution to that is you want multiple sensors working independently of each other and look for patterns in that sense of feedback mechanism. I'll give you some examples of this later. We're now using the whole of a workforce to assess business situations in real time in a non gameable way. So we can find the seventeen percent of employees who've seen a gorilla where everybody else is ignoring it. So once you get the science, you can institute the practice. If you just try and derive the science from the practice with a very limited data sets you've got, you never get anywhere. I should declare a prejudice here by the way my first degree is physics and philosophy is a joint major, which meant I was trained to have a contempt for social science from two completely different disciplines. And to be honest, experience hasn't shaken off any of that contempt. No social scientist ever has enough data to make any reasonable conclusion. Yeah, doesn't mean we can't learn from it, but we can't draw causal links. So that's key. Secondly, we need to start to realize that hindsight is not the same thing as foresight. Have you noticed whenever anybody looked backwards after a natural disaster or a major company failure, everybody can see why it happened, but they carry on getting it wrong the next time. Economists are particularly prone to this. The idea that economy is a science, I find one of the most hysterical propositions of the modern age. Fundamentally, we actually reinterpret the past to meet the political needs of the present. That's what all human beings do. So if a team has succeeded and we actually did experiments on this in IBM, we did lessons learned the day after a team knew whether they'd won an outsourcing contract or not. And we did the same process the day before they knew and we compared the results. Completely different histories. The history before people knew the outcome was rich and contingent. The history if they succeeded was narrow based on their own brilliance and their own logical processes. And if they failed, it was somebody else's fault. All of them were well-intentioned, identical process. The way we remember things is based on the context in which we're asked to remember them. Now there's another implication I haven't gotten the slide by way here. If you want to capture lessons learned, you capture it as people do things. You never do it in a workshop after the event. Moving to real time retrospectives is probably more important than instituting a perspective of retrospectives. Because we've been 24 hours, we'll actually change the way we know things. Next one. Ritual works better than rules to change human behavior. Because it changes the patterns that are available to you. I'll give you a safety example, famous one. First time we did this. New Zealand lorry drivers were having a massive amount of accidents in the first 10 minutes after they arrived at the key to unload their lorry. The reason is their entrained pattern was lorry driver. If you've driven for 200 miles, you're thinking like a driver. It takes 10 or 15 minutes for the brain and the body to reset to think like an unloader. So what we did was to give them heated belts. You can't unload the lorry till you strap on this heated weight lifters belt. It's a ritual, it's a physical act that triggered a switch of the patterns and we actually reduced the accidents to about 15% of what they had before. We didn't write rules or put up motivational posters. We basically did a ritual to change the perspective. Think about what happens. I had to put a dinner jacket in an agile conference in Lisbon recently. This is a unique occurrence having to wear a dinner jacket in an agile conference. I felt radically different after I put the bow tie on them before. Changing your dress changes the way you think. Start to think about this for the OZV in DevOps. Start to think about the different cognitive style you need to test code as opposed to develop it. And start to realize you need some ritual transformation between the team so you're using a different path for your brain and body to actually scan the data. Ritual is one of the most important things along with habits. And then the other thing a model I created some time ago is called C attend act. Most management theory and HR people are particularly prone to this, believes if you have the right people with the right competences, the right training in the right context with the right information, they will make the right decisions. If that doesn't work, they blame culture or go on a leadership course. The reality is how we see the data, whether we pay attention to the data, whether we will act on the data at three completely different processes. You can't assume if people have the right information, they will pay attention to it. And even if they can pay attention to it, they may not act on it. To give an example, we were working in DC around 9-11. And one of the big things we had to do after 9-11 is to go through complexity-based exercises with the former White House al-Qaeda team. There were some fascinating things on that, but one of the questions I asked, because if you don't know, there was a congressional order, which would have been signed by Gore if he hadn't been displaced by a judicial coup. Sorry, I'm political on this one. We should have had F-14s in permanent patrol above Washington, New York with authority to actually shoot down civilian aircraft. Yeah, the Clinton White House had worked out al-Qaeda's strategy. That was kind of like abandoned by the Bush White House. He was considered Democrat pioneer about al-Qaeda. Now, I remember asking the question of Gore and said, would you have actually signed that order? And he thought about it and he said probably not. Because the context of 9-11 means shooting down hijacked civilian aircraft is acceptable. The context before 9-11 in the mistakes in the Middle East and the Korean airliner is actually you couldn't do it. Now, understanding that difference is key. There are some things executives want to do, but the context doesn't allow them to do it even if they think it's right. Yeah, context is everything. And you need to separate what we see, what we attend and what we act in terms of the way you build processes. And again, ritual can help on that perspective come. Next major factor, how many people have got children? Okay, how many people read them bedtime stories? All right, does your bedtime story start off with Janet and John stayed at home, did what mummy and daddy said, got their childhood certification, attendance certificates and achieved the family's velocity ratio. Anybody do that one? Or do you do? Janet and John went into the wild woods, encountered evil wolves, witches, goblins, sprites, trolls, etc. Of course, we have a happy ending because we want them to sleep at night. But the reality is we tell stories of failure. What stories passes around the workshop? Stories of failure, a water cooler. What stories to the press pickup? Stories of failure. Yeah, trying to tell people not to do this challenges 100,000 years of evolutionary history, which says avoidance of failure is a more successful strategy than imitation of success. Sorry, I've gone too far on that. Basically avoidance of failure is a more successful strategy than imitation of success. We actually build worse practice systems for companies. If I go on the safety work, the story of how Bill got to strap his harness onto the gantry, as a result of which he fell off the gantry and if he hadn't fallen on his friend and broken his friend's leg he wouldn't have survived. Everybody knows that story. Nobody knows the story about the person who clipped on his safety harness and was safe because that's boring. Teaching stories tend to be stories of survived failure, not stories of success. Yeah, one of the ways you actually create organization director is not to say what you want to be, but to say what you don't want to be. If you want a self-organizing system, you have to have constraints. Constraints which identify negative patterns but allow freedom to deform the positive are much better than ones that create platitudes as a positive outcome. Negative stories are more valuable, they provide more learning. Let's say there's some implications for that. One is the one I've just said. Second, yeah, worse practice, not best practice. There are another interesting thing if you don't know it, cave painting and music come before language in human evolution. It's unique to humans out of all the mammals. As a result of which our languages are not limited by Dunbar's numbers to a limited vocabulary because our language evolved from abstractions. Now it turns out that's key to invention. That's why it survives and develops. If you move up a level of abstraction, you start to make connections between things in novel and exciting ways. If you stay in the concrete, you never make those connections. Now start to think about that because actually a focus on STEM education potentially damages a nation's ability to be inventive. Yeah, art is evolved in human beings for a very good reason. It makes us tremendously resilient and adaptive and we need to work with it. And again, we can start to look at that. I'll talk later about the work we do on high abstraction signifier sets. Finally, failure games. So the most effective work we've done in ministry decision making is set up games in which everybody fails and no matter what they do, they carry on failing. And as they fail, we gather the stories about their failure and what they've learned from it and that becomes a living database to deal with current reality. So actually we do scenario planning and contingency planning as part of the training. Putting leaders through a series of failure games, it's called anthra simulation. At the end of it, they're scanning 30 to 40 to more times data yet when they make a decision then they are at the start. If they succeed, they scan less data. If they fail, they scan more. So training people to succeed will reduce their cognitive capability. Putting people in situations where they consistently fail will increase their cognitive capability which is by the way the basis of apprentice models and everything else. Now all of this is actually common sense but we've forgotten about it. Now it leads me on to complexity. I like this image because kind of like you know there's some sort of coherence in that web of fishing tackle but you can't work it out. And if you try it apart, it gets worse. Everybody is undisciplined people who don't know how to properly coil cables. I should be taken out and grotted. The trouble is the cables are so broken by the way they've bestore them that they break in the act of grotting so you couldn't do it. Sorry there's a personal one in this one. Basically when things get entangled they're difficult to separate but it doesn't mean there isn't coherence. Complexity is about systems where you have deep entanglement between multiple factors. There are no clear pathways between cause and effect. If you pull one of those ropes every time you pull it the entanglement has changed so each pull will produce a different effect. There's no stability within the system. And the most human systems are complex adaptive systems. They have multiple interacting causal factors. They all work in radically different ways. I'll go on to this in a minute but just hold that. The key thing to understand about the complex adaptive system is it's not causal. There is no linear relationship between cause and effect. It's dispositional. I can only ever say there is a probability that this thing will happen or there is a probability that this thing won't happen but I can never say if you do this it will have that result. Now again undermines a huge amount of agile practice. Doesn't mean it's not valid but it undermines it in many ways. So again implications, retrospective coherence. Here's the classic case on this, right? If I have four dots, you've all heard the phrase why didn't we join at the dots? Everybody heard that? Okay four dots, one, two, three, four, five, six linkages. Four dots, six linkages, 64 possible patterns. If I got to ten dots, how many patterns do you think there are? Gassis please, I don't want a form and I want a number. 1,024, how many people think that's too low? Too high? Okay it's too low, anybody want to go with another one? No. A million, okay how many people think a million is too low? Too high? It's too low you get in the pattern, yeah? It's actually greater than 3.4 trillion. If I go up to 12 it's greater than 4.8 quadrillion. If you know your ancient stories, it's the emperor's chess board. A sage asked for a reward by the emperor. Said I have one grain of rice on the first square of the chess board and double thereafter there isn't enough rice in China. How many dots are there in the human system? How many possible patterns are there? Are you ever going to believe any book based on cases ever again? Yeah, not only have you got inattentional bindness but you've got people joining at the dots after the event anybody can join at the dots. But it doesn't give you the ability to do it again in the future. I like the rear view mirror image there. Next one, yeah talk about that one. The other key thing and this comes into the original agile manifesto. They don't think they were aware of complexity theory. You know the interactions are more than things. How things interact is far more important than what they are. Now that actually has major implications for HR practice. HR practice constantly focuses on changing individuals which is ethically dubious and biologically inefficient. Yeah which one you think is more important I leave to your conscious but both of them undermine it. The reality is by changing social interactions you can change people far faster than by try to change the people. Yeah linkages are far more important than people and the idea of competences by the way is really bad cognitive science. There are no such things as innate competences. And if anybody done Myers-Briggs. Okay I scientifically proved in IBM that astrology was more accurate than Myers-Briggs in predicting team behavior. Over a six month experiment. They didn't quite forgive me for that one. But astrology at least have several thousands years of deceiving people behind it where it's only a hundred years old with Myers-Briggs. Okay so hold that one because this has major implications. All right we tend to look at agile people I say people haven't got the agile mindset. That's nonsense. People are not linking and connecting in a way which will lead to agile behaviors. Trying to change people's mindset is ethically dubious and biologically ineffective to come back to that phrase. If you change his linkages interactions people change with that. Now that's probably one of the biggest things I'm going to tell you today one of the most important things that we know. Okay. So let's basically do with Kinevin. I haven't got to belong on this but I just want to draw it quickly all right. I'm going to do this because there's a lot of people using Kinevin in agile at the moment you've got codefin you've got Liz Keo. But I just want to draw something to make a point all right. That by the way is if you don't know it. That's a picture on the top of Glidavar in Wales in mist all right. Try and navigate off that without a compass and you're in trouble. There's a metaphor in there. So basically Kinevin as a framework looks like this. Destroy in purple. Now remember I said we have a world which is complex. That's a world with many intertwining causes and effect. It's a constantly entangled everything you do changes the system. Now one of the things about a complex space is you have massive conflict. Because lots of people have got good ideas about what you should do but you can't resolve who's right on an evidence base but there isn't sufficient evidence. The way you approach this and this is a major thing we do on peace and reconciliation is you don't test whether somebody is likely to be right. You test for whether their hypothesis is coherent or not because to agree that your hypothesis is coherent doesn't involve me in agreeing that you're right. So I radically reduced that pressure up front. Instead of have to select between eight ideas because you've only got money for one you give a small amount of money to any of the ideas which proves it's coherent. And you run them effectively as what are called safe to fade experiments. You probe, you sense, respond. Child people keep getting wrong. They think this means you do one iteration, you don't. You do multiple experiments in parallel. But if you do one experiment you'll suffer Hawthorne effect if you don't know that. Anything novel always works the first time because human beings respond to it but it won't scale. So you do small experiments in parallel which actually means practice is what's called X-aptiv. X-aptiv is a biological phase so a dinosaur's feathers evolve for sexual display but then when dinosaurs start to fall off cliffs by accident they evolve into feathers for flight. You can't get flight immediately because if dinosaurs jumped off cliffs in the hope of developing feathers before they hit the ground they probably wouldn't get a chance to breed or only in some sort of broken limbd agony. The basic fact is that it's called radical repurposing. It's like magnetos on microwave ovens were repurposed for, sorry, magnetos on radar machines were repurposed for microwave ovens when somebody noticed a chocolate bar melted in their pocket. X-aptiv learning is rapid repurposing of existing capability for something which is novel. And there's ways to design for that. IBM dominated the computer industry for 50 years because they repurposed punch card machines in which they were experts to actually create an operating system and we still haven't got away from that with things like Windows, right, in terms of the way it works. So that's fine, that's complex. We then get two types of order. One's where the relationship because the effect is obvious which means effectively we've got rigid constraints and what we actually do is we sense categorised response. I mean Sweden, you drive on the right, I drive on the right. I mean the UK, we drive on the left, you drive on the left. It's self-evident, everybody, you do it. That's the range of best practice. It's a valuable space. Komplicated. It's actually where I have to analyse sense respond because I don't actually know what's going on. It's not self-evident. Both of these domains is there is a right thing to do and you can discover it. In here we have good practice not best practice. And then we get chaos which is a state of no constraint so I act sense respond and practice is completely novel. We also have this which is called the domain of disorder which is a domain where you don't know which domain you're in which is where most people start. So actually you've only got a one in four chance of getting it right. Yes, that's a bad place to be. The revolutionary idea behind Kinevin by the way is you use different methods in different contexts and the methods may actually contradict each other. A method which works for complex may be a disaster in complicated and vice versa. I've done that quickly because the real point I wanted to make here was the introduction lately of the liminal domains of Kinevin which looked like this which effectively a transition devices. This one here, the liminal zone into chaos is where you deliberately remove all constraints so that novelty can emerge and that's very expensive. Chaos is only ever a temporary space. If anybody presents Kinevin and puts things like traffic in it they don't understand the framework. Chaos is temporary, the absence of constraints is difficult to maintain. If you can maintain it you get complete novelty. Is there a light nuclear fission? The energy for the magnetic fields to contain the plasma is actually more than the energy you get out of it. So you've got to be careful about that. This one here is the most important one for Agile. So this is where I've done my parallel safe to fail experiments and now I'm pretty sure this is the right direction but I need to get it right so I'm going to go a series of iterations to check my assumptions. That's called Scrum. Scrum's great strength is it's a liminal technique. The trouble is you need things you do before Scrum to decide what goes into it. And that's been one of the big problems with Agile. And I'll just list three and I will, Sanya will go through these tomorrow for anybody interested. Sorry Sanya, this is the first of many. One is what we call trios. That's where you actually put a pair programming team together with the user trained to talk to IT people. It's a lot easier to train users to talk to IT people than train IT people to understand users. Repurposing, oh your child is autistic. How do you handle it training material? It's quite easy really. Basically you send 15 or 16 trios to look at a problem and see what comes back. That costs you significantly less than an analyst interviewing people and it introduces cognitive diversity into the system. We're doing this with universities, engineering, humanities and pure sciences. When you join the university you have to form a trio with the other two faculties and you have jobs to do every term which you get rewards for but we're building a network which is transdisciplinary. Indirectly and that's another key principle. What we call triple eight. That's where you put a rapid design team on a process for eight hours then you throw the output over to another team or an eight hour time difference and they throw it on to another team or an eight hour time difference and teams two and three are not allowed access to the original user need. Deliberately forcing mutation. Every time we do it, it comes back, the user say God I wouldn't have thought of that, can I please have it? You see what we're doing here and the other big one is an articulated need mapping where you capture people's day to day experiences and when they cluster significantly I'll show you a cluster mapping a minute then you put a prototype team on it you never go and ask people what they want. Now just make this point and let's say Sanya will go into this more tomorrow. One of the big issues we got at the moment is technology is capable of doing things that users don't know what to ask for and technology experts have a narrow focus on what they think their technology is about. They don't realize how it could be repurposed. So any process which is based on a request followed by completion of the request the basis of all agile techniques misses 83% of the available data in terms of the way it works at least. So what I'm arguing for here is pre-scrum techniques in that pre-com techniques that applies to HR and elsewhere. So I wanted to introduce Caneving just to get that point across that the transitions between the domains and more important domains and we need a multi-methods approach. Anybody who argues you do Kanban, you do Scrum, you do this, you do that is basically taking what's called a mono ontological approach. They're assuming one type of system and therefore they're bound to get things wrong when they apply that technique in another system. A multi-methods approach doesn't mean that everything is equally valid. I think that's if you look at Agnostic Agile make this mistake as well. Some things are plainly wrong and some things are right in one context and wrong in another context. And we need to understand that in terms of the way it works. So moving on and starting to reach to a conclusion. Other things that you do. Mapping constraints is one of the most important things. Complex systems are defined by the constraints in play. Because you only ever make decisions based on how you've already decided to act. You all know that one? All human beings when they assess a situation have already subconsciously decided what to do. So they only see the things in that situation which justify their feeling. Now getting them to sit round in circles in Kumbaya Agile sessions saying we'll be open to people's other ideas may be emotionally satisfying for facilitators who desperately need workshops to satisfy their emotional needs but it doesn't make any difference. The reality is in a workshop anybody will say anything but reality changes matters. You just can't do it that way. Deal with reality rather than anything else. The whole point here is that we need to assess a situation in such a way that people aren't thinking about the action using the science. So what we do, I guess one thing Sonja will teach tomorrow is constraint mapping. You don't talk about what we want you map the constraints and there's a whole typology of constraints. And you say if I change that constraint might it produce a good result and you do a safe to fail experiment. And only when you start to see a pattern emerging do you go into a conventional process. And here are some aspects of the typology. Constraints can contain or they can connect. They can be resilient or they can be robust. Think about a sea wall. I don't use this in the Netherlands at the moment because it's a source of impending national grief. I mean se was a wonderful things. You can drain the land on one side. The trouble is when the water rises to a certain level or the storm reaches a certain ferocity. The wall will break. And it will be better if you'd never had it in the first place. Yeah, on the other hand the salt marsh absorbs a huge amount of water. It's not very efficient because you can't farm the land but it carries on holding that water even when it's saturated and that's resilience. So you might, for example, say this fixed constraint, let's make it semi permeable and see what happens. You see where we go with constraint mapping? It's actually you can do a lot of this on UX and system design as well. And finally, scaffolding or permanent. I'll come back to that in a minute. Mapping dispositional states. And I'm giving you kind of like the one, two, three things we now do. First of all, map the constraints. Start talking about where you want to be or idealistic goals. Map the constraints and ask yourself three questions. Which of these can I change? Out of the ones that I can change, where can I monitor the impact of change? And out of the ones where I can monitor the impact, where by failure would it teach me something and I could recover or by success I could amplify it quickly. That makes life very simple but not simplistic. And that's a very important distinction in English, all right? Being simple but not being simplistic. Yeah, talking about values that individual members of a team should have is being simplistic and idealistic. Yeah, our values are a consequence of our interactions. They don't determine our interactions. The causal chain goes the other way round. How we behave determines what we are. It's not what we are determines how we behave. And we're deeply contextual in terms of the way we work. So this is a dispositional map. This is done by getting a huge amount of narrative or observational material in self-interpreted at the point of origin into what's called a high abstraction metadata structure. These are structures that you can't gain because you don't know what the right answer is. Remember the abstraction point? We shift people up. Sometimes we'll use drawings, yeah, rather than words. Now, this is actually a safety case which I use a lot because it illustrates the point. This dimension here, sorry, changed the colour, is rule compliance. This dimension here is job completion. This is a real case. This is civil manufacture. This is military manufacture. Now, you can immediately see the problem on the left. You see a dispositional map. I'm not implying cause and effect. I'm showing a dispositional state. This basically shows there are two dominant patterns here comply with the rules or get the job done and then mutually incompatible states. Now, this is actually more common than you think. The right hand one looks better because you've got that top coaster. People are following the rules and getting the job done. But when you click on that and look at the underlying narrative, you discover it's nuclear weapons testing. Well, nuclear weapons testing creates an existential quality to following the rules and getting the job done, which you can't replicate elsewhere in the business. We then got get the job done, ignore the rules and we got this one, which is actually really bad. I've given up. This, by the way, is the pattern we're seeing among nursing staff in hospitals. In a crisis, they're brilliant. On a day-to-day basis, the measurement regime means they have to break the rules, provide empathetic care, but increase doing what to do to survive. Now, the way we change this is not to create a wonderful picture which says wouldn't it be wonderful if we're all up here. Let's have a workshop and all agree that we'll all end up there and we'll put a value statement up on the board composed of the usual platitudes. There's only so many ways you can actually express values. And they're all platitudes. What you actually do is to say this is a stepping stone. It's called an adjacent possible. So what you now do is you click on that and you say, what can I do tomorrow to create more stories like this and fewer stories like those? And that is the new theory of change. I don't talk about what sort of person you should be. I say what do we do tomorrow to create more stories like this, fewer stories like that. We do that with 360 feedback on teams. Don't get your customers to score you on a like card scale because they'll game it. You gather the stories they tell about you in a statistical way and you say what can we as a team do to create more user stories like this and fewer user stories like that. You want to engage human beings. You engage them in changing narrative not in achieving goals. And if you look at that, we're also able from that to go on and identify outliers. If we look at this map here, these people are far more significant than these people. I know about this because it's a dominant pattern. Remember the x-ray? That's a 17% to a scene of gorilla. But everybody else isn't listening to them. So that's why we present infographics about a current problem. We fire out to all employees within half an hour they respond and we draw the maps. So you can find the people who are thinking differently before they become conformed to the rest. Now again, I'm trying to give you some practical tools to finish up on but that's kind of like the point. So three things which come from that and this will conclude. First of all, vectors, remember that I said more like this, fewer like that? Measuring direction and speed of travel from the present is more valuable than measuring goal achievement. And if you don't believe me, look at that quote. That's a better study of all studies of human motivation where people are working for extrinsic goals. It destroys intrinsic motivation. Nothing wrong with extrinsic motivation in some contexts. Everything wrong with it in others. In complex systems granularity matters. You can't scale a complex system by imitation of what worked last time or aggregation into bigger and bigger structures like safe. You scale a complex system the way biology scales by decomposition and recombination into new forms. So actually getting your unit of analysis is key on that and with that we have the scaffolding concept. These are five different types of scaffolding. That's a sort of building type scaffolding and bamboo with raffia is more resilient than steel. That's a sort of scaffolding which is the whole training infrastructure, a kayaker. Yet they're scaffolding in safety by ten years of experience training development of equipment. The scaffolding is built in to the surrounding process. On the top right you've got a nanostructure. That's where actually a fabric is put in the heart which decays away and leaves micro-electronic traces in the heart which will sustain the heart in an organic way. That one is a nanostructure for burn victims. We put a gauze over your skin, a nutrient gauze. The skin regrows around the scaffolding and basically the nutrient dies in the process. And finally the bottom right one that's a keystone until the keystone is in place the structure isn't stable. But once the keystone is there all the other structures are. Stop trying to design organizations and choose scaffolding and put scaffolding in place and see what grows around the scaffolding. That's actually the big new theme in organizational change as you start with a loose scaffolding and see what happens you don't start with an endpoint design. And some cat pictures to finish, right? These are the latest additions to the Snowden household, right? They're called Lyra and Alice. They dominate the space. My wife has bought them several brilliant designer-based beds. Unfortunately my leather satchel is preferred. So I now surrendered the leather satchel to the cats, right? Because they weren't there. You can design anything you want, yeah? But the reality will catch up with you. And those are examples of the stuff we're now doing. You first of all pulse the system and you take the pulse to see what this positional state is. Yeah, then you sense the patterns in it and then you nurture change. You don't design change. It's actually a big difference. And so to finish as I started with Kipling. This is my favorite story in Kipling. It's the cat who walked by himself and all places were alike under him. And it's a story of domestication. The cow, yeah, the dog and the horse surrendered to man. The cat is prepared to look after the baby, provided he doesn't put his tail and generally be a nice cat, but ultimately it's still a wild creature. Yeah, and he remembers species variation in cats is quite limited compared with dogs. Yeah, you need to contain a wild side if you want agility, not agile. And the switch to agility from being agile is probably one of the most important things we can achieve. I hope that was useful. Thank you very much for your time.