Dave Snowden
Speaker: Miljan Bajic 00:40
So the first question Dave here, what does a leader need to develop in order to sense complexity and navigate on it effectively? What in the work has to be done?
Speaker: Dave Snowden 00:56
If I was being particularly pedantic here, I’d say if they don’t already have that knowledge, then they shouldn’t be in a leadership position in the first place, right? So that there is this really strong and I think, really dangerous tendency to try and define competencies. The idea is computer model of the human brain, if this person has these competencies, they can do this. All right. The reality is we can all exhibit leadership in different contexts. So let me be a bit more positive. All right. So there isn’t a one size fits all. This is my big debate. Well, one of my many debates with Steve Denning, when he wrote radical leadership, in that he tried to shoehorn, lots of different leaders working in lots of different contexts into a single model. And say, all radical leaders have to have these qualities now. If you look at GE, Apple [inaudible 01:53] my work with all three of those CEOs, and the only thing that they had in common is they were arrogant bastards. Yeah. And yeah, I don’t see anybody writing a book about arrogant bastard read the secret to leadership success. All right. But that’s the reality of it. So I think I mean, there’s a couple of things we’ve said, I mean, you feel goes got a lot on this. It basically says leadership is about coordination, not decision making.
Yeah. And I was sea level over my life. And one of these you learn the hard way, the more you get promoted, the more you only meet angrier and angrier customers, and the fewer decisions you actually get to make. And you’re even incompetent to make them anyway, because you’re not that close to the field anymore. So your role is really much more coordination linkage, have you talked with this person? Will come back to me when you have. Those are the sorts of things you’re doing, the differences when you get a real crisis, and then you have to make decision to make them very quickly or based on inadequate information. And that’s where we say what you do is you make decisions to increase the options available downstream, you don’t try and resolve the problem. Your role is to increase the options available to create some stabilization, by which your experts can then start to make decisions again. So that’s one thing. The other big thing, the thing I’m working on at the moment, to be honest, some people just seem to have this and some people don’t is what I call anticipatory thinking. And I’d love to find a way of measuring this or training for it.
But some people just seem to be able to do small things now, which make a big difference downstream, even though the link isn’t clear. Yeah, you learn this interesting in mountain navigation, you learn it, I’ll give the illustration. So somebody said to me the other day, all right, how the hell did you find this track? And I said, well, and I suddenly thought, well, how did I find it? I was just obvious. And I thought it through and I said, well, for the last two hours, I was looking at the hill ahead and looking for patterns, because I’ve grown up to do that. And I’m constantly looking ahead and thinking, well, that’s more risky and this more risky and if you get that pattern you get… And that comes with experience. Yeah. So there’s a key framework within [unsure word 04:13] called Ashen, A-S-H-E-N. And that stands for artifacts, skills, heuristics, experience and natural talents. And the way you look at any qualities, you say, well, what are the artifacts? Because artifacts, you can train people to use. Spreadsheets, processes, what are the skills, skills you can train people on? Then you get into heuristics and habits and rituals, which I’ve written about this Christmas. All right, and rituals and habits are ways of reducing the energy cost of knowledge transfer. So they’re normally achieved through repetition. And there’s nothing wrong with that.
That’s actually really important in knowledge. And then you get experience now, very few leaders these days have right experience, when I applied to be a general manager, this is in a software company. I had to do a year in sales, a year in support and a year in production to hit my targets before I was even allowed to be considered as a general manager, right? And if you’ve done a year in sales and you can’t, I mean, I know what it’s like not to be on bloody pay the mortgage because I haven’t made a sale this quarter. You understand selling in a way that you can’t if it’s just abstract. And what we now have is people do an MBA straight out of business school, which I don’t think you should allow, they go and join a big consultancy firm, where everything is about spreadsheets and reports. And then they go sideways into management with no practical experience. So kind of the question you should be asking is, what combination of artifacts skills, heuristics, experience and natural talent that we have? Do we have that and how do we substitute for it? And is quite critical on replacement, by the way, so… Sorry, I’ve just seen you, haven’t seen you for ages. All right. So if I say, how do I replace [unsure word 06:03]? That’s the wrong question. The right question is [unsure word 06:06] has this combination of new artifact skills, heuristics, experience and natural talent. How do we replace that? Alright, that’s a very different way of formulating the question but it’s a way of formulating the question to where you can do something about it.
Speaker: Miljan Bajic 06:24
Yeah. Maybe just to expand on framework and the work that has been done. I think, Thomas, I don’t know if he’s here. But I think what he was probably alluding to is towards cognitive development, then what is the relationship between cognitive capacity and how you sense complexity?
Speaker: Dave Snowden 06:46
I’d be very nervous about that. Because we’re embodied creatures anyway. So I mean, if you read [unsure word 06:54] Seth latest work, he’s basically arguing that consciousness is a process of controlled hallucinations and is based in the body. So we know a lot of our decisions are made. And that’s why I went down the action route. Alright, it’s not a cognitive issue per sey, right? There are some interesting things we can learn from one of the things that you see, for example, military environments is a distinction between NCOs and officers. Okay, you see the same in hospitals with nurses and doctors. And that’s quite interesting, because you have one group of people who acquire experience, then get taught theory. And you get another group of people who start with theory and then get practice. And they work in combinations. So I think is much more about what interactions and experience and context do you need people to live through, rather than trying to define specific cognitive functions, which is protect, and I’m not sure there’s any evidence really to support any context that there are a set of cognitive functions which are ideal.
You also then get into the thing that Nora Bateson and I, which a hitting really heavily at the moment, which regrettably is too common in the coaching movement is adult development models, which are deeply manipulative, right? Which actually have no basis whatsoever in any real science. They all go back to PJ’s experiments, people have tried to replicate them and got completely different results, right? And they ended up privileging the person at the top of the hierarchy. This is for things like spiral dynamics. I remember how this study came back, I turquoise, you’re an angry blue. That’s just a way of avoiding the bloody problem for God’s sake, right? So the reality, all of these interact and work in different ways in different contexts. And there are some contexts where the Army is really good at this. By the way, there are contexts for example, in a weapon Sergeant can out rank a general, right? So military environments have worked out how to delegate authority without loss of status. And that’s what we call a crew, which is something we’ve been taking sideways into industry as well. That to me, you should stop talking about people. And you should talk about roles and role interactions, it’s a much better way of talking about the problem.
Speaker: Miljan Bajic 09:13
So maybe, I don’t know if this is what I’m taking from this but based on what I’m hearing you say is that it’s really comes down to experience as far as dealing better with complexity and…
Speaker: Dave Snowden 09:30
It’s you can do a lot with artifacts. You can do a hell of a lot with artifacts, right? Tools, instruments, processes. I mean, I did a lot on head up display design for fighter pilots when I was a coder. And you do a hell of a lot with structure there to augment congresses. So we can do a lot with those. All right. So it’s not that there is one thing or one set of things, probably the most important phrase that anybody who knows anything about complexity will use a lot is please stop proposing context free solutions in the context Pacific world. Yeah, different things work in different contexts. And what we’ve had for the last 30 or 40 years, is every management movement, including agile has tried to create a context free universal solution. I mean, they were known, for example, our job is to break methods down to their lowest coherent component, and allow them to recombine and combine across different vendors. Because you scale a complex system by decomposition to the lowest level of coherent granularity, and then recombination. So DNA works.
Everybody wants to scale by getting work for [inaudible 10:43] so I’ll do the same thing. And then you get the great error of the Spotify model. Yeah. Which is made worse by the fact that nobody any good in Agile wants to work from McKinsey’s anyway. So they end up with a second rater. Sorry, I’m being deliberately pejorative to make a point here. And then they say, Yeah, adopt the Spotify model. Well Spotify lived through a complex set of journeys, which were different in Stockholm, from New York anyway. Yeah. And some of their practices are constantly shifting and changing, you can’t adopt the outcome of an emergent process, you have to create the same, you create similar starting conditions and see where your journey takes you. And that’s that decomposition and recombination. You learn from the past, but the level at which you learn it is it’s a finely grained level of learning. It’s not a total learning, not a total system learning.
Speaker: Miljan Bajic 11:35
What we seem to be like, one thing feeds another. And this is where companies want these prescriptive frameworks, they want that so like, they’ve been fed. So how do we get to the point where we realize that none of these frameworks, like he said everything needs to be contextualized. And it’s…
Speaker: Dave Snowden 11:55
I mean, it’s interesting, I have the privilege of teaching leadership with Peter Drucker. Alright, which was a huge privilege before he die. Yeah. And one of the reasons I did that is I made the mistake at a conference in San Diego in Hotel Del, and I can still see the situation. I did something a lot of agile people did, I was young and inexperienced. That’s my excuse. And I said complexity to [inaudible 12:19] on Taylorism. And if you ever remember that famous vice-presidential debate, when it’s I knew Kennedy, I got that, right? If you ever been taken apart by a 93-year-old genius on a public platform in front of 2000 people, ended up as a total of humiliation on the stage. He decided I was redeemable so took me out for dinner. And then I actually talked with him for a long period. And one of the conclusions we came to and I stopped criticizing systems thinking, because actually, when people talk about Taylorism, they’re actually talking about systems thinking. They’re talking about all the things which came in in the 80s and 90s. Were things like business process, reengineering and Six Sigma.
If you actually go back to Taylor, and you bother to read Taylor, he was trying to humanize the workforce, if you look at what it was like before Taylor, and we all now look at what Taylor produced and said, that’s terrible. It was a downside better than it came before. Alright. He was trying to humanize it by removing the mechanical side. So what [inaudible 13:19] coming to the conclusion on is complexity theory and scientific management have a lot in common. And they both differ radically from systems thinking, its derivatives, because they both respect human judgment. If you actually go back to Taylorism, management is an apprentice model of management. What happens with systems thinking is an attempt to reduce human judgment completely from the equation and make everything, processes, incompetencies and structures and measurement.
There were no three or five-year plans until systems thinking came in. I mean, the irony of US companies adopting the planning cycles, Soviet Russia, as always, I found that ironic, all right. The reality is you have people with lots of experience who are adapted to things as we went along and did some long-term things and did some risks. And yeah, they brought in new blood from time to time, but the majority of people like Japanese companies, still to this day, will have for life. So they built relationships, and they were committed to the company long term. Now we’re bringing back that type of decision making in the work we are doing on complexity.
Speaker: Miljan Bajic 14:32
So I know last time we had this conversation, I didn’t get a chance to follow up on it, but is it that we butchered the idea of systems thinking because we tried to…?
Speaker: Dave Snowden 14:44
Systems thinking was ontologically flawed from day one. I did a lot with cybernetics and I owe a huge debt to PT check [inaudible 14:54] systems and in terms of where we were with the 80s and 90s, it made a lot of sense, because we didn’t know about convexity theory then. All right. And if you look at it, and you look at systems dynamics, it’s all feedback loops and structure. And these nice little general statement about you should look at the system as a whole, right? And so what you see with systems thinking also comes in, it’s dominated by engineers and by information processes. So things like Ashby and Shannon, that’s where it comes from.
Yeah. And of course, engineers don’t like ambiguity and uncertainty. Yeah, and therefore we get reengineering the corporation and the famous thing at the start of Hammering [unsure name 15:40] book, nothing that has happened in the past has any relevance to the future. That’s what it says. And so what we want is a greenfield site when we’re building on a brownfield site. So this evolutionary, we’re now shifting into these more ecological frames. So from my point of view, there’s a huge dept. I’ve said many times that there’s no way that staff would be able to produce VSM, if he’d known about complexity theory. It’s a brilliant piece of work in the context of what was known at the time. But, so what? Yeah. And people say, well, systems thinking address complexity. Well, yes, it did. The human race addressed the gravity with canals, but then Newton came along. And we understood the science at which point we can do things differently, right?
So, I mean, that’s an ongoing debate, right? But I think the problem is systems thinking is transitionary, right? There’s still things in it which have value but it’s not a universal, right? And it doesn’t handle and it’s quite interesting. Listen to Joel Midgley. I was listening to the other day. Alright, this was [inaudible 16:48] he says, the definition of a system is something which has boundaries and is based on human perception. Well, from a complexity point of view, systems are devised by coherence, not by boundary, some system don’t have boundaries. And we also, this is materialism, we actually know that things actually exist. It’s not just about human perception. So if a human being wants to say a system is something when it’s something different, that’s rather like treating young creationists as they should be, as if their arguments should be accepted seriously. It’s the old phrase used with post modernists is, reality exists, live with it.
Speaker: Miljan Bajic 17:26
Yeah. Let’s transition to I think we could spend more time on this. But essentially, what you’re saying is, focus more on complexity and complexity management rather than the Agile community.
Speaker: Dave Snowden 17:40
The good news is, there’s totally spurious debate between social constructivists and critical realists, both you and the critical realist grew up as counters to social constructivism. Well. [inaudible 17:50], yeah aspects of the builder socially constructed and aspects are, so we got much better science now. And the trouble is, people are holding on to outdated models, they’re not moving on the model and understanding from a scientific point of view.
Speaker: Miljan Bajic 18:09
Okay, well, thank you for taking the time to answer that question. Let’s move on to the next one. What do you think will be the future of corporate strategic approaches? It’s the second one here.
Speaker: Dave Snowden 18:22
First of all, I think you’re going to move from long term planning. So Porter was the arch priest of strategy. Yeah, in the system’s period. And I think in some ways that was a pity. I mean, I think Henry was actually much better. But Porter produced a structure and a recipe which was easier to follow, Mintzberg handled ambiguity better. And I think Mintzberg gets complexity. So I think strategy now starts, there’s two or three things.
First of all, situational assessment needs to be distributed. Again, when things we outlined in the EU Field Guide, is you need your employees to be a human sensor network that you can deploy in real time. Because you need to assess test situation from a multicultural, multiple experience, multi cognitive background, and you need to see patterns and outliers in that. You can’t afford to spend three months commissioning research, because things are changing too quickly. Yeah, the new stuff we’re doing on structure theory in physics, papers just come out on that, is to say, well, this is what we call in the Asteron framework. And the [unsure word 19:34] model is a good one. Because in an [unsure word 17:37], things flow both ways, dependent on the type. And there are sort of granite cliffs which is stable and there are some banks which change constantly. And you often have to read clues from the surface level of the water. So I think the strategy is going to go is into that sort of ecological metaphor of what’s stable, what isn’t stable, how frequent do we need to assess it? Where are the outliers? And then therefore, where do we start to deploy energy? And probably, this is one of the most important things that comes out, certainly my approach to complexity, whatever has the lowest energy gradient will win.
Now, if you want to put that in moral terms, if the cost of virtue is less, is more than the cost of sin, people will sin, right? So in strategy terms, if you want customers to buy your product, the energy costs of buying your product has to be less than the energy cost of buying a competitors. And as I’m saying, energy costs, not necessarily price. And I’ll give you an illustration, is after IBM took us over, which was completely unexpected. I was sent on a mission to explain to IBM salesman why we always ask them to bid. We were a systems integrators, but we never worked with them. And so, happy to do that. So we went. And so what we always asked you to bid because you were always the most expensive, and you gave us the most material, which we could put into our proposals. But we didn’t work with you because you didn’t understand what a systems integrator is about. So you tell us your kit was faster. Well, we know. So Buddy, what? All right?
I mean, we’re going to put this together with lots of other kits, with lots of software, the differences you’re talking about just disappear in the noise, right? Where a son said, if we ever need a faster processor, because the client does it, they’ll just upgrade the processor without an argument. So they’ve shared our risk. So we’re going to go with them, even though the kit isn’t as good as yours, but they’ve taken away our risk. And I said HP put three people into our library and help us write bids. So are you surprised because our bid cost we need to reduce? Are you surprised that we end up with HP kit on the proposal? Because they’ve reduced our energy cost of bidding, I said, you’re not looking at the complete process. You think it’s just produced the better mousetrap. And it isn’t, right? It’s all the relationships and everything about it.
So look at the total energy costs of what you’re trying to do,and manage that ending. And that’s what we’re doing with the Astro mapping is map the energy gradients of the system. So you can see what’s more likely or less likely. And then that becomes the new approach to foresight, is actually to map the evolutionary potential of the present, not forecast the future. Because that way, you can see what’s likely to change and what isn’t likely to change. And it’s also links in with, sorry, I’m throwing a lot of stuff around. But it’s a long day, what we call the frozen two approach to strategy, right? So this will be memorable.
Speaker: Miljan Bajic 22:50
And Jason plays that one.
Speaker: Dave Snowden 22:53
That’s it. So basically, if you haven’t watched frozen two now, this is your excuse to go and watch it even if you haven’t got young children or grandchildren. Frozen two is a great movie, Frozen one is just Disney. Alright, but it was so successful, they have enough money to do frozen two properly, so they had fun. And there’s this wonderful moment in the middle of frozen two where the real heroine of the frozen series is the youngest sister without magic. All right, yeah, left in a position where she thinks her older sister in the snow would have lost.
Sings all I can do is do the next right thing, right? Now is Stuart Kaufman, that’s called the adjacent possible. All you can do in complexity is map where you are, and identify which next steps are coherent. And then you move into those next steps and you look again, so strategy becomes much more contingent. Now if you have to invest over a 15-year cycle, you’re taking bigger bets. And that’s a whole different process.
But for most people, particularly in software development, you’re talking about something which is much more dynamic, because stability is emerging stabilized. That is why we’re going back to a lot of the old stuff for example. But starting to talk about organizational units as objects as well as software. So you define your objects and you define the interactions, so you create stability in those definitions. But then the way that things interact with other things can actually respond very quickly to unexpected circumstances. And that’s called getting the granularity right. So you build your organization in smaller units with defined interactions and with fast feedback loops. So effectively, you’re managing emergence rather than trying to plan forward.
Speaker: Miljan Bajic 24:41
And it seems like you’re also keeping things simple or at least you try to keep it simple, right?
Speaker: Dave Snowden 24:48
It’s why I use children metaphors a lot. I mean, the children’s party story is still the best teaching story I’ve ever created, and it explains complexity. But it also makes a subtle point is everybody manages to impacts in their day to day lives. So you know how to do it. We just forget about it when we walk through the doors of the office.
Speaker: Miljan Bajic 25:07
All right and we manage complexity with simplicity. For most of us I guess, we don’t manage it with [inaudible 25:12].
Speaker: Dave Snowden 25:14
But there’s a big difference between being simple and being simplistic. And too many people confuse the two, right? I mean, it’s actually a big problem in America and the UK, is the anti-intellectualism of management education is really scary. Because if you don’t have people with sound theory, you can’t make things simple. You just go with what worked last time. And that’s been simplistic. It’s called practice, theory informed practice.
Speaker: Miljan Bajic 25:51
Mary Maru, if you could please just put the questions in the tool, I just added going again, but it’s also in the email. Dave, what is your favorite case study or usage of [unsure word 26:07]?
Speaker: Dave Snowden 26:12
As a couple, all right? There’s a really good one in Helsinki. I mean, I got this weird email in sort of Finish English, right? Which said, we’ve been using your framework and I was actually going, so I met them, right? And they used it to understand teenage violence in the Baltic states, it was a brilliant project. They just published a book. And I’m actually quite proud of that, because they did it without me being involved. I mean, they’ve been on training since. There was another one, I got a phone call from the cabinet office in the UK. So that’s the prime minister’s office. And they’d actually use [unsure word 26:52] to explain the role of religion in the Bush White House. That’s a published paper. And I never forgot that because the woman who wrote it phoned me up, and she’s dying, convinced you read Karl Rayner, because he’s all the way through this. I said, Oh, my God, I studied under him. Is it not obvious? We don’t know. Karl Rayner was the [inaudible 27:11] philosopher behind [inaudible 27:13] two. There’s some of that in [unsure word 27:14]. So I think, I wouldn’t say there’s a favorite, I would say one thing I’m proudest of is if you go and search for [inaudible 27:22] on Google Scholar, 90% of the papers there will be people using the framework without a symbol. And that means it’s got utility. So the cases will be compatible.
Speaker: Miljan Bajic 27:36
Great. Another maybe just to add to this, you talked about like making this available to everyone. Last time we spoke and you said like in context like this, and this was like last year, making [unsure word 27:49] and all the tools available to others. Could you maybe elaborate on that and how that helped maybe? Or has that changed anything since last year?
Speaker: Dave Snowden 28:02
Yeah, we went open source on the methods. So there’s a [unsure word 28:04] now. So everything there is open source progressively over the next nine months, we’re going to go complete open API on the software as well. So that we can create a development community. And the reason we’re doing that, to be honest, is simple knowledge of life cycles, when you’re creating something new, you hold it tight, otherwise, it gets corrupted. When the market starts to take off, you open it up fast, because you want lots of people adopting and copying, so we’re going down that route. The other big thing we’re doing in Agile, this concept of decomposition and recombination. And I’m working with comic agile, and also with about eight or nine other people, we’ve taken all of the different agile methods and breaking them down into the lowest component parts and producing a complete facilitation kit for that. Yeah. And we’re branding that with an independent brand, is not branded [unsure word 29:03]. So our methods of branding [unsure word 29:05] the core pack has got an independent run. And so that mean you can for example, take Scrum, I’ll give the example I keep giving, you could peel out sprint and replace it with three months time box. So and I’ll put the picture in the chat in a minute so you can see them. So the this is designed to be an alternative to things like safe, right? Which we need an alternative to the Borg because what it basically says is there are individual things in virtually all of the methods and all of the concepts and we just need to use them in different combinations. So multi method, multivendor not single framework is what we’re trying to try.
Speaker: Miljan Bajic 29:4 8
And I think that’s where we’re kind of headed and I spoke recently with the Jurgen Appello and he was kind of I don’t know if you’ve seen what he’s been doing with unfixed but essentially just saying like we need to kind of [inaudible 30:00].
Speaker: Dave Snowden 30:03
Mean the magpie?
Speaker: Miljan Bajic 30:06
I think I mean, like, I don’t know, it’s been interesting because I think he’s saying the same thing in the sense like that we need to stop getting away from frameworks. And look…
Speaker: Dave Snowden 30:14
He is trying to create his own framework. Look at the last picture he produced. What Jurgen does is he reads extensively. I mean, the problem with Jurgen is he’s got the intelligence to do it properly, but he chooses not to. Alright, so he grabs things from lots of people, throws it together, put some pretty pictures around it and sees if this one will sell. So he put up his alternative to safe the other day. And to be quite honest, it’s comical. All right. I mean, you talked about some of, that’s the trivial end of Agile.
Speaker: Miljan Bajic 30:48
All right, let’s move on. No, I agree. I think we need to keep pushing and challenging the current status quo with the frameworks and everything. But I’ll take that over the alternatives. This next question here is from Doug. Doug’s here, do well connected themes perceived complexity different than an individual?
Speaker: Dave Snowden 31:15
Yeah, and I think the essence is, we actually evolved as collective creatures. So we evolved for extended families and tribes and part of the property [inaudible 31:24] showing up in Europe and North America, which manifests really scary North American Politics, with libertarianism is the entire focus on the individual. When the key things in complexity is defined by our interactions, not by anything innate to ourselves. If you want to change, people stop talking about mindset, which is bad science anyway, and change people’s interactions. It’s cheaper, it’s more ethical, and it produces bigger change. So seeing things in terms of high levels of connectivity, interaction, the ability to change those interactions is a much better approach to change. And it’s better based in science anyway.
Speaker: Miljan Bajic 32:07
So you’re saying changing the environment?
Speaker: Dave Snowden 32:10
Change the environment, change the connectivity. What matters is, think about it, I get the children [inaudible 32:17]. What do you most worried about when you kids hit puberty? Who their friends are? Because now their interactions are changing from you to third parties and who they are will change them for life. All right? So Interactions matter more, there are no innate qualities in human beings. It’s why things like Myers Briggs are complete pseudoscience. Yeah, we’re highly adaptive, we can change very quickly and we change based on our interactions and our social interactions.
Speaker: Miljan Bajic 32:48
It makes sense when we first moved to United States and I was a teenager. That’s what my parents were mostly worried about who I was hanging out with, not how late I was staying and all of that. Let’s see here. Did your consultancy for military, UK Government add environmental positive impact like social responsible value?
Speaker: Dave Snowden 33:20
That’s a very broad question. Right. So I’ll say a couple of things. First of all, my experience of working with military and I’ve taught just war theory at West Point, I work at Quantico and on the big sea to command thing. And I’ve got no military experience. But I’m now considered an expert on military decision making which is quite scary if you think about it, is military people are more ethical than non-military people. I still remember teaching this at West Point. At West Point, they’re very bright kids, right? They genuinely worry about what’s killing people because they know they’re going to have to do it. Alright, so they’ve evolved various mechanisms on that. And I argued a long time ago, military train people on ethics, software engineers are never trained on ethics. But the implications of software engineering to society are really scary. And it’s like the role of AI. And the trouble is too much of AI.
So I use my favorite phrase is written by misogynist males on the West Coast of the USA, who take and ran seriously after puberty, which is grounds to be committed to a mental hospital. The cultural bias behind a lot of software development is really very scary. Look what happened with Google? I mean, look at [unsure word 34:40] paper, which is a brilliant paper and the woman who published it gets fired. Right, because she pointed out the degree to which the training dataset was being ignored by Google. Yeah, and I’ve been in and out all my life, right? I still remember with Poindexter in Washington suddenly said what do you think about AI and both of us said this is 30 years ago, both of us said simultaneously, they’re ignoring the training data. Yeah, I worked on submarine recognition systems, right? We knew that the training data from experience commanders was far more important than actually raw data. You needed that human element in the data as well.
Speaker: Miljan Bajic 35:27
Great. Let’s see a couple of questions here. Do you know appropriate tools to manage emergence in software/products?
Speaker: Dave Snowden 35:37
We developed one that’s called sensemaking. So I’ve declared an interest there that came out of all our work with Darfur and counterterrorism and weak signal detection. So I mean, the key thing about sense maker is, it allows mass generation of data but critically whoever provides the data as a human metadata layer to it, so it’s human interpreted data we use not just raw data. And there were two big programs in [unsure word 36:03]. So one was total information awareness, which is like modern big data. Now, we’ve got John into a lot of trouble with the congress. The other one, which I lead with Sri from Miller Park, was about human sensor networks and human metadata. And that was focused on creating better training datasets. So for example, when we do a massive engagement of the workforce, it’s not done with a sort of social media type contribute your ideas. It’s done in ways that nobody knows what the right answer is and nobody can gain the result. So you shift up a level of abstraction that gives you objective data. And then you can see what are the stable patterns and what’s the non-stable patterns. And of course, the stable patterns you want, you encourage, you give more energy to unstable patterns that you think are desirable, you try and consolidate and give them direction. So there’s a whole process around that.
Speaker: Miljan Bajic 36:58
Right, thank you. See here, Alex, is your theory embedded some evolution psychological theories like spiral dynamics? You already kind of answered that, I don’t think but maybe I don’t know, if you want to answer.
Speaker: Dave Snowden 37:15
Aerodynamics is a pseudo-science. It’s got no basis whatsoever in any research. The original material was done by, the original guy originated it worked on a limited sample of his own students over a limited number of experiments and their extract from that to humanity. Yeah, you’ll get it. And don’t get me into Jade organizations, because that’s a religious tract. And the guy doesn’t even, he only selects the aspects of his cases which support the thesis he wants to support. You completely ignores all the people his app was fired in order to make the buddy system work. And he completely ignores the fact that every single case was that a leader imposing the solution on people. Yeah, and the stuff you can learn from that but you’d be much better off going to mon dragon in Catalonia and looking to how cooperatives have evolved. Yeah, because there are better structures in that. Yeah, I mean, as I say, the whole Thiel concept is a religious movement. It’s one of the three worst Bach books ever written in agile, because he actually selects aspects of cases which support his thesis. And that’s not how you do research. I mean, it’s very nice. And it’d be…
Speaker: Miljan Bajic 38:37
Reinventing organizations and a little right. So what the current state of constraint mapping? What’s the current state of…?
Speaker: Dave Snowden 38:49
I’m trying to find the picture so I can send you. We’ve now got the basic symbols work, so we’re working on. There’s a whole process on this if you’re on the [unsure word 39:03] slack group, which is not managed by the community, not by us, there’s a video you can watch on that which will update you. So what we’re doing is basically saying we map the constraints and we use images and metaphors to do that. And we use the whole of the workforce to do it. And then we divide the constraint clusters into counterfactuals. So things which can’t be changed and constructors things we should be using replicable change. So remember, I talked about S3 map, that’s where that’s coming from. So we’re making that into a series of processes using distributed intelligence to do the mapping rather than workshops, so we get more objective results.
Speaker: Miljan Bajic 39:46
Okay, great. I’m just looking. There’s not much left here. There’s this question here. Dave mentioned in one of his podcasts that leaders coordinate decisions and rarely make decisions. You talked about this at the beginning. This is Ken, at the operational level, yes but surely the leader has to make the initial decision on strategic matters or make the key decision when presented with all the information by [inaudible 40:12]. In other words, the leader has to prove the clear vision and direction on the path forward, especially on division two be achieved. Any thoughts on that?
Speaker: Dave Snowden 40:23
You ever worked in corporate strategy? It is heavyweight politics. I mean, this idea about rational assessment, the data and clear vision, that’s not how it works, all right. It’s pretty bloody Savage, right? You occasionally get really gifted leaders generally coming out of a crisis. So your guest did very well on this, right. I worked with him. And he made two or three big decisions. But remember I said in a crisis, you make decisions to hold your options. So the two big decisions he made, is he invested in a next generation of mainframes. And that’s kept IBM going ever since. And that was a good call because everybody else was withdrawing. The second thing he did is he bought companies in each of IBM major areas and sat back and watch what happened. Okay, that’s actually how you manage strategy, you hold options open.
So yeah, he’s making decisions, but remember his decisions to hold options open. So the company I was working strategy from data sciences, we were bought, because we understood services, IBM never understood services. And our management became the management of that group. It became IBM Global Services, which was for a while the biggest group, right? But he had the sense to realize he needed to import management from people who understood the fields, right? The same buying Lotus for software, and other areas. So if you look at really good decision makers, they generally interact, connect suggests, they do multiple options. Yeah. The Vision stuff, yeah, sometimes they do, sometimes they don’t. But show me a vision statement, which isn’t a set of platitudes. It’s kind of like something you do, you go and you say, we are going to be the x, y, z part of the company, or we are going to respect our customer. There’s only so many ways you can say this stuff. And yeah, go through it, do it, it’s fine, right? But it doesn’t really make that much difference. And the reality is, you’re generally responding senior leaders are responding to what middle management are prepared to do. Again, you learn this pretty fast, right? Is you can have all the leadership direction you want, but you’re not going to fire all your senior and middle managers, because they’re doing the execution.
So you’re operating with the constraints or what they’re prepared to accept or do or whether you can replace or you promote. [unsure name 42:51] had a brilliant mechanism by the way, he had the top 300. And every six months 20 joined and 20 left. So it’s a pretty savage environment. I used to do the training of them. Yeah. And the thing he always said to them is, so far you succeeded by achieving your numbers. Now, you’re going to have to achieve your numbers and work with other people when he said very few of who you will make it. And he was right. Yeah, but one in 10 succeeded in making the transition. Right. And again, what he was doing, coming back to my energy gradient. He was managing interactions and managing context to allow things to emerge, which he can then reinforce.
Speaker: Miljan Bajic 43:38
Right. Let’s give me a couple of thoughts or things to think about. I don’t know if there are any questions, people, while people think or is there anything else that you would like to ask Dave? And I have a couple more minutes.
Speaker: Dave Snowden 43:55
I’m open. I came home with whatever you want to talk about. Yeah.
Speaker: Miljan Bajic 44:02
So Maru.
Speaker: Maru 44:05
Yeah, I have many, a whole stack of questions, a whole graph of questions that I bring in my tongue, but I’m going to select one to make it as concise and self-contained as possible. What do you think people should approach constructing useful frameworks to understand or make sense of where the company is? And what strategies they should build? The goal here is to do competitive strategy. But what do you think they should approach? How do you think they should approach building frameworks so that they can make sense of where they are and where they should go?
Speaker: Dave Snowden 44:42
I think good frameworks emerge. If you look at [unsure word 44:47], for example, it’s 22 years old now. And it’s gone through multiple migrations and changes in that time, frameworks don’t work if they’re based on a limited number of cases. Where the big problem with safe. I mean, safe is based on Dean’s memory of three or four projects. That’s the basis of it. That’s where it comes from. Yeah, in terms of its background. So frameworks need to evolve, they need to be a mixture of theory and practice, and they can’t just be put in place. Right. And framework should also allow for diversity. The problem with Agile is the frameworks are confused with the methods. Alright, so Scrum is not a framework, Scrum is a really good collection of methods. If you think of it as a framework, you end up in the method, the wars. Yeah, I mean, I still can’t see that much difference between Scrum and Kanban. But I mean, I get into trouble every time I say that. And I just quote talking back you know, that was to sheep [inaudible 45:47] shepherds. But the devil you should be working out, it’s not a big difference, right? So I think frameworks need to be theory based, they need to evolve, they need to create structure, but they shouldn’t bind you into a single proprietary approach. And I’ll give you another illustration on this. There were three things which came to form the Agile Manifesto.
So there was XP, Scrum and DSDM. Now, there are two interesting lessons of this, DSDM if you don’t know, I was one of the three founders of that, along with my equivalent of logical and [unsure word 46:24] Holt from Cambridge, right? And we met in a pub in Cheltenham. And that’s how it started. We didn’t need to ski resort for a week, dinner in the pub was enough. All right. So that came in, and that introduced [inaudible 46:37] and all sorts of good stuff. You then had XP, which is, to my mind, really the heart of Agile. But nothing could scale around XP because it was experienced based and quite esoteric, whereas scrum was codified and abstracted to the point where it scaled very quickly. And that was where things went wrong. Because not that scrum wasn’t any valuable but it created this proprietary scaling, with training, with certification thing that everybody else then followed. And nobody went back and thought, is that the right thing?
So what we’re now talking about in terms of rewilding agile, is I say, is decomposing methods into their lowest coherent components. And there’s like three things in safe, which they haven’t borrowed or stolen from other people. So we can put those in that category. And instead of taking that massive diagram, you basically take the bits which work for you and put them together in different sequences. I put some of the cards in the things, you can see what we’re doing there. And actually, that’s what people really do with the frameworks anyway, they never implement the whole framework. They can’t work out what will work and they adjust it, and they just go under the radar. And that’s a hugely inefficient approach. So frameworks need to be generic, and they need to be at least, they really should be method agnostic. Yeah.
Speaker: Miljan Bajic 47:58
So what you’re saying, and I think we’ve talked about this before, but it’s like the future and really what works is where organization constructs its own framework and evolves it, rather than saying we’re doing safe or we’re doing this, right?
Speaker: Dave Snowden 48:11
I think you’re talking assembly not a framework. So okay. And somebody just asked about Ivan’s work, we’ve reached a provisional agreement with [inaudible 48:21] jackers, and that will move, his stuff will stay as it is, but aspects of it will move across into the [inaudible 48:27]. So that’s actually underway at the moment. And again, that’s a sort of similar approach. It is trying to break things down into essences. So we don’t think there are such things as essences, but the work then we can work with, all right. So we need much more of this. Well, I call it coherent heterogeneity. What the big frameworks do is they homogenize and so you’ve got to choose it. What coherent heterogeneity does is it says things can be different provided they’re coherent. So I’ve just shown you our attempt to do that. You can make things coherent by allowing them to combine in different ways.
Speaker: Miljan Bajic 49:14
But that also alludes to organization developing competence in understanding this and involving more and developing internal employees rather than relying consultants to come in tell…
Speaker: Dave Snowden 49:27
It does, and part of what we’re doing as well is to build a lot of knowledge into the artifacts. For example, if you look at those axes, they get down to a method that you can understand those QR codes you can scan which will take you to why the source and such like. So you can build a lot into artifacts without the need for people necessary to have the same level of training. We can’t prove this in IBM. Is this an approach we adopted. You can’t expect anybody to be expert in all of the things available. So you need to allow them to become experts when they need. But make it easy for them to choose what they’re going to do in different combinations. For example, one of the packs we’ve put into here, which is a sense maker pack, we haven’t put in sense maker, we put in applications of sense maker. So do I want to do a cultural scam? Do I want to do distributed ideation? So we put those in as things that people can understand on the surface and then they can dive deeper into how they do it later.
Speaker: Miljan Bajic 50:32
Thank you. We might have time for one more question. Anybody? What about I’m interested just maybe. Last time I spoke, I was surprised that you said our response generally speaking to COVID was pretty good. Do you still think about that it was very good. And any thoughts now? I think it’s been a year.
Speaker: Dave Snowden 51:01
I think it’s amazing. We didn’t have more riots. I mean, the riots are now coming. Yeah, but they’re far less than we thought there would be. Well, I think is really disturbing is the way that… Sorry, I’m having this debate with Jim [inaudible][51:18] at the moment I’m trying to decide whether Jim Ross is just basically has bad ideas or he’s a bad actor. I haven’t decided which, right. What you’ve now got is you’ve got far right money the minute there’s something which is because there is this libertarian beliefs, our work in this field, right? There’s a libertarian belief that society has to be destroyed for something new to emerge. underpins the game a game B stuff and why my name has been linked with Gabe B. I don’t, I don’t know. I never agreed to that. Alright.
So you get these sort of things. So what happens? The minute is something like the auto thing, money flows in very quickly. Which is actually why the Canadians were right to close off the money access, because what you’re seeing is not the people have been deliberately disrupted from the far right. But they’re using money and resources and social media amplification, to actually take legitimate process and delegitimize it by expansion. And that’s where it started to go wrong there. But overall, people accepted lockdowns, they accepted restrictions. I mean, we’re about to open up completely in England. What’s interesting is the opinion polls say we shouldn’t do that. People are nervous about it. But we have a Prime Minister who needs to distract from his hypocrisy. Alright, so that’s going on that side. So now the danger sign is now do we come out of this a better species? Or do we just go back to the old way of thinking and that’s where I think we’re going wrong. I think we manage the crisis really well. And human beings in a real crisis are always good. You’ve got this myth in England, alright of a certain generation they say during the war, right? It’s always during the war. Yeah. Everybody worked together. So why can’t we reenergize that and say, Well, yeah, but okay, we’ll have to get the Germans to invade again, to achieve it, right. Again, people aren’t thinking about the context. So the context of a crisis changes behavior, what good leadership should have done and it didn’t, alright, was to find a way to use that behavior to navigate a different pathway out.
