 This is going to be a little bit of theory and a lot of experience. So my background, my name is Dan Greening, and this guy right here. I am the managing director of a consulting firm in the United States, but we work all over the world, CENIX-RECT, and it's largely dealing with high-level questions for organizations. I've been CEO of a number of startup companies in the San Francisco Bay Area, and I was the VP of Engineering for Macromedia, which is now part of Adobe. So I'm today a certified enterprise coach, which means that I hang out with larger companies and try to help them transform to more agile practices throughout the company. I was the head coach for Skype for a couple of years. I was the head coach for Citrix Online for five years. I have a PhD in computer science and whatever. Anyway, I've been around. So this is the premise of this talk. Can you guys turn it down a little more? Because I can speak louder. So the premise of this talk is that, well, this is actually a discovery that I made in doing all these enterprise transformations, that the biggest problem that we had was managers who felt threatened by agile practices or didn't understand them, and they were very different than how they used to manage. And as a result, these managers would ultimately feel threatened, not just because they didn't understand it, but because it was really threatening to them. They had built a career out of moving ahead by following a set of rules and then following those rules that they had achieved this status. And once they achieved this status, then some rabble-rousers came in and said, let's use a different practice called agile. And they said, okay, that's fine. It's just another practice. I can deal with that. They will new practices in here all the time. And the agile people would come in and they would say, hey, thanks for letting us come in. And so they would start working, and then the managers would start realizing that agile was exposing problems in the organization, and some of those problems were caused by the managers. The impediments were coming from them. And so I started, and what we ultimately saw was a number of large companies that adopted agile practices, and they became very healthy. We could measure their health. We could see that they were getting larger market share, that they were winning. Despite that, the executives of these companies allowed agile to die. And so we saw Yahoo go through about four transformations where it went to agile and then lost the appetite to keep it up, and then declined, and then they would hire a bunch of agile coaches again, and it would decline, and that sort of thing. So I thought, wow, this is a really interesting problem, right? Like, how is that happening? So this slide, sorry, you know, one option is just to turn off them. You guys in audio to turn down the sound if you're okay. If you cannot hear me, please let me know. Alright, so Gallup is an organization in the United States that does a lot of surveys, okay? And so they typically do political surveys, but they also do managerial surveys, and they've come up with a program where they will identify for you in your company talented managers, high-talent managers. They'll find managers that can be developed into high-talent managers, and they will also identify for you low-talent managers. They've found that 62% of managers in the role are low-talent, meaning that they don't know whether they can even fix them or not, alright? And here are the characteristics of these managers. They do not generate excitement for outcomes in their report. They fail to organize workflows in their groups. So if you see a really disorganized group that is constantly firefighting, I think who has seen a group like that? Constantly firefighting, crazy stuff coming in, alright? Well, it's great, only a few of you have. Practically every organization I go into has this problem. So with these unorganized workflows, you get chaos in the workplace, right? Something very unhappy Richard Sheridan would not like this sort of organization. And so they allow good teams to stagnate. So a team gets good, but then it doesn't know where to go from there. They suffer dysfunctional teams. They don't go in and fix teams that are broken, and they seek convenience over best. And this is, I sometimes call this, phoning it in. Who knows what the term phoning it in means? Alright, I'll give you the answer to that. There were reporters that would go out into the field and they would have to write a story, but they had been drinking heavily the night before and they kind of half knew what was going on, and they would phone in to their editor the story that they wanted printed in the paper. And usually they were pretty mediocre, right? So that's seeking convenience over best. This is what our managers look like. They say, oh my God, something is wrong. You know, we just introduced Agile. It must be Agile. That's what's wrong with our company. We're seeing all these problems. We didn't see these problems before, right? And here they are, right? Agile is the thing that exposed them. It didn't create them. But if we kill the messenger, maybe we can keep our job for another couple of years. So what I'd like you to do is, let's see, we might have some sticky notes, but if you don't have sticky notes at your table, yeah, so sticky notes in your bag or there's this paper in your bag, if you would please pull it out. And then you have a pen as well, hopefully. And I would just like to take a couple of minutes and ask you to write down your company's mission. If you are working, if you're a contractor working at another company, that's fine to write down that company's mission. I just want you to write down the mission statement for the company. If you don't know the mission of the company, try to think about what it should be, what it is, right? Okay. So thank you for doing that. Right now I would like, can one of you share the mission that you wrote down? It doesn't have to be. You can tell me. Today is all about tomorrow. Okay, that's our mission. Okay, that's a pretty good broad mission. Anyone with a more specific mission? Making a travel easy, an airline travel? Okay, that's cool, making airline travel easy. I like that mission. Any others? All right, so I want to know from you what percentage of you think that the mission you wrote down is something that your colleagues would agree with. So if you're in a company that has the mission statement on every door, you walk in and you see it every day, you would probably think that close to 100% of the people would agree with you on the mission if you wrote the mission down verbatim. But if you made something up, but you think everyone would agree, it still might be close to 100%. So what I'm going to do is I'm going to ask you by quarters. So how many people think that 0 to 25% of the people, the colleagues in their company would agree with the mission that they wrote down? How many think that less than 25% would agree with the mission they wrote down? One, two, three. How many people that think that 50% or less, 50% to 25% would agree? Two, few more. How many people think 75 to 50% would agree with the mission? Excellent day. And how many people think that 75% to 100% would agree with the mission? All right, so about somewhere between 50 to 75%. That's pretty interesting. That's pretty high. So how many of us had to guess what our mission was? Please be honest. Let's just raise our hands. Not too many. That's interesting. So that's a good sign. You work for great companies. My job is done. You can go home. All right. So the reason I wanted to bring up this question, I want you to think about this question, is because it has to do with managers who are good. So good managers measure leading indicators. And they measure leading indicators for something, right? What do they normally measure in companies? What do you normally want to measure? What do you want to measure in a company? Performance. What do you mean by performance? What you expect and what you deliver. And why do you measure that stuff? You keep your customers. And by keeping your customers, they do what? They give you money, right? So usually, people, you're one of the things, the key, well, some people call them key performance indicators, but I'll call them success metrics. Your company's success is based on the ability to get revenue. And is that it? Just making a lot of revenue? Is that the most important part of a company? What else is part of it? Value. Okay, what does value mean if it's not revenue? As a sustainable entity, maybe? Like long, long-term revenues, maybe? Because you're achieving a mission or something? I agree with that. That is an important thing. What else is important? So that's kind of value, but there's one more thing that, you know, people either miss one big thing or the other big thing. So you guys have focused on the revenue side. What's the big thing that hasn't been discussed yet? Talent pool. That's kind of okay. Getting close. What? Keeping employees engaged. That's kind of close. What? Value, value, value. You guys are like, wow. All right. Say again? Loyalty of the customer. That's a long-term value thing. Impact from users. Okay, guys, I have to give it to you. It's cost. All right? There's two parts of every business. There's value and cost. And together, they make what? Profit. They make profit, right? So if our costs... Go ahead. Yeah. People in the organization. Yes, making the people happy. But for an organization, for a company, I just want everyone to be clear. Not everybody's been in the CEO role, but I'll tell you, profitability is number one. Okay? It's number one for the CEO. It might not be on the mission statement, but it's there. And that's because if we're not making a profit, what happens to the people? They're gone, right? Like we have to fire them, right? If we aren't making a profit, how are we doing on our mission? Like we have this lofty mission that we're going to make air travel more fun, right? Well, we can't do it if we're not making money because we're going to go bankrupt. So cost is a major factor. What's interesting is that strong and agile practices, as we normally define them, is cost-centric. It's all about getting velocity higher, which is producing more functionality, but we don't really have value attached to that functionality necessarily. And it wasn't until lean startup happened recently, like 2011 or something like that, that we started realizing that we're going to apply similar rules, similar agility to the value creation part of it. So we shouldn't just try to make our teams faster and faster and produce higher and higher quality code because that code might not be valuable, right? We just might be producing more crap faster. So that is... So leading indicators, I've just been talking about not leading indicators. I've been talking about lagging indicators. So profit is something that doesn't appear until after your investment has started to achieve revenues and that sort of thing in the marketplace. And that often takes time, sometimes years, okay? So in order to start thinking about that, you have to measure leading indicators. They're not the success metrics we're after, profitability, but we want some indication that we're gonna make a profit someday. If we don't have that, we'll make bad decisions right now. So there are four other patterns here, and I put them up here just to give you an introduction. I'm gonna go through each one of them separately. These are basically the agile-based patterns. And those agile-based patterns are present in every single agile methodology, including Scrum, XP, and Lean Startup, and a whole host of other things. So let's talk a little bit about these leading indicators. So the first thing that agile managers do that is different from what I would call standard managers or even high-talent managers is they focus on metrics. They focus on success metrics, profitability, how well we're achieving the company's mission statement, and of course with profitability is revenue and also cost and a bunch of other things. And so those success metrics are interesting, and we ought to know what they are and they ought to be specific. So the more general they are, the less they guide us. So if our mission statement says we'd like to make a profit and we'd like to satisfy our shareholders and we want our stock price to be high, that's like the mission statement of every single company on the stock market, right? So we want it to be specific because the narrower we can make it, the better focused our workforce can be. And so mission statements that are, you know, you just take a template off the web and you say, okay, this is our mission statement, that doesn't work. There's actually a practice that Toyota pioneered called Hotion Connery that is about taking these things and turning them into shorter-term indicators. So let's think about some... I'm going to go to the bottom here and I'm going to say let's talk about some leading cost indicators. So on a strong team, we often measure velocity. It's really a cost metric. I mean, it's kind of the inverse of a cost metric, but it's still a cost metric. The faster we go, the more functionality we can produce in a particular time period, the lower the cost we have for developing software. And so that velocity is a cost metric. And similarly, quality as defined by a scrum team usually is also a cost metric. It's how many bugs went out to the field and then had to come back and be fixed again. So there's a cost to that that isn't reflected in the velocity. What we need to do is we need something to account for the cost of the bad quality stuff going out in the field and that might be the bug counter reported by customers. We also have cohesion might be an interesting thing to measure, but the thing that people like to measure that's easy to measure and that's a good thing is happiness. So how many people measure happiness in their scrum team? Does anyone here do that? Some people do. So the way you do that is you go to the team at your retrospective, you say what went well, what went not so well and then you don't ask what are we going to change yet because you want to ask the happiness question before you do that. You say, okay, how happy are we? I want everyone to think about are they happy, unhappy or I don't know if I'm happy. And then you say one, two, three, boom and everybody does this and you average it out. It's one minus one and zero and you average it out and it's some number, right? The happiness of the team is found to be related to its productivity. So this is something that Jeff Sutherland wrote about and I've used it ever since I learned about it, I've used it and I found that it's really helpful for getting people to think about what would make them happier in their workplace but it's a cost indicator. It's how well do we communicate? We measure these things and they're aligned together. We get the outcomes that you would hope for a well-run team. People work in synergy. Sorry about that. People work in synergy because we're all aligned up to deliver the same sort of thing. So there's coherent value delivery so we are understanding what we're delivering by working together. We deliver packages of value that are more interesting and then we can make, we can conclude things about the market and so this brings me to the leading value indicators. I talked about leading cost indicators now we'll talk about leading value indicators but they're things like reach. How many people did our product touch after the last sprint? How engaged were our users? How many features did they use in our product? Did we retain our customers in the last sprint or did they drop off? And did we have any monetization that appears to show some success? And virality finally is one of the most powerful metrics and one of the hardest to think about or measure. It's when we touch the customer how many other customers that person bring in. So if that's a high number like if it's over, you know, 0.1 per week our business is golden and we might as well just park and go home and run our money or run off like DJ Malia or whatever, right? Take it home. DJ Malia has a house in my hometown or my current town, Lita. It's a really huge house. And he can be a jerk there too. Just letting you know. So it's not limited to you. So here's a mission statement from Hec. That's what I like to call it. Hec is not hell. It's like one step above hell. It's kind of mediocre hell, right? Like if this was hell this would talk about we're gonna kill people and do all this other stuff. But this is Hec because we don't really do anything. So at the Metro department we strive to enhance the quality of the life of our customers by working with the citizens in our community as well as public and private entities and professional, effective, efficient and sensitive matter to promote public health and preserve the environment. What does this company do? What? It makes, it loves people. It's a loving company. Actually, blah, blah, blah, right? So here is a better statement. I actually looked up who these people were and I'm going like what the hell is that, right? So these people reliably supply clean water to the citizens of Lynchburg. And they also capture and treat its sewage. So they didn't put it in their mission statement. They might be embarrassed about their mission, but the bottom line is it's really valuable, right? And then now we understand this part. We promote public health and preserve the environment because that's gonna be an issue for these folks. And then they have to thoughtfully balance the needs of citizens, companies, society and future generations, right? So that's interesting. Guess what? We can start to measure this stuff, right? We can measure the cleanliness of the water. We can measure our capacity to clean water. We can measure how much sewage we treat, all that stuff, right? So here's our little guidance. Metrics, what metrics matter? The metrics that align as best as we can to the success metrics that we're trying to achieve. So the leading metrics have to align well to our success company. How do we better align them? We have to have a lot of conversations about it and I'll talk about some of those. Perversity is when you have just one metric usually and that one metric sounds good like velocity but you apply it to your team and guess what, they start generating a lot of buggy code. They optimize for that one thing and we forgot to measure some other thing and so we got a lot of buggy code and then we see, oh, we better add quality to our collection. And then we're gonna ask how to measure managers and you'll see some of that stuff in here. Here's some references. This stuff is online. You don't have to desperately write it down. Just look in the talk and you'll see all these slides. All right, so I'm gonna give you an example of how to apply this stuff in a big company and here is one way that I did it at Amway. So we had a situation where we had project proposals before that would go in front of the finance department and they would say, hey, how much do you want for the next 12 months and they would say, oh, this is how much we want and sometimes it'd be 18 months or 24 months, right? And they would say, this is how much we want and then the finance department would say, well, you better write down what you're gonna do because we don't want to fund you if it's some kind of stupid thing, right? And so what they would do is they would actually write a project plan, a waterfall project plan, right? Maybe it wasn't as hard as doing it really like PMI would do, but it was pretty close. So we said, hey, you guys need to do something else, right? You need to fund teams. You need to fund an intent, a customer base and then maybe what you ought to do is have this ability to cancel projects or redirect the funds in the midstream and the finance department said, oh, that's weird because normally when people underspend their funds we think that's bad, right? Because they had a capital expense funding package and then they gave us back money and it looks bad on our balance sheet and we can't spend it on something else and I'm going like, are you kidding me? Like saving money by not delivering a bad product, that's a great thing, right? We need to take that money and do something else. So here's what we did. We said, okay, so what we want this group to do is we want it to treat agile project proposals slightly differently. We want you to look at what are the mission metrics that are improved by the product. So is it revenue or is it cost savings? Is it something related to the mission? We want to know that. And then we want you to tell us, you project people, tell us what leading indicators you're going to measure, they're going to tell you whether you're getting close, okay? We want you to tell us that because we're going to come back and ask you quarterly how you think you're doing, how are your leading indicators matching what you said you thought you could do. And then finally we want you to give us a rough project roadmap that shows us that you're going to deliver something to users and you're going to actually test this up front. So much so that you should be able to tell us, hey, this project, we thought it was cool but it kind of didn't work out. And then the leading indicators show us that and then might be able to be healthy for the company where we can say, okay, thank you so much for creating this great project. We love you that you're measuring these things and we love that you failed early because look at all this extra money we have, we can spend it on something more profitable. That's a very healthy discussion to have. It's a discussion that only an agile company can really have. And you see as I'm talking through this we're talking at the highest levels of the organization. This is a senior review here. And this person is going to help drive agility in the rest of the company just through their policy of how they fund a project. That's kind of cool. Anyway, that's an example. The leading indicators you probably aren't, you're getting there in the agile world, right? So I was going to do value stream analysis but I've taken too long so I'm going to try to speed up a little bit and help you along with these patterns. So the second pattern that agile managers do is they limit work in process by trying to discuss it. Yeah, that doesn't work for agile, does it? Why don't lagging indicators work for agile? Yeah, it's like a year ago, dude. I mean, what am I supposed to adapt to? Right? It's kind of hard, right? Actually, this is a long discussion. You have to have this discussion. You basically want to talk to as many people as possible. If you only talk to product managers or even lower product owners you're not going to get very far. You need to have these... I mean, this is one characteristic of agility in organizations. The highest level that you're talking to is where the conversion ends, right? It goes up that far and it doesn't go much further. And so if you can have these discussions with senior VPs you're doing well. I mean, it's going to help the organization. In fact, if you can... I'm going to assert this. This is a hypothesis, a theory, that if you can convert the CEO to be agile him or herself you don't actually have to work very hard as a transformation consultant because your CEO is going to make that happen, right? If they're measuring leading indicators, if they're trying to adapt rapidly, if they're limiting work and process by time and cost at the highest levels, guess what? Underneath them have to be agile or they will not be able to do that. So the CEO can bribe the whole thing. And actually this is what we often see, right? Like we see CEOs like the CEO of Facebook deciding that they're going to be agile and really embraces that. And guess what? The whole organization is pretty damn agile. Same with Spotify, right? They do have a bunch of agile coaches there and that was probably really useful and that is, but they're an awesome company and I use their product. So anyway, back to here. Second pattern. First pattern was leading indicators for success. Second pattern is limit work and process by time and cost. So we have to organize teams. Here's what an agile manager does. Of course they organize teams to rapidly deliver a portfolio of products and a bunch of products. That's standard scrum and scaled scrum and whatever. But what I don't often hear people talk about but I've done many times is organize teams to deliver strategy. So when a company decides to change something, how we're doing recruiting, how we are promoting the company, what our branding is for the whole company, how we do marketing, how we do financial planning for example. Those things take executive time. It might not be VP executive time, it might be directors or something like that. But those take time and it's creative time. We heard a previous speaker talk about creativity and design and all that stuff. Well, working as a manager, designing programs in the company to support promotions and performance reviews and all that stuff, that is creative. That is creative work. And if you can get teams to work together, you generally get better outcomes. So I have organized strategy scrum teams in virtually every company I've worked in since I became an agile guy. And in doing that, we get a lot of great outcomes. So I'll talk maybe a little bit about that. We also abandoned doomed work early. So the minute we think that this is doomed, we have a heads up and we say, hey, you know what, we're going to take that money and we're going to use it for something else. And then if we have money, we can produce other products. The other thing we can do is if we have extra time from our engineers and that sort of thing, we can use that time to help them learn to create more capacity. So we get lots of great outcomes. Fewer death marches increased innovation appetite because we're limiting work and process by time and cost. We're shipping very frequently, once every four weeks at least, and we're testing whether the market really likes it or not. Okay, because we're doing that, our innovation appetite increases. It increases because failure is not a disaster. That's the big thing. The reason people fear failure in large companies is because they have been multimillion-dollar disasters and they end the career of managers and VPs at least at that company and sometimes at a lot of other companies too. It makes them difficult to hire. And so by having these short time and cost things, you're actually making the world of your manager better. And you can have this conversation with them, right? These are the sorts of conversations you would want to have with these high-level managers so they will support your agile transformation. It's advantageous to them. You discover more about customers and stakeholders. Work becomes more fun because they're working with teams on strategy. And there is a risk that some project-centric people become more mission-centric. If your project-centric people do not become more mission-centric, you're going to have discord, right? But this is the typical thing with transformation. We always have these problems. So I wanted to talk about limiting some examples of applying this principle in real life. I was at Citrix. I was the head of agile coaching there. But one of the things I did was I said, okay, this is crazy. We're funding 30 different projects running simultaneously because all these product managers have their pet project and we want to fund them a little bit and they're all feeding desperately to have at least two people on each project or whatever. And then, of course, the most profitable projects suffer because they're understaffed because we're trying to staff everybody's happy play. So I said, you guys, this is a crazy thing. We need a backlog of cost and time-limited projects. So we said every project, no project can be longer than three months. It can have no more than three teams on it. And now we're going to put it in a backlog and we're going to staff it every quarter and we're going to put up to three teams on each project. And then when we run out of teams in the engineering department, sorry, we are not working on any more projects. So that was limiting our work, right? And the nice outcome, which is a little bit hard to see, I suppose, but you can see these dots. This is the duration, the release duration. So from the time we started a project until it finished, got as high as, are you kidding me, 42 months. That is a lot, right? Like three, almost four years, right? So I got hired, I don't know, like here, and I went, oh, this is crazy. And then I actually, one of my employees said, hey, we should just scrum in our little team. It was like a team of 70 people. And I said, oh, I don't know what that is, but okay, you know, like, and we started doing it. And I went, oh, this is really awesome. And I started talking to people and saying, hey, you know, we really out of your scrum everywhere. And then we hired Shweber, I think, to come in and train. And then we started doing this enterprise scrum idea where we did projects and that sort of thing. And it drove the release duration down to about one and a half months or so. So we were basically releasing every one of the months. And this was my first indication that managers, it's dangerous to have non-adult CEOs. The CEO hired a person into the organization that was an expert at business organization structures. And this person said, hey, you know what? You know, there are too many competing product managers who are always like begging for the CEO to fund their project, whatever. What we really need to do is reduce this discord. And the way we do that is silo the organization into three separate organizations with three separate engineering departments. And we're just going to do, you know, budgeting up front every year. And that's her thing. And guess what? Guess what? What do you think happened? No one wants to say it. Here's what happened. It got less agile, significantly less agile. Suddenly this number, the release duration, started going up. What a surprise, right? We weren't limiting things by time and time. Here is the third pattern. So agile managers proactively experiment. What I mean by that is that they don't just sit there and run these time and cost limited projects and then use leading indicators to figure out how they're doing and then hope the variance in the marketplace provides enough kind of experimental data to tell them how to adapt. So if we only did the first two patterns, it would sort of be like the science of astronomy or astrophysics. We can't actually move the stars from one place to the other. We can't stick two black holes close to each other and watch what goes on, right? We can't do that. We have to observe what's going on and we have to hope that the variability or the variance of the universe is going to tell us something about how stars behave and how black holes behave and all that stuff. But conveniently, when you're in an organization, you can speed up your ability to learn. And the way you do that is you proactively experiment. You say, you know, every sprint, we're going to try something new and we're going to create a hypothesis around it to say what do we think will happen? So how many people use a definition of ready in their strumtons? How many people have a definition of ready? Okay. If we beef up your definition of ready, what do you think will happen to your velocity? Let's say you're not worrying about dependencies. You're taking anything in and then you add something to your definition of ready where now you're not going to take in work that has dependencies on another team. What will happen to your velocity? Say it louder, please. Come down. If you're ready, criteria is stronger. So if you say we're not going to bring in, so let me make sure. The ready criteria determines whether you're going to take something into your sprint backlog. So you have a big backlog with 100 things in it. They're all estimated. And so now what you're doing is you're saying, okay, we're not going to take that thing in because it depends on some other team. But we are going to take this next thing in because it's something we can do without depending on another team. We're going to take this thing, and we're going to take it in something else that's not dependent on another team. Velocity is the sum of the number, the story points that can be completed by the team in a sprint, right? So I'm not sure this crowd passed that test, but I'll tell you what happens. Your velocity will go up, okay? Your velocity will go up because what happens is a lot of teams take stuff into their sprint, and they don't think about the fact that they need something from someone else that isn't done yet. And so when they go out and they say, hey, guys, we need that thing. You told us what was done, and the guys sometimes say, okay, yeah, here it is. Great. But that's not what often happens. Often you go out and say, hey, we need that thing. And they go, what thing? Right? And you go, hey, you know the thing we told you about three months ago that we said we were going to really need? And they go, oh, I forgot. Sorry. And so now your velocity drops because you are depending on another team. That's actually an important... I'm glad we had that little quiz because it's important to something that's about to come up hopefully if I can keep my time together. So I'll just talk through the rest of the slide until it appears as it just did. So when we experiment, we actually create hypotheses. So that's what I recommend that people do in retrospectives. I recommend that they have these leading indicators like velocity and quality and happiness and that when they make a change, when they say, hey, let's try something new, they make sure that whatever they're going to try is going to affect one of the metrics that they think is important. If it doesn't, if it's not going to affect one of the metrics, what are some things we might do? If someone proposes a change that doesn't affect the metrics that we think matter, what might we do? Get rid of the team? You might adopt a change, you might adopt a change, but it's not going to affect the metrics, right? So how would we know whether it was a success? Say again? But don't we have to measure something to have an experiment and the metrics that we thought were important are not in the... That's true. We might say, hey, let's try this random thing. Let's change the length of our sprint. We go from two weeks to three weeks or something. And we don't have any idea. We might just want to experiment just for the information's sake. Fair enough. And that you can do occasionally. I would suggest, though, that most of your experiments should be motivated so that you have a theory, right? So many experiments are open-ended. Hey, we just want to figure out what's going on. And that's that kind of experiment. And then other kinds of experiments are, hey, we want to find the Higgs boson and we think it has the energy at this level. And so let's construct an experiment that either confirms our theory is true or invalidates it, right? Like, either way, we get some information, but we're going to try for something, right? So that's like the ready criteria thing. Hey, our velocity is kind of crappy. Maybe we could try an experiment. I heard that improving your ready criteria can increase your velocity. So then they say, oh, what's our theory? Our theory is we're going to try this, right? We're going to try keeping up, right? You know all you know. Anyway, so taking action as an agile manager, there's a couple of things that come out of that. One is that we don't learn anything if we always succeed. Well, we might learn a little bit, but it's very little, right? So sometimes when we succeed, like we might have a theory we're going to make a profit. Well, sometimes we make this much profit and sometimes we make this much profit, right? And so if we're making this much profit, well, if we can't tell the difference, right? If it's always success, like, you know, we made our profit and then someone actually comes and asks you, well, how much should we make? Well, 0.1%, right? Well, that's not a very good outcome. Or it's 5,000%. That's a really great outcome. So what you want to do is say, I want to understand the situation well enough that I'm going to hypothesize something that I think is going to have a 50% likelihood of failure. And the reason that you do that is because you think more deeply about the problem. So you might say, I think our profitability is going to be 18%. And someone says, OK, that's an interesting hypothesis. And then you run it and it's 15%. And you might say, hmm, that's really interesting. I wonder why it was less than I thought. That's how scientists work and that's how we can work and more rapidly improve. So by doing these kinds of 50% low risk failure thing, we actually start reducing the anxiety of failure in the organization. If you want culture to happen, you have to do things that ease the culture in. And this is one of those little tricks that does that. So if we start experimenting and we're experimenting with strategy, so if we're doing strategic activities that are bringing people in and having them work together to achieve strategies, we get a bunch of stuff. I put this here and not all of these things are guaranteed to improve. What I'm saying is, these are things you can experiment with as an agile manager. You can say, you know what, we need faster and better hiring. I had to deal with that. I'm going like, God, I hate hiring. It's just such a pain. And then I realized that I could stick it all in a Kanban board and then I could actually get my team to do most of the work. And then suddenly we were able to turn through lots of resumes and lots of people and we started getting better people. We were able to get them offers earlier and that was really nice and I didn't have to work that hard and that was really good for me. So that's good. So this is kind of a slide talking about scrum and retrospectives. And I just wanted to show you this little unusual thing that I like to do and I like to put it as the first meeting because it is the first meeting. It is the meeting that sets up the experiment you're about to run. So there's two little pieces of paper here. One is this list of data. That's the data from the past. What was our velocity last time or the previous three friends or whatever. This is the quality and so forth. And now let's craft an experiment. When we craft that experiment, let's run it like a real experiment. What's our scrum master for in our experiment? Anybody know? If they have an idea, why do we have scrum master if we're running experiments? They do facilitate this, right? Here they have a particular role. This is an interesting challenge, right? Like they have one role here and they have a different role here. And so the role here is they do facilitate the creation of a great experiment. So they do want everyone to participate and get involved in it. But when they finish this retrospective, what could they be asked to do? Anybody have any ideas? So in a retrospective, they facilitate the team coming up with an experiment. And if you're a scientist, let's say you're a biology scientist or something like that, and you've got your bio team together and you've come up with this experiment, well, now you're going to run the experiment, right? What are you worried about when you run the experiment if you're a biology scientist? It might be better, it might be worse. Here's what biology researchers worry about. Contamination, right? They've got all this great stuff going on. They have to control the experiment. They have to make sure that the experimental conditions are so well controlled that they can get data out of the experiment at the end. And that is what good scrum masters do after a retrospective. They become something that sometimes people get a little bit horrified when I say this. They become police officers, right? So what happens is you say in your retrospective, you run the experiment, and by the way, Sanjeev, we, you the scrum master, we authorize you to help us enforce these rules. So if we are going to try pair programming, we'd like you to let us know that we're not pair programming and could you please get us to pair program. And if we're not able to do that, we probably ought to get together in a meeting with the whole group because we agreed to pair programming and we're not doing it, right? So this is what a good scrum master would do. They would say, oh, you're not programming, pair programming. And we said we were going to do that. Could you do it? And then the person says, okay. And then they do it, right? That's great. The other outcome is, no, I won't do it. It's stupid. And then Sanjeev says, well, why did you agree to do it in the retrospective? Because I didn't want to make people mad at me. And then Sanjeev says, well, we've got to get together for a meeting and figure out what to do. So we run that. So here is a thing I experimented with. So how many people know what MBO means? Management by objective. Demings. Kind of that statistician that hung out with me for a while. And for a while. He said MBOs are the most horrible thing in the universe. Right? So it's a set of goals that you're going to try to achieve in the coming year or the coming quarter. And they're supposed to be smart and smart is S-M-A-R-T and I forget what it all means. But anyway, at Citrix we had MBOs. And I said, wow, I kind of hate doing these MBOs at the beginning of the year. And then by the end of the year, I'm a totally different person and my responsibilities are very different and I've learned a lot. And then I'm evaluated at my achievement of these MBOs. It just doesn't make any sense to me. So I said, okay, so let's run an experiment with a different style. Let's do user story goals. So I said every quarter and you can see, look, they look a lot like user stories. As a product owner, so my staff was, in this case, was the user experience department. So I had the user experience department at Citrix. And I said, hey you guys, we have this problem that there's not enough of us even though there's 24 of us in this little company, user experience was really, really important at Citrix. And so everybody wanted time from us. So we didn't have enough capacity to satisfy the demand. So let's try to fix that. So I can always get reasonable user experiences. So users don't stumble using our products. And so Citrix focuses the great UX talent on key work. And there's more to it really because the acceptance tests are this. We wanted to teach, I was the product owner effectively and I'm not this product owner, but I was the product owner of the strategy scrum team that did this. I said, you know, I want us to teach every engineer how to do at least okay user experience design. And I know that sounds like a really hard challenge and it was probably, we didn't achieve all these acceptance tests, but I, but these were MBOs for different people in my team. I would meet with them once a month and we would adjust these acceptance tests and I was always very generous about it, especially if they were sincere about trying to achieve the general goal. And what I thought was really funny was at the very beginning when I had these user stories written by UX people, they wrote these user stories like this. As a UX staff member, I can go to the Agile UX conference so I learn more about Agile UX and so Citrix has people who know more about UX. Right? And you know, if you're a good product owner, the stakeholder here is not a person on your team. Right? It's never as a developer I, you know, have a better API so I can write better code and so my company has better code. Right? Like that's just a meaningless user story. So the user story I was talking about has an external. So I had to teach my UX staff that we were here for a purpose. We were serving the company. Right? And so someone in the company is benefiting from our work. So I had to teach them that what happened was they became much better employees but much better contributors to the organization. And so these were a couple outcomes. One designer trained UX for dozens of engineers. That was really awesome. Another UX designer it wasn't this user story but he invented a different user story which was the UX department should all know how to do reasonable JavaScript. So they could actually design UX using jQuery I think was what we had them use. And then my happy part was one UX designer wrote me great reviews saying I was a great boss that I had really open designs about how to contribute and how to think about his role and he's now I think he's a senior director at Amazon school. I'm not going to ask you to do this so so I've described three patterns so far leading indicators limit work and process proactively experiment if you do those three things you're going to be agile. What I mean by agile is you're going to be able to adapt the things coming in on a rapid basis and you're going to be able to produce value and that's really awesome. But one of the problems with that is those three things are all three are hard finding leading indicators that's hard limiting work and process you make people mad because you say hey you know I'm not going to work on your thing because I have to focus on this other thing and then experimenting creates this kind of trend of failure success failure success and if you have really jerks in your organization they're going to say hey you have a lot of failures in the group even though you're the most productive group in the whole company. It's sort of like I'm from the United States and Obama is our president and he's been very successful at hunting down and killing terrorists. So we have this weird thing in our country where we have a lot of people talking bad about our president and they never acknowledge the great stuff that he has done but they're always picking on this little stuff. Sorry for that distraction. Let's talk about the fourth pattern. So the fourth pattern is something that helps us keep agility. So we've got those three patterns and we can be agile there but part of the problem is the workload may shift and when the workload shifts we might not have the capacity to handle that workload. So for example if you're on a scrum team and you've got five developers and one tester and then you start implementing stuff that requires a lot of testing not very much development but a lot of testing. If people have their role you're going to get situations where you can't finish the job because you don't have enough testers and so what you want to do is you want people to shift their thinking so they work on whatever is necessary to get the job done. And so a guy named Christopher Avery came up with this thing called the responsibility process and I'm going to encourage you to read about it. I'm not going to talk about it here but I will talk about the solution that emerges from it. So this idea that when we join a team we are collectively responsible means that when we join a team we essentially sign a contract with the team where we say I will be personally responsible for the collective output of this team. So now if you've got a testing load that's coming into this team and everyone on the team feels personally responsible for the collective output of the team the team members are going to say well we're going to if we don't shift what types of work we're doing we're going to have to ship crap right because we don't have enough testers to handle this thing. So I guess we could howl at the managers to say we need more testers that would be a responsible thing to do but if the managers say hey you know you don't get any we don't have any budget for that and the team will have to say well what are we going to do I guess maybe we developers should test a little bit or a lot right and so now what's nice about that is you get this team that's more elastic with the work load that's coming in and the team starts wanting to learn more so that they can satisfy this collective responsibility idea if you're an agile manager you want to radiate collective responsibility and that means two things first of all that you yourself should act personally responsible for the collective output of the department and the company so that means you start worrying about more than just your little department or your little group and then what you also do is you start saying hey you know I was assigned to do this retention project with HR but I'm going to share that strategic goal with some colleagues because we should all be collectively responsible for the retention of our employees you teach Avery's Christopher or Avery's responsibility class and then you create these strategic strums teams they just sort of happen if you have this sort of collective responsibility idea you start forming these teams in virtually every situation where you have kind of a lot of creative stuff to do anyway a lot more gets done and there's a funny thing that starts to happen if you create these strategy scrum teams people want to hang out with you I mean people like I've had strategy scrum teams for agile coaching and I rope in every scrum after I can find and all the agile coaches in the company whether they're external or internal and then you know what happens like people from other departments kind of float over and go wow what are you guys doing over here I go yeah you know we meet every week and we're working on these strategic things that someone says well you know I got some extra time I want to hang out then they join the team they're not even my employees and so and the reason they join is because they develop new skills these things are really great for creating a lot of happy eager employees the other cool thing is you get some free time so if you're in the boss here you're the product manager of this strategic scrum team the cool thing is you're basically delegating your normal work your strategic work about improving retention or whatever you're doing there's a bunch of people who will help you and you can use that time you can use that freed up time to do some more sophisticated things and that would be our fifth pattern so I mentioned this every single scrum coaching group I've never done has been an agile coaching you know I didn't list it here but the UX department was also organized as three strategic scrum teams anyway these guys all these people and I mentioned all of that and it was really awesome and I got a lot of papers out of it actually because we started writing about metrics how to scale scrum and stuff it was fun so we're going to talk now about the fifth pattern this is the last pattern I'll talk about external factors limit our agility so remember I talked about changing your ready criteria so that you had a stricter definition said we're not going to bring in stuff where we depend on other teams right and that increased the velocity of the team well that does work but it's not practical in many organizations right I mean you can do it for some stuff but you start doing it and then you realize systems start seizing up a little bit because other teams are slower than you want and then you can't do something because some teams need to bring it to you and they're not a scrum team they have like six months delivery cycles or whatever it is and then you're desperate to get on their list and you're changing what you want to deliver to your customer every sprint right and so you changed what you wanted and now you need your dependencies to change what they're delivering and they're on their six month cycle and they go hey leave me alone man you know I'm not agile so they mess us up so they make us slow they make us test more you know they're not agile so they don't have automated tests and they hand us stuff that has got bugs galore in it and then we have to boot it back to them and look there's a whole bunch of delays that happen there they mess and so what this happened to Toyota right so Toyota was a pioneer here and they said we want to be just in time manufacturer and so you order a car and we're going to start building all the pieces for the car and so they said oh chassis you know we know how to do that the body we know how to do that now we need the tires oh we don't make tires let's go to the tire manufacturer you know we need four tires and the tire manufacturer said hey you know sorry you're supposed to order that like 90 days ago we have a product cycle here we're busy building inventory for snow tires and you want some other kind of pair and so they said okay we got to do something they went out and taught their suppliers how to be just in time manufacturers and by doing that they actually created this great infrastructure in Japan that transformed Japan from in the 1940s it was producing poor quality stuff that all of us well not all of us but at least in the US they had a reputation of producing crap right my grandmother was always talked about Japanese crap cars right but then you know look what happened like they became the best quality most profitable car manufacturer part of the day anyway in the world and around the same time GM the US homegrown thing it was all about you know mass manufacturing went bankrupt so that tells you what happens when you start thinking this way there's some other people that we that limit our agility too so this is one side the people that we're trying to get stuff from to build our product but then on the other side when we're delivering stuff we want to get feedback from our customers and sometimes they're slow so if they're not a dependency really they're sort of we need feedback I guess we're dependent on them for feedback and if they're slow we don't get it and that whole thing with you know Richard Sheridan saying hey you know we love it when our customers love our product and want more stuff well sometimes our customers are not very fast and it causes problems for us and worse you know it can it increases the competitive threat because we don't test the market faster so this fifth pattern is starting to is saying move out of your team and start thinking big right so now we're saying hey our agility is limited by the environment we're in and so let's fix the environment so now you start worrying about converting bosses to be agile you start having this conversation with them that said hey you know Dan Greening did these strategic scrum teams and it was very successful for motivating his folks let's try that right or you know there's all sorts of let's have some leading indicators right for you guys in the finance department same thing you guys in the finance department really ought to start thinking about these leading indicators and ought to be a helping us be more agile rather than keeping us less agile alright so now we start seeing these sorts of things so leaned Kanban is a different strategy that can be used for operational stuff we see theory of constraints who knows what theory of constraints is okay that's good this is a systems thinking thing and so sometimes when we go to scrum courses we'll teach theory of constraints because it helps explain some of the stuff in scrum but it's actually this bigger picture thing where we start saying here's how we spread agile to the rest of our world I did some dependency mapping A3's are how you solve problems and big organizations the problem with this pattern is executing well actually the fourth pattern and much more so this pattern start executing at exercising skills that we didn't have to do in the first three so the skills we have to deal with here are holistic thinking kind of new right actually have to think about the ecosystem around us and the second thing we have to think about is diplomacy you start doing this stuff and it can be a little bit risky right so you have to be very careful about how you talk to people you have to quietly work with them you have to get their permission to talk to their departments and things like that otherwise you can find yourself on the street but nevertheless this type of skill is very sophisticated skill is characteristic of great agile managers so there we go lots of great outcomes let's see well this is a systems thinking problem I think it's going to be hard to explain but I'll do my best each of these lines is a level in the hierarchy of Skype building a product and you can see there's like 40 teams well there really some of them have three scrunchings in them but there are 40 modules right so what's happening is to get something from here of getting so to get some can you push the button and say stop so if we make a change in this team it has to propagate to this team to this team to this team etc the change until finally it's delivered at the end and so that is a limiting factor on the agility of this system this is systems thinking right now so what could happen so if you change this you have to wait until the sprint is over now it ships it's thing to the next team and so forth when you have these things up you actually get about 27 weeks worth of time that's required to propagate a change at this lower level all the way to end users who are using this client over here does that make sense what I'm saying so so what happens is if you look at this system there's going to be some interesting dependencies here that if you could squish them the world would suddenly be just a little bit faster right so this was a case where this team depended on only this team and then what we figured out we could do is we could merge these two teams together and create what are called um feature teams so we ended up creating I think two or three feature teams that all ran in parallel on a shared code base and we saved something like five weeks out of that 27 weeks of delivery time so this is the kind of thing that you start doing you start thinking bigger outside your world um this is kind of a comparison of the agile base patterns which are actually those base patterns that I was talking about agile managers are actually characteristic of every agile team thing whether it's an agile organization an agile team or an agile person they have these characteristics and so I paired them up with Gallup's assessment of managers and what I found was that these patterns matched up pretty well with high current managers but they were more aggressive so agile is more aggressive so measuring leading indicators to improve is more aggressive than challenging ourselves and teams to improve right like we're actually saying a little bit more it has to be lined up with the mission and the limiting work in process is a more aggressive statement than just creating structure and processes to help teams deliver on expectations here we're just getting a little faster here so that's kind of cool you know I talked a little bit about how organizations lose agility so they basically stop doing one or more of these patterns so they stop measuring so those people who do pound no estimates and stuff they all have this sophisticated discussion about basically if you're not measuring what are you measuring some other crap but the bottom line is if you're not measuring anything do after reading the propaganda from the you know pound no estimates people we can't tell if we're improving right so now none of this experimentation is useful suddenly it's hard to tell if we're limiting work in process a lot of issues with that so when we stop measuring that's not good also if we have lazy teams so that I think sometimes you have teams that don't want to push themselves to have a higher velocity and then they I think of that as sort of phoning it sort of you're not trying to improve stretch goals that managers okay bad managers non-agile managers come in and every sprint or every quarter will say hey that's a great goal you've got there but let's make it a stretch goal right and then they give you something that's 50% harder virtually impossible to achieve so the first quarter they do that you go wow okay I'll try and then you bust the gut and you don't make it and then the next quarter you go okay I'll try and then you don't make it and the next quarter you go like screw this man like there's no way I'm gonna try so these stretch goals are just horrible and they interfere with this automation if you don't automate testing you're basically gonna lose on agile I don't know I mean I can have that distinction if you if you're interested in why that is true I will show you mathematically why that is true if you do not automate manual testing you cannot be agile period so that's interesting and then all these other things alright 10 more minutes okay I'm gonna wrap it up instantly so I'm not gonna talk about that here's some references that are kind of interesting this work there's actually all of these I mean these papers are mine right so but they have references in them I'm kind of an academic or a secret academic or something anyway I have a PhD and that has made me write all these detailed references so if you go to any of these papers you'll see references to other staff including these blog posts but then I also want to talk a little bit about one of my favorite books ever in the entire universe is a thing by Christopher Alexander called a pattern language and I am just telling you this because I'm just telling you this because I love the book not because it's directly applicable to agile it is about how to design cities, towns and well that's it cities and towns and countryside how to design it's urban planning but it is so beautiful I'm not an urban planner it is such a beautiful book you just can open it to any page and you can read something really inspiring about how to build a house so when I wrote these things especially this one it was modeled after this pattern language idea and we've seen pattern languages in computer science as well so we've had a lot of people in computer science who have done it anyway there you go that's me so I would like it if you would provide feedback on this talk in conference engine I am a feedback driven guy as you can probably tell you can also send it to me in email here are things that might not be in conference engine so if you feel like sending it in email do it in email would you recommend this talk to a colleague or a friend and then you give it a score from 0 to 10 does anybody know what this is called that survey question what is that promoter score very good so this is the simplest possible survey you can ever do I'm not going to tell you how it's scored just if you feel like doing it send it to me or you could even I suppose if you're really bold you can twitter it right so and then the best thing the best second question is what would improve your score so if I did something differently would you have liked the talk that would help me improve these two things are great leading indicators for the success of your product pretty cool huh so now open question open question yeah both this is applicable to all levels more diplomacy do you have a mic make sure it's on yes so the question was initially when you adopt agile the team is enthusiastic about stuff and so it's more productive and then you create expectations in the organization about that productivity being overly high I think that is true first of all but the second thing that's true is when you start a green field project that has no functionality in it and you're building functionality the first few sprints will require virtually no testing and this is going to the thing that I said I would talk about which is adding testing capability or automated testing so what happens is if you don't have any automated testing you actually have to do regression testing at every sprint which is very new right like it's very different from waterfall and then if you don't automate it what ends up happening is you're not capable about midstream into a project that's maybe four months in you're not able to ship a tested product if you don't automate and so this is another factor that causes lots of problems with executives so we have to basically have some serious discussions with executives in fact what I now do when I talk about agile is I tell people there is a cost to it there is a serious cost to it and it is a cost in delivery right so we have to ship it 18 times instead of once every 18 months and so all those things are multiplied and they're going to have a cost factor that will become more and more apparent the benefit is that you can adapt faster if you don't think you need to adapt don't do agile right like you don't need it because it's so costly and then I always try to prep them and say hey you know what there is a cost so let's just make sure that we know we need to automate testing and we have to invest in that and that is actually something your team can start doing right from the get go right but you're right it's a difficult conversation have a few more minutes yeah yes how can we best utilize agile how can we utilize agile well the best thing is we have these patterns right like limiting work and process and stuff like that and so you will have challenges you know like this is a hard thing what I like to do is I like to get the whole team together and say how are we going to change what our manual testers do over time right as a scrum team that should be our collective responsibility we should help we should help testers manual testers become more scripting people we should teach them how to do exploratory testing so instead of doing manual regression testing in every sprint let's start doing like more sophisticated exploratory testing that hasn't been done before or one thing that they can do which is really great when a customer reports a bug what would be awesome is if there was an automated test that would replicate what the customer did and so now the manual testers can be part of that automation team it's actually you know using selenium for example if you're doing a web based application selenium is a relatively easy tool it doesn't require significant programming skills so you don't have to be you could be a tester and still have a lot of great fun but I say let the team help the team understand the problem and the transition and have the team love all of its members and help their members careers improve that's solid good yes cucumber and selenium work together really well yeah it's all awesome and you know what happens a lot of testers it becomes from masters and coaches and that can be good too a lot of options anything else yeah say again oh collective responsibility yeah right this is a big challenge right like this is a cultural shift so I I totally get what you're saying I didn't say it would be easy but but it is a pattern that we see an agile manager that they have this characteristic and the same 14 so I one of the things that I like to do how much time do I have I have zero time otherwise I would tell you all about Christopher Avery's methods for dealing with I'm happy to talk after I think we're out of time and I really want to thank all of you for coming and staying and having a great time