Watch now (here is a direct link)
Listen now
Living with complexity and exploring the Cynefin framework
Douglas Squirrel, Author of Agile Conversations
About this podcast episode
🌪️ Living with complexity and an exploration of the Cynefin framework
Douglas Squirrel, author of Agile Conversations, joins Bill Raymond on the Agile in Action podcast. Douglas discusses the misconceptions around agility in tech teams and the importance of experimentation in complex environments.
Bill and Douglas discuss the benefits of rapid experimentation, handling complexity and chaos, and avoiding common pitfalls in software development.
In this podcast, you will learn the following:
âś… The Cynefin framework for managing complexity
âś… The importance of experimentation in software teams
âś… How to avoid common development pitfalls
🎉 Strategies to make tech teams more profitable
Transcript
(transcripts are auto-generated, so please excuse the brevity)
Introducing Douglas Squirrel
[00:00:00] Bill Raymond: Welcome to the Agile in Action podcast.
[00:00:02] I am your host, Bill Raymond. And today we are joined by Douglas Squirrel. Douglas, how are you today?
[00:00:07] Douglas Squirrel: I’m doing great. It’s a sunny day in England.
[00:00:10] Bill Raymond: Wonderful. You are the author of Agile Conversations, and you’ve been on the podcast before to talk about that with one of your colleagues. And you’re an expert in making tech teams profitable. We’re going to talk about living in a world of complexity. But before I do, I always like to ask this question.
Defining Agility in Software Teams
[00:00:27] Bill Raymond: What does agility mean to you?
[00:00:28] Douglas Squirrel: An overused term for software teams, so I’d prefer not to use it too much where I really want people to be agile is in experimentation. I want them to find out more about the space they’re in, and software is uniquely, wonderfully placed for that, because it costs you nothing to try again.
[00:00:48] And what’s so sad is so many of us have a very high bar for how good our software should be, and I wish we would all put that bar right at the bottom so that we could experiment more.
[00:00:58] Bill Raymond: Yeah, I think that’s great advice. This is something that we sometimes just jump into. We have this idea, let’s implement that idea as opposed to let’s try out some of those concepts and make sure they work for our customer and they’re adding value.
[00:01:13] Douglas Squirrel: And the great thing is we can find that out quickly. And so many folks have a negative reaction to the idea that an experiment might fail. But in fact, in the quadrant we live in, that’s where we’re going to get into this complexity idea. The quadrant we live in. That’s the only option. The only thing we can do, is to do experiments, but we fool ourselves into thinking we actually know how things work and we don’t.
Introduction to the Cynefin Framework
[00:01:35] Bill Raymond: Before we talk about complexity, obviously we want to talk a little bit about the framework that you’re going to focus this around, which is the Cynefin framework. .
[00:01:45] Douglas Squirrel: You’re Welsh is really good, Bill.
[00:01:46] Bill Raymond: Thank you very much. I have it. I have it here spelled K U H N E V I N, which is exactly not how you’re supposed to spell it. It’s that’s my spelling to make sure I get it right, but I think it’s actually C Y N E F I N.
[00:02:02] Douglas Squirrel: Look, you write you, you speak. You’re perfect. You should move to Wales.
[00:02:05] Bill Raymond: Okay. Perfect. That sounds good. So let’s see if you could just maybe give us an overview as to what the Cynefin framework is.
[00:02:13] Douglas Squirrel: I certainly can. So it was invented by a guy, guess what? He’s a Welsh guy and his name is David Snowden. Most people call him Dave and Dave had this wonderful idea when he was at IBM and he saw repeated issues coming up in his consulting work. What it basically gives you is a way of orienting yourself. It’s a way of understanding if you’re dropped onto a mountain, like Snowdon in Snowdonia in in Wales, which I think Dave may be named after a big mountain in Wales, you might say well, I don’t know if I’m in the North face or the South face.
[00:02:42] I don’t know where I am. I don’t know how to go up or down. The framework gives you a way to figure out where am I and what should I be doing next. And let me explain that because you’re probably not climbing mountains, but you are dealing with problems in your product and your software that require this orientation.
[00:02:58] I’ll illustrate. If you were trying to figure out how to deal with a vehicle that you had, you might be in several different circumstances. I’m going to give you different vehicles and you’ll see how the different quadrants of Cynefin work. The quadrants are just four areas. You don’t have to worry about where they are physically.
Exploring the Simple and Complicated Quadrants
[00:03:15] Douglas Squirrel: But the lower right quadrant, is called the simple quadrant. And that’s like the car I have in my driveway and it’s an electric car.
[00:03:22] It has a battery. I don’t know where the battery is. I don’t know how the battery is connected to the wheels. I don’t know how the car works. All I know is I get in it. I turn the wheel, I push pedals and buttons and things, and the car takes me places. Anybody could figure that out. You don’t even need the manual.
[00:03:37] It’s very simple to understand what’s happening in that car. With me so far?
[00:03:41] Bill Raymond: Yes. Thank you.
[00:03:42] Douglas Squirrel: Excellent. So the next quadrant is the next most complicated, and it’s called the complicated quadrant. Imagine now instead of my car, I have a Formula One car. It’s ready to go on a race at Monaco or wherever. Now if I were to climb in there, I couldn’t find the gear shift, because gear shifts in Formula One cars aren’t where they normally are.
[00:03:59] And figuring out even how to put fuel in it or anything like that. Completely beyond me. I wouldn’t have the first idea. However, there are people who know everything about formula one cars. There are super duper experts and I could hire Lewis Hamilton or Max Verstappen or one of those people I don’t follow this sport, but I could hire them and they could come to my house and they could drive the car.
[00:04:18] So the complicated domain is the world of experts. There are rules for what you should do. They’re not obvious or simple to people like you and me, but you can get an expert and you can follow the rules. This is the most attractive quadrant. This is the one we keep thinking we’re in because we think, gosh, how hard could this software be?
[00:04:36] Wait a minute. People have been building e-commerce websites all the time. I remember building one in which we had 4,000 new products every week. And it’s looked like you could just use Shopify. Boy, you could not use Shopify. Boy, that did not work. And it looked like it was complicated, but in fact it was complex.
Understanding the Complex and Chaotic Quadrants
[00:04:51] Douglas Squirrel: So that’s the next quadrant over. And that’s the one where almost all of the viewers and listeners here live in, but you don’t know it. And it’s easy to fool yourself here in the complicated quadrant. In the complex quadrant, you have a new car that no one has ever driven, or even better example, you’re a test pilot in a new airplane that has never left the wind tunnel.
[00:05:12] Now, the problem with a new airplane is there are so many different components, it’s ten times more involved and intricate than a Formula One car. There’s, it’s moving in three dimensions, and it has all kinds of control surfaces and things, and no one’s ever flown this thing. So there, there might be a manual that says how it’s supposed to work, nobody’s checked it, and you take the plane up, and you have to figure out what does it do when I dive steeply?
[00:05:32] What happens when I turn and try to, do a roll? And you find out suddenly their plane falls apart and you have to eject. So you don’t know that these are the characteristics of the airplane, and no one in the world knows. You can’t hire an expert. No one’s ever flown it before. That’s what we are normally in, in software, even though we fool ourselves that we’re in complicated.
[00:05:53] Now, I’ll say more about that in a minute, but let me get to the final one, which is if we go all the way around chaos. Most people never encounter chaos. It would be unusual to encounter it. That is like an airplane in a tailspin. So if you just see the plane just spinning around there’s no connection to what you do with the controls.
[00:06:09] You pull back, you push forward, something else. What you do has no effect on the situation. All you can do, there are no rules, there is no connection. All you can do is try stuff and hope to get yourself into one of the other quadrants. So that’s unusual. Although lots of people think they’re in a chaotic situation.
[00:06:24] They think there aren’t any rules here. My organization is just insane and doing crazy stuff. So all I can do is put, keep my head down. But in fact, there are rules. In fact, there are things you can discover by experimentation and you can discover how your organization works. Now I can say a lot more, but how am I doing so far?
[00:06:43] Bill Raymond: That’s great. For those that are listening to the podcast, I will provide a link to this graphic so that you can see this and follow along. And if you’re watching on YouTube, you’ll see that on the screen when it’s brought up. So I’m guessing that we have these four different quadrants.
[00:07:00] We have clear, complicated, complex, and chaotic.
[00:07:03] Douglas Squirrel: And by the way, Snowden keeps changing the names of them as well. You can tell that Dave and I are we’re, I’ve met him. He’s a great guy. But he makes this work more intricate than it really needs to be. I call it simple. He calls it clear. He’ll call it something else next week. We got the idea.
[00:07:16] It’s like my car. Nobody has any difficulty using it
[00:07:20] Bill Raymond: And I think what I’m hearing you say is if you’ve got this clear or a simple then you’re in a pretty safe space. You’ve maybe built a website for your company. If you’re getting into complicated, then maybe you’re starting to create something that’s unique, but you’re maybe pulling together other technologies that exist.
[00:07:39] Whereas when you get
[00:07:40] Douglas Squirrel: and people have done it before. That’s the crucial thing with complicated is that it’s well understood how to do it. So building a website in 1999 was a pretty complex activity. Nobody had really built an e-commerce website before, Amazon and a few others had made it, but buying stuff on the web, really not well understood 2007, eight, nine.
[00:07:59] I remember looking up websites that said things like so you want to add a button, we were all trying to figure out how this stuff worked in the aughts. And now we have things like Wix, we have Squarespace, we have stuff that’ll just make it for us. And we all have Facebook pages and such.
[00:08:12] So building a website has made that full journey from Hey, nobody understands how to do this. We’re discovering the rules. They’re there, but we don’t know to complicated where you need experts. And when you hire them, they can do it for you. And now anybody can do it. It’s now moved into the simple realm.
[00:08:26] Bill Raymond: I would like to understand this chaotic environment and what that looks like. Also if it’s a good thing or not?
[00:08:34] Douglas Squirrel: It can be a good thing. Entering chaos can be tremendous for creating very new innovative ideas if you do it intentionally, if you say, all right, I’m now going to enter this domain where not only no one understands the rules, but there are no rules. And this is new and unshaped that we’re going to generate a whole bunch of new things, but we don’t know where we’re going to wind up we don’t know what outcome we’re going to get.
[00:08:56] Now I’m trying to think of a positive example. While I’m thinking of that I’ll give you a very simple negative example. I remember being woken up at one in the morning when I was working at an e-commerce company. The same one with the 4,000 new products every week. And the, I was woken up by our hosting company and I groggily tried to figure out what was going on.
[00:09:12] Eventually I figured out we were under distributed denial of service. So there was so much traffic coming to our website, we’d actually knocked out the data center. So the entire hosting provider said nobody who buys our services can use their services because your website, so you better do something about it.
[00:09:30] I didn’t have control over it. There were things that the hosting company was doing and I more or less was an observer and I could just watch them try to filter the traffic and stop things coming in. But I, those weren’t my servers.
[00:09:41] There weren’t my Firewalls and so on. It was upstream from me and I could observe that our website was definitely down. Nobody could buy anything. Luckily, it was one in the morning, so we didn’t have quite so many people purchasing cheap handbags on our website. But in that circumstance, I really didn’t have any control and no action I could take would have an influence on what happened.
[00:10:02] Now. I’ll try to come up with a positive example for you. When somebody like Pixar, really creative, really innovative movie studio gets a bunch of really clever people together in a room to come up with a new exciting movie. They really don’t have much control. The people who put them in the room can’t say we’ll make sure we put this person in the room and then it’ll turn out to be a cowboy movie.
[00:10:23] We need a Western, we’ll put the Western expert in there. Because when you get really creative people with a really fantastic process, like Pixar has, they may come up with a completely different movie, not the one you expected. And that’s a good thing. So you want that disconnection between the actions and the results, when you can plan for it and you’re happy with completely different results.
[00:10:43] Because they’ll come up with an amazing, super duper movie, but you can’t control it or say by putting these folks together in this way, we’ll come out with this this result. So there’s a positive and a negative example of chaos in software, at least adjacent realms.
[00:10:56] Bill Raymond: Because we deal in these complex environments and sometimes we do find ourselves shifting into this chaotic environment. Sometimes it’s hard to get us ourselves out of chaotic environments very often, if it remains that way too long.
[00:11:13] So I’m curious, what are some of the things that we need to do to deal with burnout?
Dealing with Burnout in Complex Environments
[00:11:19] Douglas Squirrel: Burnout is a different topic. So I’m very happy to touch on burnout. If you were to be in either of these environments without control for a long time, you would definitely get to burnout. Simple and complicated usually don’t have burnout associated. They have boredom that people get ennui, they get oh man, this is anybody could do this.
[00:11:37] And we got an expert here telling us what to do. Why am I here? So the kind of danger on that side of the equation is it’s just too boring and people leave if they want to be excited by new, clever things. But on the other side, you’re right. If you allow chaos to reign for a long time, people say, there’s just nothing I can do.
[00:11:54] The most important thing to do, especially in complex, but it also works in the chaotic environment, is to try things.
[00:12:01] And so that’s why I’m always coaching my clients and helping people on my forum and so on to do more and more experiments and to praise results, whatever they might be. And to get their cycle time as low as possible so that the time between when you have an idea and you say, I’d like to try this experiment.
[00:12:18] Maybe this would improve our conversion rate. Maybe this would improve our customer satisfaction. Maybe fewer people would churn or complain to customer service. If we had this feature, you want to be able to get from that idea. To measuring that it’s a reality or not to, to sensing whether it worked or not in as fast as possible.
[00:12:37] I usually teach people to do that within one day.
[00:12:39] Bill Raymond: That’s a fast turnaround.
[00:12:41] Douglas Squirrel: It’s an extremely fast turnaround and it’s so fast that engineers and others often say well, that’s impossible. We could never do that. And I have a standing bet on my website, which I’m happy to fulfill with your listeners and viewers. If you can find something that you think couldn’t possibly be tested at that speed, please get in touch with me and I will buy you a beer if you can find one that I can’t do. I have never had to buy anyone a beer because if you work closely enough and thoroughly enough at it so that you can experiment, you get a meaningful experiment, maybe a tiny experiment, maybe the smallest possible change, but it will tell you, am I on the right track or not?
[00:13:17] And that’s the exploration philosophy that you need in the complex domain, which you almost certainly are in. And it’s very different from the let’s go from A to B philosophy that you have in complicated and simple where you’re simply saying, look, everybody knows how to do this, or an expert knows how to do this.
[00:13:31] We just have to follow the steps. This is a territory with no maps and we need to make the maps as we go.
[00:13:37] Bill Raymond: Yeah.
Delivering Value Through Experimentation
[00:13:37] Bill Raymond: Now, I’d like to explore this a little bit more because we just did a short series on delivering value within product teams. As a matter of fact, when this publishes there will have already been a person named David Pereira, who is an author of the book, Untrapped. And he postulates that a lot of teams aren’t focusing enough on delivering value to the customer. Reduce your backlog, get things off of that backlog that don’t matter.
[00:14:09] Rethink how you’re going to do your product in the here and now thinking about how you’re going to develop in the future, but only the short future to make sure that you are delivering value to the customer. And we’ve had a few other podcasts similar to that, where we’re talking about delivering value.
[00:14:28] And In my mind, what you’re saying makes sense to test and validate with the customers. But we also have this drive to make sure that anything and everything we do has value. But what you said is maybe try a few different things. Now, I don’t think they fly in the face of each other, but I would like to hear your take on that.
[00:14:50] Douglas Squirrel: Let me use an analogy for you. And I think it’s relevant to the country you’re in. There were two guys who went and explored the center of what’s now the United States called Lewis and Clark and what happened to them at one point is there, they’re out there with their expeditionary force and they’re going through hostile territory with people who don’t like them and don’t want them there and there are bears and unknown territory and there’s no map and they come to a fork in the river.
[00:15:13] Part of the river goes this way and part of the river goes that way. Now, if they wanted to add the most value, their goal was to get to the headwaters of the river to get to where the river originated. If they wanted to do that in the most direct way. What they would have just done is go uphill because you’re trying to get to the top of the river.
[00:15:28] You want to go highest up, and so you just say, hey, look, that one’s more up than this way. We’re going to go that way. And if you have a map, if you have knowledge of the territory, if you know exactly how things work, if you’re in the complicated domain, that’s the right thing to do. Hey, let’s look at the map.
[00:15:44] Let’s get our GPS. Let’s figure out what’s here. They’re the ones drawing the map. They don’t know, and it is impossible for them to know. What you need to be doing in this circumstance is what Lewis and Clark did, which is to try different experiments. So they took little groups of three from where they were at the fork in the river, where they didn’t know what the right direction was.
[00:16:02] Maybe not to follow the river at all. Maybe they should hike for a while. They didn’t know that, so they sent out little groups, and the groups went out for a couple hours to climb local hills and see what they could see around them and figure out, should we go this way or that way? They came back.
[00:16:14] It wasn’t clear. So that experiment of sending people out gave them a little more information, but not enough to decide we’re going to take this fork or the other fork. That’s okay. That was a good result. They celebrated that. Say great, good job folks. You found out this and that you found out we shouldn’t go that way for sure.
[00:16:28] That’s definitely wrong. And now we’re going to do more. So they narrowed it down to two possible directions and they sent out a larger force one way and a larger force the other way. And a few folks stayed back at base. And a couple of days later, now that they’d explored more. They came back and it became very clear that one fork was the better one.
[00:16:43] It wasn’t the one that was more obvious. Most of the people in the group thought we should go down one fork, because it looked better. It looked more like the rest of the river that they were following. The other one said, this one’s headed in the right direction, and you’ve explored it far enough that I think we’re going to find a waterfall there, and we’re going to see that the water drops a lot.
[00:17:00] That’s what they found a few days later. They found Missouri Falls. And they were able to continue. So this is the story of decision point. If you want to go look this up, it’s a wonderful story. You can go visit the location and see what they did. The point of the whole story is they did experiments that did work, but give them negative results.
The Value of Negative Results
[00:17:16] Douglas Squirrel: And those negative results were really valuable, and they increased the value of Lewis and Clark’s eventual map and that their progress, of course, toward the West Coast, which is what they were trying to get to. Had they not done those experiments, they would have produced less value, even though each individual experiment produced negative value.
[00:17:33] He’s don’t go this way, bears this direction, right? Wrong way. Don’t go this way. That’s useful data. And the danger is that far too often we fool ourselves into thinking there is a right way. We know the right way. We can consult the experts. We can find it out. And therefore we want to punish people who don’t follow the right way.
Challenging the Concept of Best Practices
[00:17:50] Douglas Squirrel: So this, for example, is why I outlaw with all my clients the phrase best practices. I say, you do not know what the best practices are. There is no chance that you could possibly know that. So stop talking about it. What we want to talk about is incremental value. How are we showing that we are headed in the right direction, which may mean going down the valley before we go up the hill.
[00:18:08] Bill Raymond: That’s a great way of describing it. Thank you very much for that. I think that was a really great story and a way to clearly articulate that we are still delivering value, even if some of the things that we thought we were going to do are even just thrown out when we, and we do, we go another way.
[00:18:25] Douglas Squirrel: Yes.
Explaining Failures to Clients
[00:18:26] Douglas Squirrel: And one of the hardest things to do, and you need really good account managers for this, is to explain to your client why that’s good. Hey folks, we just tried a thing and it was a miserable failure. It definitely didn’t work. Now, aren’t we so happy? Your client says, my god, I just wasted, a week’s worth of development time on this thing.
[00:18:40] Why don’t you guys build the right thing? So you really need to explain that carefully to the person who is paying the salaries, or else they’re going to cut you off and try to find an expert, and only when they can’t find one will they discover, oh, that exploration, that was right.
Embracing Complexity in Innovation
[00:18:53] Bill Raymond: Now, we did title this podcast, living in a world of complexity, and we talked about the Cynefin framework and you’re talking a lot about complexity. So I guess I should ask you this question because I didn’t upfront, is this the world that you believe that teams should be working in?
[00:19:14] Douglas Squirrel: Yes, absolutely. For the reason I was alluding to before. Which is that it’s boring. It’s not so interesting to be in the simple or the complicated domain. So the very best folks are not going to stay there. A couple of folks have just, as we’re recording this, have just left OpenAI.
[00:19:29] And I don’t know what they’re going to do, but they’re doing some very interesting ideas and new stuff because this territory is starting to get conquered. We’re starting to understand how ChatGPT and all its friends actually work. And to get a handle on it, get some manuals, get some ideas about what good practices are.
[00:19:45] I think you can get a degree in prompt engineering now. If you can’t, you will be able to soon. This territory is getting pacified and getting understood and mapped much more quickly than it might have done in the past. And I think it is really important for organizations as a whole to not be afraid of complexity.
[00:20:03] We’re going to enter this area. We’re going to explore it. The reason there’s value in it is because other people haven’t discovered it. If we were to just go build yet another e commerce website, if we were to just go, SpaceX just kept building Saturn five rockets, That would not be a very valuable company.
[00:20:18] That would not be we wouldn’t have Starlink, and maybe we’ll finally get to send Elon Musk to Mars, and I hope he stays there when he gets there. We would not have all the progress that SpaceX is making. What they’re doing is, as we’re recording this, they’re blowing up rockets every few weeks, but they keep going farther, right?
[00:20:34] Those are successful experiments with negative results that tell SpaceX, yeah, this part of the rocket works. The part that blew up? Not so good. Don’t do that one. And they are making huge progress very quickly, precisely because they’re entering the complex domain and somebody who’s just sitting back, Boeing has not been nearly as innovative as them and they’re four years behind because not that they’re having doors fall off airplanes, which is also happening to them, but their space exploration has just been much more by the book.
[00:21:01] They’ve operated much more in the complicated domain. They’re behind. They’re not as innovative. They’re probably not going to get to Mars first.
Signs You’re Not in the Complex Domain
[00:21:07] Bill Raymond: What are some signs that you are not living in that world of complexity?
[00:21:12] Douglas Squirrel: So that’s the must much less frequent mistake. Let’s make sure we talk about the other mistake, which is much, much more common, which is to think you’re in complicated when you’re actually in complex but the mistake you refer to also happens where somebody thinks, great, we need to be super innovative.
[00:21:26] We need to come up with new, amazing, super duper stuff here. And in fact, somebody will hire me to come along and review it. And I’ll say, you could buy this. There’s no reason why you should be building this software because this is actually a solved problem. This is much more egregious, but more memorable for your listeners, but you can map to all kinds of things where people are typically the error is to build where you could buy, where you’re doing not invented here. And you say, we can do it better. The most extreme example of this is that somebody came to me. And said, Squirrel, I want to tell you about my amazing new startup.
[00:21:59] We’ve spent years working on this. We have really clever people. We’ve invented this fantastic technology. It’s really going to revolutionize mobile devices. I said, great. I have a phone. This is when phones were relatively new. iPhones were still a kind of nifty novelty. And I said I got one, so what, show me what does it do?
[00:22:16] It turned out that what it boiled down to, they had fancy words for it, but what it boiled down to us is that you could select some text from one app. And then you could cut that text and you could paste it into another app. I said, I hate to tell you, iOS now has that. Apple beat you to it. And you know what?
[00:22:33] It’s kind of commodity stuff. You don’t have anything special here. And then the poor guy I was telling this to, just his face fell. He was really depressed. Once he figured out that what he built was basically cut and paste. With fancy names for it. Now there’s a good ending to the story. There’s a silver lining, which is once I kind of woke him up to that, he figured out some other things he could do to make sure that he was manipulating texts in a better way, had some security Aspects to it and he specialized into an area where he actually did have something innovative where it actually was complex But he hadn’t noticed that this was pretty well understood that it was something I could do on my phone without needing his fancy special software and therefore he was going to have no customers whatsoever.
The Danger of Not Testing Innovations
[00:23:13] Douglas Squirrel: Of course what he hadn’t been doing was to test all the way along. He hadn’t been going quickly around the cycle of probe, sense, respond of try something, get customer feedback, and then continue because the first customer would have said, I can already do that. Why would I need one of those? He hadn’t done that.
[00:23:28] He’d done it in secret. And that’s one of the dangers that leads you into this realm. So if you’re operating in that way, if you’re saying to yourself, we have new innovative stuff, it’s better than everyone else in the world. And you haven’t tested it. You haven’t gone out to actual customers and checked it.
[00:23:42] Red flag, danger, stop everything you’re doing. Go and check that with customers because you may be in the complex domain. You may be operating where you need to explore and make decisions like Lewis and Clark, but maybe not, maybe this is well known or not desired. Then you should try something else.
[00:23:57] So that’s the mistake that direction
[00:23:59] Bill Raymond: Yeah, I can see that. And I think if we were to go into your average company meeting where, you have the leader of a team and we’re sitting down, I can’t tell you how many times this happens and this is nothing against the teams. It’s just something that you have to really think about and almost embed into your process.
[00:24:20] Is, very often I’ll sit with these teams and they’ll say, yeah, the, we’ve rolled this feature out, is there on our backlog? We got it out to the customer.
[00:24:30] Douglas Squirrel: after months of effort, we had the whole team where we had different people working the back end, the front end, the data scientists, everyone did the designs were beautiful. And,
[00:24:37] Bill Raymond: and do we know if the customers like it? And sometimes we don’t ask that question. Sometimes that gets lost in there and we find out very quickly when the customer doesn’t like it. And that’s when you have to unravel and say, oh we just kind of rolled out the feature. We had lots of meetings to talk about it.
[00:24:55] So we figured we nailed it, but we never actually gave it to the customer to truly test it because that would slow things down.
[00:25:03] Douglas Squirrel: There’s a worse outcome.
Encouraging Frequent Customer Feedback
[00:25:04] Douglas Squirrel: And I see this very frequently, nobody ever tests. So they get the feature out, they celebrate, they have the pizza party, they have the exciting, they release, they go to the all hands meeting, they say, wonderful, new feature, excellent, version two, it’s so stupendous, and nobody ever checks with the customers, and so they never actually find out that they wasted their time, and they think, great, now it’s time for the next one, and the amount of money you can waste on this is why it’s so easy for me to help tech teams be insanely profitable.
[00:25:28] Because they’re so unprofitable when they do this. And so undoing this is something actually engineers know how to do. Engineers know how to release frequently, how to get frequent feedback from customers. They just don’t believe anybody wants them to do it. Once I can get them doing that, they often discover that they have wasted huge amounts of effort on things that nobody actually wanted and that they didn’t check.
[00:25:47] So they actually never knew it. They thought they’d had a great success and they don’t.
[00:25:51] Bill Raymond: Yeah. And it does take leadership to, if you will, I don’t know if allow is the right word, but to enable that behavior to let
[00:25:59] people know wrong as bad is okay.
[00:26:02] Douglas Squirrel: We want to get the negative answer. And in fact, if you’re a good scientist, you bias away from the answer you want. If you want to prove that apples fall down when you drop them, you make sure that, it’s a particularly light apple and it’s the, a day when their gravity is light or something, you’ve done something so that the experiment is biased away from your result.
[00:26:20] And only if, every time I drop the apple, it still falls down. Okay, now I know that gravity operates downward. You want to do the same kind of thing with your experiments. You want to bias away from customers liking it, bias away from it being a successful result. Because that way, you’re even more certain that it’s going to work correctly for you.
Designing Effective Experiments
[00:26:39] Bill Raymond: What are some things that I can go back with my team to start doing today?
[00:26:44] Douglas Squirrel: And I’ll come to the other direction, the much more frequent error as a way of answering your question. So you brought me onto it perfectly. Thank you. The most frequent error is that you think you’re in the complicated domain, because we’re engineered, our brains evolved, so that we thought we can think that the universe is ordered, that things are working the way we expect.
[00:27:03] We form theories, and then we believe that the world works according to our theories. Inconveniently, it often doesn’t, and things surprise us but they surprise us because our brains are wired to think, I’m in the complicated domain. I understand this. I’m in the simple domain. There, there’s a set of steps I can follow that will make sense here, but you’re not.
[00:27:22] And if you think, that your software is simple, that your software is known and well understood. And you’re just doing something again that you’ve done many times before. It is very likely that belief is catastrophically false. And you can find out very quickly whether it is false or not.
[00:27:38] You can undo that assumption by doing some experiments. And you want to do those experiments quickly, you’d want to try, just like Lewis and Clark sent out the little parties to go all around and North, South, East, West, everybody try go look around to see what’s here. You would want to do very quick experiments, which means getting into production fast.
[00:27:56] It means doing things like continuous integration, continuous deployment, all the tools your engineers know how to do so that you can get that experimental data very quickly. And you can rapidly prove whether the path you are following the next five sprints or the next 20 sprints that you’ve laid out in your roadmap are actually taking you toward a good outcome or a catastrophically bad one.
[00:28:17] So design experiments that let you shortcut that. Hey folks, what I’m going to do is I’m going to give you the report. And it’ll have no data in it. What on earth good would that be? If you got a report with no data in it, but the columns are there, you could check if the right columns are there.
[00:28:30] These columns are all wrong. Why are you working so hard to get all this data that we don’t need? Oh, that’s interesting. I had one report somebody asked for, and I calculated that the, if it was printed out, which is what they wanted to do, it would form a stack that was taller than the building we were in.
[00:28:45] And it would take human beings at least one year to read all the data in the report. And I said, do you really want this report? They said now that you tell us that, maybe not. Maybe we could do something else. It turned out they wanted analyses. They didn’t want the report. That kind of experiment where you can find out from your customers much more quickly whether you’re on the right track will tell you very fast whether the path you think you’re following, the best practice, the complicated process that your experts are telling you to follow is actually complicated or do you actually almost certainly live in the complex domain. You’re building something that’s never been built before.
[00:29:18] You need to be much more experimental. And then the great news is there’s wonderful things you can do there. We have fantastic tools that allow us to go through really fast cycle times. Even for my biotech customers where they’re building medical devices that have massive consequences, people cut off parts of their bodies.
[00:29:33] If these things give the wrong result, I get them releasing every two weeks. So if they can do that, surely you can release your e-commerce product, your finance product, something like that. You can get feedback from your customers every day. And if you can do that, when you’re operating in the complex domain, the profit is huge and the orientation you get, the mapping that happened, that goes on is so much faster than any of your competition, you’re going to be way out ahead.
[00:29:57] That’s the sort of thing you can do. Design those experiments, make sure you have the process in place that allows you to run those experiments very quickly and praise people when they do experiments that have both positive, good outcomes until you go this way. And the ones that say, here be dragons, bears, don’t go this way.
Real-World Examples of Experimentation
[00:30:13] Bill Raymond: I have a really good example of that. I was working with a group of scientists and they, these are labs, beakers and things like that in the
[00:30:20] Douglas Squirrel: room.
[00:30:21] I know that world. Yeah,
[00:30:22] Bill Raymond: Yeah. And we had this need to put labels on these beakers. And so we were actually building some software that was collecting the data for the the studies that they were doing and all that good stuff.
[00:30:34] But one little tiny piece of this that we’re going to roll out to all of the scientific community around the world in this very large company was to print labels. And so we were like we’re going to get those thermal label printers because they just print out the, the perfect size.
[00:30:49] You just wrap it around the beaker and we tested it. The team was sitting there like trying it out. It, it printed out the formulas and things like that. All these special characters that were needed and
[00:30:59] Douglas Squirrel: great to you. Your experts told you this works great. This is the path. We already have the thermal printers. Everything’s great. What went wrong?
[00:31:06] Bill Raymond: Yeah. Yeah. Then it goes into the lab and I don’t know if you know how a thermal printer works, but it uses heat to print. And so when you put the label onto an actual beaker with warm liquid, it turns black and you don’t get a label.
[00:31:19] Douglas Squirrel: Fantastic. And if you had done the experiment by printing one label by hand and putting it on one beaker, you would have found out very quickly whether that works or not. Can I tell you just one more story? Do we
[00:31:29] Bill Raymond: Sure, please.
[00:31:30] Douglas Squirrel: There are many stories like this, but this is one of my favorites.
[00:31:34] Airbnb is called Airbnb, because they walked around, the two guys who started it, they walked around with air mattresses, literally things you blow up, and they went to people’s doors, they would follow somebody else into a block of flats, big high rise apartment building. And they would go around knock on the doors.
[00:31:52] A lot of people didn’t answer their doors. A lot more people said, you guys are nuts. Why are you walking around here with an air mattress, but a few people said, that conference happening on the road. Yeah, I under I know about that. And I’d be happy to have somebody stay in my front room and they say, here’s the air mattress.
[00:32:07] And then they hadn’t built any software. They would just write down the person’s name and they would go back to a spreadsheet where they had advertised, come stay for this conference and they would match up the people. So they did this completely without technology, right? Hardly anything, basically pieces of paper and spreadsheets and things.
[00:32:24] Nothing fancy. And they discovered that, in fact, people wanted to do this. But they could very easily have discovered, and they would have been very happy to discover, that nobody would want to do this, and it was a completely insane business idea, because then they could do something else to try to solve the problem that they were working on.
[00:32:38] And I’m sure they tried some others that we don’t know about, that didn’t work out so well. That kind of experimentation is exactly what you need. And it’s exactly what you could have been doing with your thermal printer and your beakers to, to try it by hand with no technology and say does this actually work?
[00:32:53] Can I use it in a lab? Aha. No. Let’s here are dragons. Let’s go somewhere else.
[00:32:58] Bill Raymond: Exactly. Yeah. Those are great examples. Douglas Squirrel.
Conclusion and Contact Information
[00:33:02] Bill Raymond: This has been a great conversation. I really appreciate your time. If anyone wants to contact you, reach out, have a further discussion. How might they be able to reach you?
[00:33:12] Douglas Squirrel: So the best place to go is someplace called squirrelsquadron.com. I’m sure you’ll have the links in and everybody knows how to spell squirrel and squadron. And if people go there, I want to make an offer to all of your listeners and viewers. I have a nice little one page summary that takes you through some of this.
[00:33:27] And I would love to send that to you that describes some of these techniques and how you can track them and so on. Which we didn’t have time to cover here. Please get in touch with me at squirrelsquadron.com. You can join my community and get my emails and so on, but basically get in touch with me and I will get you that additional material that builds on some of the things we were talking about here.
[00:33:46] Bill Raymond: I appreciate that very much. And thank you for contributing to the community. I really appreciate that. If you want to reach out to Douglas Squirrel, of course, he just mentioned a few links here. Of course, we will have the full links. on the https://agileinaction.com website, where you’ll also be able to read the entire transcript of this podcast, watch the video and listen to the podcast.
[00:34:08] But if you happen to be in an app or on YouTube right now watching this, then just go to the show notes, the description, and those links will be there as well. Douglas Squirrel. Once again, thank you so much for your time today.
[00:34:20] Douglas Squirrel: Thanks, Bill.
[00:34:21] Speaker: Thank you for listening to the Agile in Action Podcast with Bill Raymond. Subscribe now to stay current on the latest trends in team, organization, and agile techniques. Please take a moment to rate and comment to help us grow our community. This podcast is produced in affiliation with Cambermast LLC, and our executive producer is Reama Dagasan.
[00:34:43] If there is a topic you would like Bill to cover, contact him directly at Bill.Raymond@agileinaction.com.