Simon Wardley – Wardley maps
Join Murray Robinson and Shane Gibson in a conversation with Simon Wardley about Wardley maps. Wardely maps are a discussion tool that surfaces assumptions about competitive landscapes. Starting with the customer, we draw the value chain of components that we need to meet their goal and then identify what stage each component is in the evolutionary lifecycle. Is a component a custom build, a product or service or a commodity or utility. When there is a competitive market everything moves to become a high volume, low cost standardised commodity over time. Understanding this allows you to get ahead of the wave of market change and it allows you to identify enormous cost savings by using utility services like AWS instead of building your own data centre.
Read along you will
Shane: Welcome to the no nonsense agile podcast. I’m Shane Gibson.
Murray: And I’m Murray Robinson.
Simon: And I’m Simon Wardley,
Murray: Hi Simon. Thanks for coming on.
Simon: Pleasure. Absolute delight. And thank you for inviting me.
Murray: we wanted to talk about Wardly maps with you today.
Could you introduce yourself and tell the audience a bit about who you are.
Simon: So I’m Simon Wardley and I’m the inventor of Wardley maps. I run companies. Work for governments. I advise large businesses. I do lots of interesting stuff in the open source world. I speak I speak at a lot of conferences and I teach people how to.
So the whole story behind this was actually I was running a company this was back in 2004 2005. The company’s doing very well. Revenue growing, were being profitable and I’m the CEO, but to be blunt, I was clueless. I didn’t have any idea what I was doing. I was making stuff up as it went long. I was terribly worried that people would rumble that I didn’t know what I was doing. And so I started reading everything I could find as a strategy. It was getting nowhere ended up in a bookshop. And the book seller said to me, had I ever read sun SU art war? I hadn’t I bought two separate copies. They’re all translations. And it was in the reading of the second one. I noticed a particular pattern where sun Sue talked about five factors, one of which was landscape. And this led me into the idea of how do I understand my Landscape. This led me onto a journey of discovering maps. I was using these maps to explain a problem. And somebody said what they called. And I said they’re called maps. They’re maps. And they were like what sort of maps are they? And I went, oh, I don’t know value chain maps. And somebody said, oh no, there’s something called those. All right. Strategy maps and know there’s somebody else who called some, I don’t know, help. And so I just said, worldly maps, there you are. And then it stuck. And I assumed there may be me maybe a few other people, maybe my mother or somebody would use them. It would be very small number of people. My mother’s a economist by the way. But it turned out that. Other people hadn’t learned how to map competitive landscapes and so they’ve just grown from there. .
It’s really difficult to explain this stuff without actually looking at a map. If we can get a map, I can take you through it.
Murray: Good idea. So how about if you talk us through the high speed rail map. So for the audience, this is at a website called Wardley pdf.org high speed rail map.
Simon: Okay. So I’m looking at this map here, and it’s a sort of monstrosity of different blues and other bits and pieces. So it says top right hand side pioneers, settlers, town planners. There’s lots of dots on it and everything else. So what we’re gonna do is strip it away first of all, and then build it up.
Murray: What are the axes?
Simon: Before we get to the, a axis I need to explain a difference between a graph and a map. So almost everything we have in business which cause itself a map is in fact a graph. And the way you tell this is if you take something like a mind map, if you just move any component up or down a little bit and ask yourself the question, does it change the map? The answer is no, because the map is defined by the nodes and the connections between it. Now, if I take a map of the world and I shift Australia and put it next to the UK, does that change that map? And the answer is yes. So the first thing you have to learn with a map is that in a map space has. Okay, so you can’t just simply move things slightly without changing the meaning of what it is.
So almost everything written in business we have, which is called a map is in fact a graph. We have very few maps now in order to create a map. And in order to create space to have meaning you need at least three components, you need an anchor such as magnetic north. You need a position of pieces. This is north, Southeast or west of that, and you need consistency of movement. So if I’m going north, I’m going north, I’m going south. I’m going south. So I started this off in 2005, trying to work out how to do this. And the first thing is the anchor. So the anchor that I use on a map will be the users and those users could be the public. It could be the business, it could be the government. You could have multiple anchors on a map at any time. So if I look at this map here, they’ve got three different types of customer. They haven’t specified what those customers are. One requiring 3d visualization, one requiring graphical information system, one requiring access to the website.
Okay. So they’ve got three I’m guessing this is engineers, architects. They could probably do with defining those customers a little bit more. Okay. Once you’ve got the anchors at the top, but now what you need is position and you get this through a chain of needs. The customer needs a graphical information system that graphical information system needs a project information management system, PIMS that needs a land registry, which needs an enterprise resource planning management system, which requires a risk analysis.
So what they’re doing is they’re basically going through and saying, this needs this. So the way to think of it is like a cup of tea. What does that need? When it needs a cup, it needs tea, it needs hot water hot water needs, cold water, and a kettle needs power. So you can create a chain of needs.
So that’s what they’re doing here at the bottom of the chain of needs. They’ve got things like data center, compute and power. Okay. So that gives you anchor and position, but now you need movement and it turns out that all of these nodes are forms of capital. These forms of capital evolved through four common stages.
Now, technically we should call it stage one, two, three, and four. But that’s not very meaningful. And those stages have labels. So if we’re talking about activities, we talk about the Genesis of novel and new custom built examples products and rental services, commodity and utility services. If we’re talking about practices, it’s novel, emerging, good and best.
But as shorthand people just put Genesis custom product commodity. So now what we’ve got is anchor and position and movement. I can look at this and go, the customer needs 3d visualization needs a graph kill information system needs. Land registry needs E R P M needs. Risk management system needs a platform needs, compute needs power, and we’ve got power as a commodity computer, as a commodity platform, getting towards commodity risk management more in the product stage land registry is custom built, etc.
Murray: How do you decide where to put things in the vertical dimens?
Simon: The first thing to note with the map is there is no Y axis. The only reason that Y axis visible value chain exists is because in the early days of mapping, cuz it doesn’t need a Y axis at all. People would go what I can’t cope, what’s the Y axis. And it’s just like, all right, we’ll give you some scaffolding.
We’ll put value chain. And the further you go up, the value chain, things are more visible and the lower down they’re more invisible. But as soon as you get good at mapping, you realize that Y axis is complete gibberish thread away. What you actually need is the evolution at the bottom and your just describing the chain of needs.
Shane: So the evolution at the bottom. So it’s a form of maturity model. Is that what I heard?
Simon: So it basically describes how things evolve. So if you think of a thing like compute. The first compute was 1943, the Z three. Then you get custom built systems like lines, electronic office. Then you get the first products like the IBM six 50 and a constant evolution of products until you get things like Amazon EC two, 2006 computes becoming utility.
So it’s this common phase, the Genesis of the novel and new custom-built examples of things and products and rental services and commodity and utility services. So when I’m looking at this map, I, immediately I’m going right. We’ve got land registry business information management system.
I’m immediately going well, land registry. Why, are we custom building that? Surely other nations have that surely this is a problem that has been solved many times before. Is there something unusual about our land or whatever it happens to be? And there may, but the point about the map. The person writing the map is exposing their assumptions in a form that you can now challenge by not challenging the person by challenging the map. And that’s critical. So most organizations run on what we call store is narratives. And then we have an entire industry which runs around telling people that great leaders are great storytellers.
So when you challenge somebody’s story you’re actually saying you’re not a great. And that’s why it gets a lot of conflict. If I can get them to put a story on a map, I can challenge the map without challenging them. And a typical example of this, I had an insurance company in 2011, they had a wonderful process flow about introducing robotics to improve the speed at which they installed computes in their data centers, that a bottleneck, and they were gonna spend several million in capital expenditure to get rid of this bottleneck.
It’s fantastic. And of course they asked me to have a look and I can’t say why you’re using robotics. They’ve spent six months, they’ve done all the return investment calculations and all this sort of stuff. So I said, could you quickly map it? And it took about 15 minutes. Most of that time was I don’t see why we’re mapping. And then they just got on and mapped it. And the interesting thing is when you follow the map down, they put computing product. Okay. I think it’s more of commodity. This was back in 2011. We need to order server and it needs to come in goods in that’s more of a commodity. And then they went, we need to modify the servers and Mount rack ’em and they had racking custom bill. And I simply asked, where do you put rack in custom bill? And it’s because they had custom built racks. And one of the modifications you’re doing to service well, the service we buy don’t fit out racks. That’s why we have to take cases off them. Draw new holes, add new plates, to get them to fit our racks. And that’s why we need robotics. And somebody in the room just went, why are we using standard racks? And these people aren’t daft, there was at some point in the past, it made sense to have custom-built racks because there was no such thing as that of racks, but they had been tracked by that narrative and that story ever since.
And so what their focus on is improving process flow and improving process flow makes perfect sense to go and spend millions on robotics to get rid of that bottleneck. But when you map it and you look at evolutionary flow, you go, what the hell are we doing? Why are we using customable racks? And simply by using standard racks that saved the millions, except for we look to the map and just somebody went surely computes a utility now with cloud. Yeah, of course it is. So we don’t even need racks. We don’t even need the data centers. So it’s saved them a bucket load of money. And that was a simple 15 minute odd discussion.
My favorite example is one in government Liam Maxwell saved about 450 million through mapping. It’s gonna save about 1.5 billion. Most organizations always shocking how few understand who the users are, the components involved in building something how evolved those components are. And a lot of when we talk about legacy, what we’re actually talking about is custom-built examples of things which became commodity long ago, and we should have removed them. But we haven’t.
Now to your question of agile and back to this map. So in 2006 so I adopted my company, we were using extreme program, which was written by Ken Beck as a friend of mine back in the 19 very late 1990s. Just before the agile manifesto came out. I think I can never remember 2000, 2001. And we quickly learned that agile doesn’t work everywhere. If I look at this map here what you can see it’s circled various of the components and the ones on the left hand side built in house with agile techniques. Extreme programming is very good. And the reason for this is the stuff on the left hand side Genesis custom built is what we call the uncharted space. Things in that space are gonna change. And we don’t really know what we want now in the middle, where we get more to the product space. It says use off the shelf products and lean where it’s, because we’ve got a better idea of what we want.
And now we’re about learning and producing waste. Now what it gets to the far right hand side, and remember everything evolves across this. On the far right hand side then you’re outsourcing to utility providers, or if you’re building it yourself using something like six Sigma, because what you are about is reducing deviation because you know exactly what you want. You just want volume operations off. Exactly. I think good enough with as minimal deviation. So you’ve got these two extremes which is the the desire for deviation you need on the left. So extreme programming, cuz it reduces the cost of change is good. The stuff on the right you know exactly what you want, but you want it volume of operations.
You want no deviation and the stuff at the middle, you are all about learning and reducing waste. So you use multiple methods and that’s important to remember cuz there’s a world of difference between using agile. So using agile would be like using extreme programming, which you would do here on those components on the left hand side and being agile means using appropriate methods. So on the right hand side, it means outsourcing using, six Sigma in the middle, it means using lean and on the left, it means using extreme programming.
Murray: So in this example, we are looking at, I’ve got some platforms like an enterprise resource management platform, a CR gIS. And so on. I’ve had quite a lot of experience with companies implementing these. And there’s always a considerable amount of customization of these platforms. Generally about 30% of the features have to be built. New about 30% is customized and the rest is configured. For example, companies putting in SAP, an SAP module to do their finance are gonna be spending 10 million on labor and licenses
Simon: And the rest. I love this. I love this. I sat in a room with about 140 CIOs, different companies. And I, I said, okay. Let’s pick something. Enterprise resource planning. Who’s got a big ER P system and every single hand went. Okay it must be a commodity then if you’ve all got it. And there was lots of no, it’s not a commodity. We have to do lots of customization, et cetera, fit our business and blah. Okay. So I asked one of the CIOs in the room, one of the customizations you’re doing and they said, all right for example, we have this. And so I said who else in the room has got this 440, your hands up, went up. And then we just started going through the list. And what it transpired is everybody in the room was customizing the same system in almost identical ways, creating no differential advantage for any of them. Sure. Creating huge amounts of money for the vendors and the consultants and all the rest of it. What should be a commodity. They’ve got self-interest parties, who’ve got an interest in not making a commodity. And of course, they’ve all everybody’s been telling themself. It’s a secret virtue, which is differential. The customizations are exactly the same. And of course you go why don’t they do that? Put that into product. They’ve got, self-interest not to, they’ve got an entire ecosystem to support. So it was just like, why didn’t you people talk to each other?
So literally we were just going down through this list and everybody had exactly the same thing. And I think I remember we, we toted up some figures there and it was ridiculous. You are talking many billions being spent on customizing the same system in almost identical ways. And creating no competitive advantage for anyone. And could these things be turned into a commodity? Yeah, of course they could.
Murray: That’s what we’re seeing with the software as a service movement.
Murray: That these things are all going online and you can use them in the standard way, but I think people are still configuring the software as a services quite a lot as well.
Shane: Yep. You pick up the big, the sales force. You’d think that would be commodity, no customization, but there are some very large consulting companies making some very large amount of money, customizing that out of the box solution.
Simon: I absolutely agree. What is Salesforce, they are more of a rental model. They’re not really a utility model, it’s a rental model. And of course they’re doing the same old product practices of encouraging people, you can customize and all the rest of it. God most of this stuff should be highly industrialized. It should be, literally come out of the 10, all done pre rebates, everything else. You can bet your bottom dollar that people are customizing it in identical ways and being ed to do so.
I have one organization, gosh, they were running around government departments. So this is lifeboats rescue police or ambulances. All have an issue of, you’ve got resources, you’ve got an incident, you’ve gotta get those resources to the incident. And so it can requires you to map it out or the geographical maps in this sense and where things are and all the rest of it. But it’s the same problem. But you get this one company going through, 40 odd, lifeboat agencies, selling identically, the same system, customized for every single one at huge cost. And it’s the same customizations. It’s shocking. So this is also why some of the things that Amazon does cuz Amazon just brutally focuses on industrialization. It’s just we’re gonna take this stuff, which is a product and turn it into a utility and we need more of it.
Murray: In my state in Victoria, the government spent 330 million customizing a train and bus ticketing system. And they chose a system because it had already been done in Hong Kong. And then. Spent 330 million customizing it.
Simon: I’ve gotta say the worst duplication I’ve seen ever in government is actually to do with prisoner registration, where we managed to build prisoner registration systems about 118 different ways. How many times do you have to rebuild that this is nothing compared to the level of waste I’ve seen in the private sector? One very big pharma company had 350 teams building enterprise content management systems and five global efforts to build the global enterprise content management system. None of which were talking to each other.
Now you think, that you can’t get worse than that. Oh, you can just go into the banking world. You just look at risk management system. We stopped counting in one particular place. Once we got over a thousand risk management systems and the forever complaining, they can’t innovate and you go if you’ve gotta keep on rebuilding the wheel in slightly different shades of blue.
Yeah, I’m not surprised. So obviously there’s a lot of vendor interest in this cuz it’s all money to people. If you think about the Hong Kong example there, you have to ask, why is the government talking directly to the Hong Kong government, if it’s a nationally owned system and that’s why it’s encouraging to see lots of open efforts and lots of collaboration between governments.
And we started to see this with UK gov. One of the things I helped write was something called the better for less paper back in 2010, for Francis Moore. And this led to the introduction of things like spin control, supported something called government digital services, which was created subsequently.
And so there was a big sort of focus on use of open and collaborating with others. A to get more information, to share that information, stop rebuilding stuff which is already, been done. I think that, you just have to grab a couple of thousand developers, put them in a room and ask, who’s built a user lock in system and everybody’s hand goes up.
You say right. How many people have built it more than once and most? Yes, they so, in a single room, you’ll get the same system having been rebuilt thousands and thousands of times. There’s so much waste there.
Shane: Yeah, and I agree. I’ve got a startup and we had to do the user login system. So we adopted Google’s authentication because
You, click on it. It comes up, it does the work for us. They’re much smarter. They’ve got much better machine learning models than we would ever build. For identifying where you are, , I’ve just come back from the UK. And every time I logged onto a new service in the UK, Google came up and said, Hey, you’re looking a bit weird. I need to check that it’s still you. And that was interesting. And the gov.uk is an interesting one for me. So in new. New Zealand. We adopted a government marketplace the boffins over here saw the UK one and then decided to build their own for 10 to a hundred millions of dollars, rather than just adopt the one that was built for the UK government that was open source and pretty much fit for purpose for a country of our size.
Quick question on the map though. So is that where you see the Woodly? The Woodly maps used a lot is to map out the estate of an organization around their systems. And then the goal being moving from left to right. and identifying the step change that we need to take to move each one of those things further.
Simon: So the first first point with the map is getting everybody to focus on number one, who the users are and what their needs are. The second part of the map is understanding the components I, the supply chain whether it’s a physical, whether it’s a additional supply chain, is it material, but it’s understanding what components go into making that.
So that’s understanding what we need. The next part is course understanding how evolve those components are because the methods and techniques we’re going to use, whether it’s project management, whether it’s pioneer settler tower planner.
That’s an organizational model that I put in place in 2006, because what we discovered was you’ve got some very good engineers in the uncharted space where, there’s a lot of change going on, some very good in the learning space and some who are very good at industrialization.
You need brilliant people in all three. And so you’ve got different attitudes. So what you learn by putting that evolution stuff in is that you need different methods and potentially different attitudes, and that leads to different ways of organizing as well. The other thing, then you start learning is you’ve got duplication. So the maps help you start going, why are we building this over here when we’ve got it over here? And so you start building profile diagrams, and once you start doing that, then you start discovering why most of our contracts are flawed. So if I take that HS two map that you had, if you imagine we said we’re gonna outsource the whole lot.
I can tell you that contract’s gonna fail because the stuff on the right, we can define the stuff on the left we got. So obviously we’re gonna get massive change control, cost overruns. Because the stuff on the left is going to change.
So you realize that most contract structures are flawed. And then of course, then you’ve got the issue that people improving the process flow by the, looking through the lines in the map. And of course things have evolved. And so you’ve got evolutionary flow. So other than the fact that most people don’t think about users don’t actually understand their needs, don’t actually understand the supply chain use the wrong methods, use the wrong contract structures, focus on process flow rather than evolutionary flow. Don’t actually organize themselves. And a massively duplicate everything’s hunky Dory in it.
Shane: So as we look to move the notes right over.
Shane: what I think I’m hearing is within the evolution model you’ve got in there. So within that idea of custom built versus product versus commodity,
there’s a concept of jumping the chasm at each time. Isn’t there. So if I was moving something in the commodity area slightly the techniques I’d use would be very different than if I’m trying to jump from a custom built evolution. across the chasm to product it. It’s a bigger jump it’s riskier. I have to use different techniques than moving within the evolution column.
Simon: So there’s a dangerous trap here. Diffusion is not the same as evolution. So let me explain why it’s a dangerous trap. If we start with something which is novel and new, the Genesis of something like compute and then custom-built examples, and then products and then rental services. What happens is you’ve got many instances of that evolving thing.
So many different constantly improving products. Each one of those runs on a diffusion curve. So the evolution of something usually encompasses hundreds, if not thousands of diffusion curves, and this creates a big problem. So something’s in product stage, late product stage, you get the vast majority of their market using that late product, but it’s evolved to more of a utility.
And so the utility version though, it’s the more evolved has early adopters. You can’t avoid evolution, we’re gonna go there, but if you do CIO surveys, and this is the sort of stuff I saw in 2008, 2009, 2010, and our CIOs, what the future was, they told you it’s all about virtualization in the data center.
It’s all about VCE. It’s all about building more data centers, getting more efficient equipment, and you can look at the map and say, no, it’s not, it’s all about using cloud, but they turn around, they can look at all the other people in, on their space, on their diffusion curve using products and say everybody’s using it.
Everybody’s building data centers. That stuff over there is just for startups, which is what they said. But you can see from the map, it’s the more evolve form it’s unavoidable. And there’s reason for this efficiency, speed and access to new sources of value. So be very careful with the chasms and the diffusion curves because they’re great, but just remember the evolution of something involves hundreds, if not thousands of diffusion curves each with their own chasm and it can cause problems.
mix those up.
Murray: Does everything inevitably evolve to a commodity?
Simon: If there is competition, so supply and demand competition, it will drive there unless there are specific constraints physical constraints or informational constraints, preventing it, getting there. If you wanna go the other way, just get rid of supply and demand competition, create a monopoly, and then you can manipulate the market in any way you wish to
Murray: Is there anything further to the right after commodity?
Simon: Now the maps themselves deal with the space of competition. So there’s because they’re driven by supply and demand competition. Now there’s three basic forms of competition. One is conflict. One is collaboration. Another is cooperation. So competition is the act of groups seeking something. And we can either do that by fighting each other conflict or working together cooperation or laboring together, collaboration. The maps deal with that space there a window into that particular space before the maps, you’ve got things which are not governed by competition. So you’ve got things with social value and after the map, you’ve got things with social value as well. So decommodification, so you can think of it like it becomes a commodity and eventually we go, ah, right government, we need to turn it into a public utility. And so it is no longer governed by the rules of competition is governed by a whole bunch of different rules, political rules.
We reference that on the map by highlighting something and just saying it’s become a public utility, but we can’t tell you much about that space cuz it’s outside the boundaries of competition. Does that make sense.
Murray: Yeah. Yeah, that makes sense. So when I’m looking at this example here of the high speed rail map,
Murray: It looks to me like what it’s doing is helping you to think about all the different components of your solution that you might put together. And it’s helping you to say which ones are really customized and which ones not. So instead of building my own data center, because Amazon web services already has, and it’s Microsoft ones and so on, we don’t need to build our own, but I do need to build my own land registry and my own business information management system, because that is very specific to me and my government department.
Simon: It allows you to express the assumptions that you’ve got around a space and which allows other people to look at it and challenge those assumptions. And that’s exactly what happened here. They actually ended up with a much more complicated map than this originally. This is for building HS two in a virtual world. Cause it turns out that it’s easier to dig up a virtual world and go, whoops, we’ve got that wrong . It is a lot cheaper to do that than the English countryside. They built this system. They start off with a systems graph for building HS two in a virtual world, turned it into a map and used the map to manage the project. And this project ended up in front of the public accounts committee, cuz it ended up way ahead of schedule and way under budget. Which was pretty unusual for a government project. And James Finley was the CIO who was in charge of this. And then they started using mapping in other parts of the it function there. Unfortunately they didn’t use mapping for the whole railway. I say unfortunate, cuz it’s a very expensive project, but they used it for building HS two in a virtual world. But fundamentally, yeah, you are exposing your assumptions. It allows other people to look at the assumptions. You’re making question them, ask about, why is land registry, custom bill? And in this case they use the map to think about some of this stuff should be more commodity. Maybe we should open source. Some of this stuff, maybe drive some of this stuff in more of a commodity. Maybe we should divide the contracts up in a different way. From what they originally had envisaged.
Murray: surely there must be a thousand land registries out there in the world.
Simon: You’d hope so so the point about this is the maps opened up those conversations a bit like the map for the insurance company, which had a rack and custom built, enabled somebody to go, why have you got custom built racks? These people weren’t DF. They were just trapped by narrative and track by stories. And of course, if you’re doing a process flow, you’ve got none of that information. You can just see, like thing goes into rack. We’ve got a bottleneck, let’s improve the bottleneck process efficiency. Once you map it, they just go, why are we doing this?
Murray: When you put up a map like this with a room of people they’re making their assumptions visible, as you said, but they’re still assumptions and those assumptions might be wrong and they wouldn’t know.
Simon: That’s the point about you put the map up, you allow other people to challenge. In fact, so couple of things you need to know about a map, all maps are in perfect representations. Secondly, they’re all built on models, so they’re all wrong. So they’re all imperfect and they’re all wrong. But they turn out to be useful because. You can put your assumptions down in a way that others can look at it and go look, I’m looking at your map. And I’m seeing, you’ve got compute being a commodity. You’ve got data center being more of a product. Exactly. Fine. So what are we gonna do? Are we gonna build a data center? No. Or are you gonna build computers? A you just gonna use Amazon, right? Fine. Okay. Land registry. You’ve put land registering custom built and say we’re gonna build it with our HR. Okay. That’s good. Don’t others have land registry. Surely land registry should be over at the commodity. Can we not go? Okay. Fine. What’s happened is we’ve got consultants running around selling custom built examples everywhere. No one’s built the, there is no, maybe you should go and talk work with a few other governments cuz land registry must be a common problem. And see if we can’t create, standardized version between multiple governments, it’s a bit like student registration. How many times do we need to build that? So I can start taking out the components and asking those questions.
Now it’s really difficult to do that in narrative form. If you imagine the story to try and describe this, it becomes really difficult to challenge the story. And of course, you’ll challenge what somebody says. And of course we have this industry sayings, great storytellers, great leaders. So it’s easier to do it with a map.
Murray: I think partially this is due to the process we use to come up with solutions. Maybe that’s the cause of it. I know working in software and product development for a long time, that people will always start by saying, who’s the customer, what’s the outcome, what are the requirements? And then they brainstorm all the requirements. And then they go to the architects and say, what’s the solution. And the architects come up with all the solution. Nobody has thought about cost during this process. There’s no design to cost happening.
Simon: When you look at a map, it is saying about what the users, the user needs and what the components are, et cetera. So it forces you to think along those lines. Now, when I used to run a software company, I used to get these companies coming up to me and say, we want to build this. How much is it gonna cost? And 2004, 2005 we’d actually converted our entire data center into a virtualized environment. We had something which was not dissimilar to EC two. In terms of you could spin up machines through an API, we’d actually built a platform as a service offering called Zike, which was serverless, but it was back in 2005, 2006. Because of our efficiency, I was using this to introduce something called worth based development. So companies would come to me and say, we wanna build stuff. And I said that’s great. We’ll build it for free and they’d go, what, but I want us to agree a metric of value, an outcome, not an output. An output is like how you get that the product. We’re not interested that we want an outcome a metric of value and we’ll take a percentage of that.
So the more value you get, the more value we get and we’ll build it, we’ll do the risk calculations and everything else. And it was interesting. That’s a couple of times we did this there were real issues. I can go through that. It was quite funny, but most of the time people say, oh, no, we just want a cost for the project.
And it was just like what are the metric of value? Oh, we don’t know. Okay. Where do you want a cost for the project? And I had to Badger and it got down to, they wanted a cost for the project. So they could put in an ROI calculation to work out how big the market needed to be to put the business case together to support what the executive wanted to do. So basically they were just collecting information just to support what the executive had already decided to do. And it was like, do you use marketing to challenge that decision? No, we’re going this way. Executive decided we need whatever it is. You see it today with people going, we need AI. We’re gonna install AI. AI is the future. Do you have a map of the space? Do you know where AI? No. No, but we’re gonna do it. It’s so when we start coming from, companies think about outcome, they think about needs. They think about components. They think about business value. No, they don’t.
Shane: So sounds a lot like people implementing agile, right? I wanna implement agile, not, I wanna change the way I work to achieve some other benefit. The benefit is supposedly agile and it’s that’s not quite right, because you need to get business change and business value. And agile may help you do that. But it’s an enabler. It’s not a thing.
Simon: So if I go back to that building HS two in a virtual world map, use extreme program, left hand side, use six on the right hand side. Now being agile means using multiple methods. Using agile is, using extreme programming where it’s appropriate and people get this wrong and they talk about, we wanna be agile. Okay, fine. But that means you’re gonna use methods, which aren’t agile because they’re more appropriate. You’ve got ANZ. Didn’t ANZ, go on some sort of big crusade. We’re gonna be agile everywhere. And as it ended up in a complete horror show
Shane: it’s funny. You should say that because there’s just been some articles on LinkedIn around some people’s view that the aNZ bank didn’t deliver their goals that they stated when they went down, their scaled agile
Simon: oh, did they use safe? did they
Murray: I know something of the insight story to this from gossip within the consulting industry. So, new CEO needs to do new things, to make your stamp, needs to cut costs and improve service and all that sort of thing. Talking to various people, a lot of people saying you should do agile cause that’s the popular thing. And they got Boston consulting group in and they said, we can cut your cost dramatically. By implementing agile, they set up a target of a reduction of 25% of their management on the basis that, their bureaucracy and agile is much more streamlined. And so, then they implemented the Spotify model and safe because the big consulting firms don’t really understand any of this. They don’t really have any practical experience. They just produce packs and the bank paid them vast amounts of money and they came up with new organization designs and then implemented it with a target of firing 25% of their manager.
Simon: Do you know who doesn’t use the Spotify model? Spotify,
Murray: We had Spotify on, they, they said it was an ideal, but anyway, it’s true that agile as an organizational system should be able to make you less bureaucratic, but doing it overnight, like that , is not the way to do it. So everybody said after it happened, it was just chaos. And there was lots of work for other consultants to come and try and fix it, which was good for them. AAN said, it’s getting a lot of criticism for not being more effective in the loan area. And the loan area was the main area they did not implement agile. So it’s actually just very political. People are trying to push the current CEO out and get a new one. So that’s really what’s going on.
Simon: If anybody wants to sync them, just encourage them to go bimodal. And have you heard of bimodal? I presume that’s a big horror story. So back in 2004, if I go back to that again, that ages two map is from 2011, 2012 is somebody else’s map. But back in 2003, four with my own organization I didn’t have mapping, but I realized there was something wrong with the way we structured. So I actually split the organization into two parts basically future focused and more core services. And that almost collapsed the company. I’d previously built sales structures. And then I made this division. It always collapsed. The company caused absolute warfare because what I failed to realize is there are the extremes but there’s the transition in between.
Which is why, when I have maps in 2005, we introduced 2006, the pioneer settle, the tail planner model, and that stuff has picked up. So what you have are people good at, creating the novel and new and failing a lot and everything else, then brilliant people who can take that and turn it into something useful. And then you have the people who can take that they still, from that industrialize it into common components, because otherwise you have these two extremes they’re too far apart. And you have these people building the novel new, and they go to the people on the other side and say you turn that into a utility. And they go, where’s the documentation. We haven’t built any in it’s big fire argument. Everybody ignores each other. And it causes all sorts of problems. So this was a known problem in 2000 four and of course we had the solution by 2006. So in 2011, I just couldn’t believe it. When Gartner came out saying bimodal.
Murray: That’s because the big consulting companies were selling it. There’ll be something better. Public private partnerships are still quite popular. That was a big money maker.
I just wanted to go back to what I said earlier, cuz I don’t feel like we really covered it, but when we are developing software or implementing SAP or whatever every organization they do have, we don’t wanna spend more than X and we wanna achieve certain things. But after they say that, then they forget all that and they immediately go into just go and ask everybody what they want, write it down and then ask the technical people to design a solution, to give everybody what they want without any idea of cost. And they come up with things that they would like to use.
May or may not be commodities, but because nobody has really looked at it from your point of view, like if you said, okay, there’s all these things, there’s all of these commodity services you could use. Then you could say let’s structure our requirements in our solution around that, because we know that it’s going to be much lower cost. There is a way of developing software, which just says we don’t even discuss cost until we are already to find everything we want and by that time it’s all highly specific to us.
Simon: The missing bit is value because it doesn’t matter how much it costs as long as it creates more value than the actual cost itself. The problem is we don’t often know what the cost is, and we certainly don’t understand the value. So we don’t have the chain. We don’t go from the point of view of here’s the user here’s the user need. And because when you think about that chain, what you’ve actually got is each of the lines are actually exchanges of capital. So if you think about a cup of tea needs a tea cup water, you’ve got exchanges of capital. As in the cup of tea’s got value because we’ve combined tea with hot water, with a cup and the hot water comes from cold water and Kele needs power.
So you can actually trace. Flows of capital within a system. And so what you need are the two things you need to be asking, what’s the cost of this change and what value is it actually going to give us if you did that? Most of those customizations would probably disappear. Now, there is a place where that’s going on. So 2014 the runtime lamp.net, all this sort of stuff shifted to more of a utility, which we called serverless, which is Lambda. One of the beauties about Lambda is we have that concept of billing per function. So for the first time, when you build an application, you can actually start to look at where money is going within the application.
Most people, their application runs in a cloud. They can get a generic value of the total cost, not cost per function, or, they’ve got it in their data center, in which case them got a clue, you just ask people what’s the cost of that particular function. And they have no, no, idea. They don’t have the metrics or anything. You ask them what’s the overall cost. They still don’t know. But you get this for free basically within the service world, you get billing per function.
So now we’re building systems where we can actually look at capital flow and there’s a whole practice developing around this, which is called fi ups. So if you start mapping out systems, we said about computing going from Z three. To product, to a utility, as things cross that boundary from product to utility, then you start to get a new set of practice.
So with compute, we went from high meantime to recovery to low. So we went from a world where it took weeks to get a machine to seconds. So now we could distribute systems design for failure, chaos engines, and we got DevOps. And the same is going on in the run type space where we’ve shifted from product to utility.
So it’s new set of characteristics. One of those is visibility over actual capital flow and we’re getting an entirely new set of practices. I use that map cause I used to run Straty for a company called canonical. They provided Buntu. And so I used that map to attack the cloud space. In 2008, we were two to 3% of the operating system market up against Microsoft and red hat. It cost half a million, took 18 months and we took 70% of all cloud computing. So if you were involved in computing in 2008 to 2010, you might have noticed it was basically all red hat, Microsoft, and then suddenly it was Buntu everywhere where you were mapped.
Strategy is iterative that worked in 2008 to 2010. Now you wouldn’t attack that space because now the run time has shift to a utility. The place to focus is on serverless and the fops, everything below that is now heading towards the new legacy, including DevOps and cloud. If you started a cloud project and DevOps project today took my friends at Netflix, what eight years to get rid of their data centers. By the time you completed at 2030, you’ve just built the new legacy, cuz everything’s shifted to serverless and fops. So one of the things you use maps for is also not any anticipation where to attack, but also thinking about how that change forces your strategy change as well.
Did I answer your question about the cost and the value?
Murray: A strong theme we’ve had from guests on the podcast is that we really need to focus on outcomes and the value of those outcomes and hardly anybody does. And the reason is because the managers making the decisions are too inward looking and too focused on, their own campaign against the guy in the office next door.
Simon: If you look at that map of hS two, building HS two in a virtual world, the thing about that map is there’s a lot of components in there and we can have a lot of good discussion about how we should build things, how things are changing and all the rest of it.
So there’s a lot of complicatedness and a lot of complexity exposed in a map, and that’s a problem. And the reason why it’s a problem is something called Ashby’s law of requisite variety. So in a controlling element of a system, you have to have as much complexity in the controlling element as there are the inputs and outputs of what’s being controlled.
And so when it comes to management, there are two ways of doing this. You can either make management capable of coping with the complexity of what is being managed, or you can pretend that what is being managed is simple. And we do the pretending one through what UN owner’s KPIs. So we introduce these wonderful things called KPIs and pretend that somehow they embed reality. In fact, what they’re doing is they’re making so. easier to manage, because now I have to just look at one figure it’s like the chat bit, when there was the Cheral disaster and there was the person inside saying, no, the power plant’s running, I’m looking at the dial and his friend is outside saying, no, I can see it’s blown up. So we tend to not look at reality. We tend to look at the dial and we manage by KPIs. I, think we’ve got too much focus on that and we train layers of managers to constantly search for the ultimate dial, the ultimate KPI
Murray: Everybody’s looking to optimize locally. If you look at it from a systems point of view,
Simon: Without understanding the space.
Murray: Without understanding the system that they’re in.
Shane: When we talk about the evolution scale, you mentioned that there were two, so there’s the one we’ve seen around custom build. And then there’s the novel emerging. Good and best. And so from what you said, I’m erring that the novel emerging, good investors for services, orientated maps, and the other ones for technology ones. Do I get that right?
Simon: Not quite. And that’s not your fault. That’s my fault. The evolution access really should say stage 1, 2, 3, and four. That’s what it should stay. Because what you’re talking about is the evolution of capital and there’s something called the cheat sheet, which describe all the characteristics of stage 1, 2, 3, and four. But that all came from text analysis and it turns out there are labels. Which are good for describing different forms of capital in those stages. So if we talk about the capital activities, things we do the labels of Genesis custom built product rental, commodity utility fit. And so that’s what we use the labels at the bottom. If we’re talking about practices, the labels of novel emerging, good and best fit. If we talk about data then it’s like a model divergent, convergent modeled fit. If we’re talking about knowledge, it’s concept hypothesis theory, and accepted fit. If we’re talking about ethical values, then we get concept divergent, convergent, accepted, fit.
All of them are stage 1, 2, 3, and four. And the point about this on a single map, because I met things like cultural systems, nation, state systems, all this sort of stuff. You can map activities, practice data, knowledge, ethical values on all on a single map. But rather than putting the entire list of labels down the bottom, and not putting stage 1, 2, 3, and four, because that’s meaningless.
We start just getting Genesis custom product commodity and occasionally I will bring the other labels in. So I might go Genesis concepts and custom built and diverging product and converging and then maybe commodity stroke accepted. So I’ll bring multiple labels in just if it makes it easier to read a map,
When I started with this mapping. I just thought, this is what you learn how to do at mBAs. And I never done an MBA. I teach a multiple business schools now, which is just so for me, so weird. And I thought this was Simon’s cheap and cheerful way of mapping out the space.
Obviously other people have much better ways of doing it, and it took me at least five or six years of sitting in rooms and going this and people going what’s that it’s like being a carpenter, you’ve gone and made a chair and somebody’s looking to what’s that it’s a chisel, what’s a chisel. What do you mean? You don’t know what a chisel is? How do you make things for that chisel? And after about five or six years, I decided to really realize this must have been about 2011, 2012 people aren’t looking at the landscape. Again, it was just like this really bizarre moment.
And this stuff is exploded. It’s all creative commons, help yourself and everything else. There’s conferences. So there’s like map camp. There’s just when one map camp, Germany there’s map camp London coming up there’s conferences there’s people of written books Amazon web services, their second ever book is called reaching cloud velocity. It’s got 17 pages of mapping in there. They have one of my old models, the ILC model, which is basically how they rip up industry after industry. You’ve got new books coming out, Liberty mutual use of serverless. Dave Anderson’s got a book coming out, which is full of maps. And then you’ve got flow book by James ER, oh, that’s maps. So there must be a dozen Eric. Sean’s got lots of books with maps out. That must be dozen or more. There’s just slow. The UNS global data slashy book has maps all the way through it. And there’s a huge community they’ve built. Wikipedias, they’ve built awesome lists on worldly maps. They’ve built tools. It’s just amazing, absolutely amazing. The things they’ve done and what they share with others. It’s just spectacular, but we’re all still learning.
Shane: All right, summary. So I’ve heard of Wordly maps, but I had no idea what it was. So for me, what I got from it was it’s really a way of sharing a story, right? It’s a shared language. It’s a way of bringing something up and then having a discussion around it. It allows us to challenge the map, not the person. So it gives us permission to validate the assumptions without attacking the person that wrote those assumptions down. I thought that was valuable. And it also seemed a good way of identifying duplication, putting the nodes on there five different times in five different places is gonna start a conversation about why what’s the context for that. Is it just the way we used to work or is there a really, a good reason why we should be operating that way?
For me that bottom axis seems like a maturity mapping exercise. So as we move from left to right, depending on how far we are moving and how fast we wanna move, we have to pick up different ways of working to make that move. So you talk about, lean versus six Sigma versus some agile techniques.
So that was me Murray. What have you got.
Murray: Well, you should go and do this on your own startup, Shane.
Shane: So yeah, I’m definitely gonna go and apply this in a couple of ways and see what happens. So one around where we see ourselves in the market and we’re moving to but also the agile database of working, looking at the practices that we share with people, if you’re a data team and saying, okay, how would I worldly map it? To talk about that. So yeah, I have to do some more reading, some more practicing and some more giving it a go and see what happens.
Murray: It certainly explains why every new concept like data mesh becomes a product. And then after that becomes a service. There’s a strong commercial pressure to turn, a consulting idea, which is very custom into a product which can be sold at high volume for higher revenue and much higher profit margin, at least initially. And then as everybody gets into the space, so the profit margin drives down. There’s a strong incentive then to move into a software as a service model, which is more economical and brings the profit back. So it’s a good explanation of how things become modified over time. And I agree with you Shane, about how it’s a very good discussion tool and problem solving tool. This is a really good tool that I can add to discuss problems with people and to talk about what’s going on and where the assumptions are. I think it’s very effective way of thinking about when you’re buying solutions buy versus build type of discussions. I could see that being very useful.
I think what you’re saying about focusing on value and outcomes is really important. And people don’t do that enough. And then putting the cost against it as well, really important. Everything just gets turned into deliverables and contracts, and it’s a big weakness in procurement .
Simon: I have to defend procurement here because it’s not just project management, it’s also financing as well. So once you start mapping, you start to re discover that on the left hand side, you need to break things down to small components and use a much more VC approach. And then you’re using much more off the shelf and then much more utility based pricing. So you’re financing your procurement methods, your project management methods need to change as the thing evolves. And the problem, because we don’t show people how to look at that landscape and how to consider things evolving. If you don’t have that, all you have are graphs. It’s very difficult to, how are you supposed to work that out?
If I’m normally looking at some sort of a systems graph and I look at a component it’s just What are the cues? Where is the information telling me how I should manage this? And of course, once you start mapping it, you start to realize, Hey, look, this is how evolved we think it is. So that’s going to impact how we manage it. And also it’s going to change. It’s going to evolve. So the methods of managing it later on will not be the same as now.
Murray: Yeah. I really like it as a collaborative decision making tool. As a way of surfacing assumptions and, discussing them. I think that’s really good. And I’m gonna start using it now.
Simon: Oh, well do tell me how you get on and do, tell me any horror stories as well, where it fails, where it didn’t work, all that sort of stuff limits are such an important thing to understand what the constraints what are the failures of the system. And also. All maps are imperfect. They’re all models. They’re wrong. It’s only a way of discussing as you pointed out.
Murray: All right. Now you mentioned a few places where people can go and find out more information. So there’s Wardley pdf.org.
Simon: That’s a community I’ve actually written an unfinished book. It’s about 600 pages long. That’s all free. There are a number of published books. Reaching cloud velocity is one by AWS there’s a new one coming out with Dave Anderson which is all about the flywheel effect. There’s James Eckhart’s the flow. There’s also a community hub. So list dot Wardly maps maintained by the community and it has links to videos and people have training courses and it lists some of the reading and books in there.
So you can see it’s got reaching cloud velocity in there. The artist strategy is also by Eric Sean. It’s also got links to courses you can do. Then there’s people like Ben building courses, learn Wardly mapping. He runs little courses on this. I just use it as a tool for my the research that I do. Then there’s things like map camp uK.
There are actually many different slack groups. So you’ve also got things like map camp. So this is a virtual conference where people from all over the community come together. And this year they’re talking about, oh, there’s a whole range of speakers, lots of interesting issues from what sort of financial system can support humanity best.
And so they’re using mapping to look at that problem to look at mapping and healthcare. That’s an interesting one. They did some work on sexual health in Brazil use of mapping for that. And also looking at mapping healthcare systems as well. It’s all sorts of different fields all sorts, different people as well. So I think that’s a starting point.
Murray: There’s a lot of resources there. That’s great.
Simon: It’s all creative commons. The community is amazing. They’re incredible people, very kind. And I find them wonderful, and I also do research projects as well, and that’s all creative commons. So there’s a number of tools out there for building maps. One that I I use a lot is online board maps. When you open it, it’ll just be blank. You’ll see, there’s something called example map, and it just shows you, and you can actually code the map itself. There’s several others coming in along this sort of line as well. And so what I do is I build maps in there and I store ’em in GitHub. My GitHub repository, if you’re interested I’ve got 12 research groups looking at everything from agriculture to construction, to defense, to finance, to government. Those maps are all open. That’s what we’re working on at the moment.
Murray: Very nice. Excellent. Thank you very much, Simon for coming on. been great.
Simon: Oh, thank you for inviting me. It’s been a delight.
Exit: That was the no nonsense agile podcast from Murray Robinson and Shane Gibson. If you’d like help with agile contact Murray evolve, that’s evolve with zero. Thanks for listening.