Transcript
Vass: This is about helping software teams improve. We’re going to start with a little bit of history. The Co-op that I work at today has its roots in 1844 when the Rochdale Pioneers got together to try a different way of doing business. There was lots of different co-operatives followed from them and merged into one bigger one. At the time it’s the kind of stereotype you can imagine of the rich landowning classes oppressing the working people. There was lots of low wages, hard to get high quality food, lots of like flour cut with things and price gouging going on. These people got together, you would put a pound in for membership and try and have a different way of doing business. Those rich landowning types hated it. Over time they tried to improve. As well as being a successful business, the co-operative ideals are about improving what’s going on in society. Some of the reasons we’ve got not just food stores today comes from this history.
Big challenges in those days were funeral poverty where you hadn’t planned for that, and you’ve got that expense, or getting into legal troubles. Starting businesses to help with those in the olden days is the reason you’ve got Co-ops confusing a collection of businesses nowadays, as it does funeral care, it does insurance, it does legal services, and a whole collection of other things. Some of the stories I really like from those days about how they were at odds with the traditional model of doing business there. Lots of industry said, you’re not allowed to give a discount, we have price fixed this. If you want to sell medicines, if you’re doing aspirin or anything, you can’t make it any cheaper. They wanted to do member pricing, which is still a thing today, so they started making their own.
Similarly, they couldn’t offer electronics at a discount, so they opened factories started making their own radios under the brand name Defiant. When I joined the Co-op, the building we worked in, was one of the traditional Co-op warehouses from the 19th century. Our meeting room, one of them was called Defiant. The history is all around you. I like that. Where it came from was about rebelliousness, about not going along with the status quo, about changing things. I think some of that’s still there today.
There’s a quote from Parkinson, you know about Parkinson’s Law that work expands to fill the time available. He’s also done work on the growth of bureaucracy. He’s got a formula about how it accretes about 5% to 7% per year for no other reason. The Co-op is coming up for 180 years old, and it’s got 65,000 employees. There’s been a lot of time. Some of that’s necessary, just keeping track of what’s going on and understanding what’s happening. There’s a lot of bureaucratic processes.
If you’ve worked in any big organization, you absolutely know the frustrations of, how do we keep on top of this? Who’s changed that now? What’s the priority? What’s this team doing? In addition to that new and different and rebellious side, and try to do good for society that I love, there’s all these traditional big company balances. I do think when I find myself getting frustrated with, they don’t know what they’re doing, there’s lots of software and agile teams to do this. If only they did something different, you’d be better. I think, what are the chances that I could have a business almost two centuries old and keep it still successful? Nothing. What experience do I have of running somewhere that’s got multiple billion pounds turnover every year? Absolutely none. It puts me in my place. They’re obviously doing something right. I don’t know everything as I walk about trying to tell people how to change their jobs.
We’re almost coming back to the present day. I will talk about software. Let’s just use a timeline, zoom us with a tech focus up to modern times. Here, 1844, that’s those Rochdale Pioneers, and the first Co-op starting. It wasn’t a software company then. If you think about, traditionally, where do you think about when you think of tech companies? It’s Silicon Valley, is the famous one. A few years after the Co-op started, this is the state of California getting founded. We predate Silicon Valley by a fair bit.
Zooming on to 1911, I think, you can guess our next tech milestone. I’ve just popped on, this is when IBM got founded. That’s International Business Machines, because computer wouldn’t be a word people really used for a long time after that. That’s IBM’s punch card business. Zooming on again to 1991, this is CERN, the physics lab where the world’s first ever website got made in 1991. Thanks to the Wayback Machine and some ICANN digging, I think just 1996, just a couple of years after the website was a thing at all, there’s the Co-op’s first website. Our last stop, very important landmark, is 2018, where I joined Co-op Digital. Looking at the length of the history and trying to understand what’s going on was difficult, but I think I’ve actually been at the Co-op for 3% of its existence, which is more than I thought. It’s not nothing.
Scope
This talk is going to give you a tour of modern-day Co-op. What are the software teams? I’ll show around what are the ways of working you see there, and how have I tried to do my best to give advice and help and support where I can. You may get ideas on how a big company with such varied types of business gets things done. You might be able to steal some ideas about how you organize or how you use your own things.
How It Started
How it started. It was 2018 when I joined the Co-op. This is not how the Co-op started. When I joined, Co-op Digital was a fairly new thing. Co-op has such varied landscapes. Just in food, there was different parts of tech and different parts of the big food business. They were almost not talking to or aware of what some of the other businesses did. There’s agile. There’s cloud development. There are other things that have been going on for a while. By and large, lots of parts of the Co-op, their experience with tech was do a big RFP, get some vendor like IBM or one of those other big companies to come and make something. We don’t really get involved, and you specify for years in advance what it’s going to be and it gets ready later. In about 2015, Co-op felt that it was interesting trying something different. A lot of the team who set up gov.uk and the government digital service came to the Co-op to try and install some of the ways of working there and show us what went on. That’s when we moved. Co-op Digital was born, a little bubble with some teams to try out something different that hadn’t really been seen at the Co-op much before.
If you’re familiar with GDS, you will have recognized some of these things before. Some of the things they introduced was like service design, working in the open, doing discovery work, to like, what is the problem we’re trying to solve, and let’s slow ourselves down before we jump to giving out a contract and build something two years from now. They had teams working closely with the business. Something I really enjoyed was some of the teams that got someone when we were redoing how admin and funeral care works. I think it was the first change for the funeral industry in 400 years, to get rid of lots of paperwork and lots of overhead. They had funeral directors who come and sit with the digital team to explain at length like, we can’t actually do that. Do you know what I’m doing when this comes up? To go back out to the business and say they’re solving real problems. Here’s a new version. It was good. It was fun.
The other thing that got introduced from GDS was the different phases. The discovery, alpha, beta, live process, as, do not jump straight, to, I think we should have an app. Can I get millions of pounds of budget and go and launch one? You should start way over here, as, are you sure there’s a problem, and how can we define it? Who should we get involved in doing that? At one point in Co-op stores, if you worked in a Co-op shop, the only way to find out what shift you’re doing is to look for a bit of paper in the back room of the shop. That wasn’t too many years ago. This was a huge step forward as we brought out shifts and understanding what people wanted, what would be useful for them for the business and the colleagues, was really good. In your discovery phase, you work out what are the challenges, what do people do.
The business and colleagues would love to be able to do things like mark yourself, “I’m happy to work in any stores in a particular area. Just shout if something comes up”. That’s a huge unlock for stores flexibility, and the colleagues love it, which just wasn’t something that was possible before. You put them on a backlog, and from your alpha, you don’t jump straight to, so we should make this then. Your alpha, you’ve got lots of options. You try things out ideally in paper or throw away prototypes until you’re sure you’ve got it right. That’s the process that got introduced quite a lot in different bits of work we did.
One of the first teams I worked on was Co-op’s One Web team, which is still a team today. Thanks to the Wayback Machine, I’ve dug in and I found out that first one website we made in 1996 optimized for Netscape 3.0. There had been a lot going on since then. For quite a while, I think anyone in any part of the Co-op who had the budget could go and pay any agency to go and make them a website, and we didn’t quite know how many websites there were. It was a lot. It was far too many. I don’t know what it looked like when you were as a user Googling Co-op and just finding all these different top-level domains. I’d just moved from the BBC when I joined here and they had iPlayer and Sport and News, and they had one website.
I think at most Co-op probably needs one, was what we went for. We started. We started moving over the recipes site. Co-op’s food website looked like one but it was actually I think four or five different top-level domains that would change as you did it, but skinned up to look the same. It was different contracts with different agencies. Moving the recipes website over was good because we got someone from food marketing to come and sit with us and understand the problem, the mess of updating it, and things going on. It looked quite nice. For lots of the things I just skim over here, if you do want to hear about them you can go and have a look on Co-op Digital blog because loads of these stories are still there worked in the open.
One reason that I was so pleased with the one Co-op look of it was the design system which has now grown into the experience library. This was one place where the designers and other roles had got together and said, things look different everywhere you go. Things are worded differently wherever you go. Can we agree, as a group, what’s a sensible way to do that? The approach they took that I really like as with those different businesses, people were understandably quite worried. They’ve had a lot of flexibility before in the business, “I want this to look amazing. I want this to look different”. Is it all going to be the same? Is somebody looking for Easter food offers going to see a similar thing to like legal services forums? Familiar not consistent is what they talked about doing, like recognizing the variety of what’s going on here. Make the simple things repeatable and do them again. There’s web components and stuff you can build there.
The other thing I really liked from this was the content guidelines. There was a quote, it’s probably still on there from the head of content design, about everything you write on your site makes it harder to find everything else. I think that’s really stuck with me. I think there’s an instinct sometimes when you’ve got a website, look how much amazing stuff we’ve got? Getting people from the team to sit with us as we do content audits or I printed stuff out and put it on the wall, they just look at, that’s too much, isn’t it? I think as well as the technical side of things, we’re doing things in containers, so that you had full control over your own website without affecting other people’s, but you could benefit from any new features like scheduled publishing and stuff like that. That was a good model and it was quite an interesting technical challenge to work on, as much as it was a culture shift internally.
This idea that you show me three designs and I’ll pick one, just shouldn’t go because we’re looking for a big picture. How does a Co-op make sense? I like more words, can I get more words, is a difficult thing to talk about what you’re doing.
It went all right. We started taking down and moving things over. Over time, it started to get complicated. There were some cases where the team just needed the ability to change pages. The wine team at the Co-op, at one point their contract was, you have to keep going back and paying to make changes. They didn’t need much technical support, apart from, we’ll give you a wine site and you can log into the CMS here and you can add a new page anytime you like. They were delighted. Other places have more complex ongoing needs. The web team grew a bit and it evolved into, there’s some sites we look after and change because there’s things that need changing. There’s things that don’t have a digital team, but they do need things updated. Maybe we can get some skills in there. Then for other digital teams, they were taking on what we’d done. I was starting to get confused because I was thinking it feels like we’re more like a platform for other people to use.
At the time, I’d never really worked in that team before. I was wondering about setting objectives, about, how do we measure ourselves and what direction do we go in? It felt like an anti-pattern. You know in the old days where there’s a database team and there’s some other component team, and there’s somebody who does the frontend. Shouldn’t we care about the whole end-to-end stack? Don’t we need to be closer to the users? I wasn’t sure.
Then I saw on Twitter this advert for something called team topologies. These slides I was looking at was, these are the slides I want. Look, it’s got metrics. It’s got things we need to think about, adoption costs. That was my first encounter with team topologies. I got to go on a two-day course with this where I could ask specifically, it was Manuel Pais, when you say things like don’t make it mandatory, do you actually mean not mandatory? I can absolutely see something like that. Give people a choice, like, if you do it like a product then in other teams, it forces you to make it good. I can see that in theory. Who really thinks about the big picture in a complicated organization? It’s often easier for some other team to say, I don’t know Python, so I’d rather write my own thing from scratch in React over here. Don’t we need to? It was really good to speak to them about crossing the chasm about they might have a point about that, and 100 other things that were super useful.
Another idea that we got from this as a team was about, you shouldn’t start off as a platform team thinking about, what’s the fastest we can go? How do you get those DORA metrics up? How can we deploy more stuff? Because you can just get yourself very quickly into a mess. The hard problem is about interaction modes. This is something that really stuck with me. Lots of us haven’t had to practice that at all. As you work with other teams, collaborate. Just get everyone together and work it out. It’s fabulous in lots of cases just because it’s not done enough. If you do that for everything, it’s super wasteful.
Before you know it, you are spending your life, especially as we were supporting more teams around the place, collaborating with them all the time. Thinking about interaction was really important. What do we want to offer as a standard service? You ask this, you get that. Who are we supporting for a limited amount of time? The idea of putting a deadline on it, how many teams are we supporting now with what, and when do we want that to end? You can move that deadline if you want. Bearing in mind that we expect this to stop at some point, prompted some really good conversations. Some other digital teams felt like for that platform, for the AWS side of things, that’s always going to be you doing that for us? We shout when something needs changed. They’re like, no, the intention is, we’re training you to do it, and we’re going to stop soon. That prompts a different conversation that I just hadn’t realized needed to be had. We got all kinds of value out of that. Working on that team day to day as it grew and changed over time, I’ve worked at a few places, I’ve seen a lot at the Co-op, I feel like I’m helping here.
Improving Things
I moved on. I moved to a role where I was managing people on teams rather than non-teams directly. I’ve got ambition to help. I thought, I can improve things. I talked earlier about not getting too big for your boots, thinking you’ve got the right answers everywhere. What I saw was lots of teams doing quite well, but then falling into the same problems and pitfalls that I thought we’d solved some time ago. I talked earlier about having an on-site customer, like who from the part of the business you’re working with is with your team day to day. That’s a practice that just fell away. There was a few different things about prioritizing work. How do you choose what you’re doing? There was teams I talked to who were struggling with, we’re not sure about our agile process, like how we put work on the board and stuff. I said, do you talk about it in your retros? Says, we don’t have retros. I assumed that’s the trouble. We’d never written it down. There was wide freedom.
That idea of just get qualified people together and let them work it out, sounds great, but then it’s doing people a disservice. I think how we want to work here is a question that doesn’t need answered. At different companies, you do something different. There are probably hundreds of different ways you could do agile well. Helping people with that. Also, for lots of people, just learning how to program or learning how to design can be a pretty much full-time job in your other part of your career. You look up, there’s a Kanban board and sometimes you have a retro. You’ve never really talked about why are they there? How do we do that well? That was something that was getting missed. I thought, and we talked to other senior people around teams in Co-op Digital, like we could improve things. We’re doing well, but we could get a lot better.
It’s important to hesitate before you steam in and tell people how to work. I think you can tell people anything and lay down standards, but you can’t make them find it useful. We’re looking for what’s a challenge that you’d like some help with, and do I have anything that helps you? I think it’s difficult sometimes when you’re working out what practices would help people. Sometimes I think you have to adapt it to your context. I’m not really going to be on your team for a while. I sometimes feel like the advice I give people is, read these books, watch these talks, and then try it out for a year or two and you’ll work it out. That’s not as helpful as I’d like.
The other thing is giving people bad advice. There’s a good quote from Ron Jeffries, talks about what it’s like when you’re a bit experienced and you’re coaching other teams, and you try to give them advice. You’ve got to be super aware of that thinking. Lots of coaches think, we did X, so you should do X. That doesn’t imply that X helped you at all. That may have just been something going on in your team. Actually, it does not imply X is the best choice. There’s loads of things people could do. They could happen across something much smarter than what you would have picked. It really does not imply that X is something you should do. I try to keep this in mind. On this, how do we help people? I think there’s whole talks involved in all of those things. Does anybody want your help? Do you have good advice for them? The topic for today’s talk is specifically about making change stick. What I found again, was people did want some help. We did coach them or talk them through ideas, or put them in touch with someone else. That did solve the problem they had.
Then you look back just a few months later, and either some folk have moved off the team, or the type of work had changed, or they just took their eye away from it. They’re saying, we are struggling with that same thing that you helped us with before. I said, and what about that thing that helped with it? We stopped doing that. What happened? Sometimes you go back, it’s just six months or so, a couple of team changes, some people have joined and left, different ideas. They don’t even know what you’re talking about. I thought, as I cycle round and other people are trying to help, how do we make that last? One thing that will either give you cheer or give you the idea that nobody can do this, so we shouldn’t bother, is it’s not just a me problem, and it’s not just a Co-op problem.
There’s a fabulous book from Henrik Kniberg about, “Lean from the Trenches”. This is a book-length case study about some people who’ve really solved it. They had a big piece of work to do. As an organization, with our stakeholders and bosses and the teams involved, they got Kanban working for them. They go into loads of detail about how they broke things down, about how they avoided the big traps of big-bang releases and not knowing what you’re doing. It’s really detailed, and it really felt like everyone involved came to an understanding of, we won’t do those silly things anymore, we’ll fix it. In his very next presentation that I saw him give, once he was at Spotify, he put up a slide that stopped me in my tracks. This is the same team with all the same stakeholders in management who had just learned all the projects. The very next thing they went on to, everything they’d learned had gone out the window and it’s all gone wrong. I don’t know how you can actually make change stick, but we can try.
On the topic of improving things, we got talking to lots of different people around Co-op Digital as it was at the time, about what challenges are you seeing. I think some things people hadn’t really taken into account. Some of the things we’d solved about, how do you start a team off? How do you get your stakeholders engaged? How do you set up your processes and set a goal and stuff? Was for those earlier stages. Alpha and beta are time-bound things. Once you’re live, there’s a whole lifetime after it. It had just been just about 18 months before we were talking about this challenge, I think every single thing that Co-op Digital made across its 12 or so teams was no further along than beta stage.
Now, over time, we’d suddenly evolved. You just turn around and things have shipped, everything’s live. Everything’s been going for a while. We’re a long way from the initial team kickoff and initially getting show and tells going and stuff like that. We’re just in a different phase now. We tried to draw up who are all the teams. What do you mean by a team is something that confuses people quite a lot, so drawing it out. I think another good bit of advice from the team topologies course was a team, their definition is five to nine people with a shared purpose. Larger groups are called groups. That’s an important distinction. I think that gets missed. The design team, the digital team, and things like that, confuses people sometimes.
Who are the teams we’re talking about helping? What challenges do we think we had? We did some user research. We talked to different people. We started seeing themes, the same things from on teams and the senior leadership team around it and other things like that. We pulled it together. Some things about accountability, that issue about, is it the business that decides what direction we’re going in or what we should change next, or is it the teams that do it? Some things we’re interested in was consistency. I do like the quote from Ralph Waldo Emerson, that foolish consistency is a hobgoblin of little minds. It looks neat if everyone’s sprint starts on a Thursday or they all do the same number of story points or something like that.
Teams have such different ways of working and different challenges. I don’t think doing exactly the same thing everywhere is the right answer. Make it familiar enough, but not totally consistent. That idea of easy things becoming easy and what works for the Co-op was good, because there’s lots of debate. There’s hundreds of different ways you could set up teams. There’s all kinds of what’s the perfect way to do a software team. It doesn’t actually matter. We should pick something fairly sensible and get onto the interesting hard problems, like, how do we understand our users and how do we make some business value happen?
From this and from talking to loads of people, it was as fat as a phone book: lots of big ideas about what we should change. I’d love to start with the leadership around team, taking some responsibility and understanding what’s going on. What tools you can use and what’s going on. If you want to know how do software teams work effectively, I think you’ve got this leadership thing. You’ve got those ways of working. You’ve got technical approaches are important. I think user research. It was so much. People agreed this was good, but when you try and look at it and use it, it was just too big a first step. We stepped back. Just seven things. We’ll keep it small. Each one was just a sentence. I think this was useful because it’s the size that fits in your head and it’s something we can say, are you doing this or not? We got one thing there. I’ll zoom you through some others, a few more here. You should have regular retros.
Sometimes teams say we don’t need them because we just talk about problems as they come up. That can be true sometimes, but not as often as people are just saying that. Each of these was something we were seeing just teams not finding a way to do. No one’s going to our show and tells. Stakeholders don’t really talk to us. Then they come up annoyed later when we’ve done something. I think a team charter is important. I kept banging on about team topologies to see who would pick it up. Things about understanding team health, like we’re not saying that for fun and we’re not doing it to be nice, we think these are the ways you deliver more effectively. The reason I’m not spending too much time on what the seven things are is that I don’t think which seven things you pick in particular is that important. There’s advice on how to do things well, but I don’t think that it’s these seven are magic. I repeated this again. I think it’s pick something sensible and see how you can make it stick.
We did lots of visits to teams and walking around, warning people about traps. It is the kind of thing you put in place and then you’ve forgotten it. You turn around, did we have a talk one time about things that were important? Other traps is mistaking the measure. It might be that everyone has a planning session, everyone has a show and tell, everyone has a retro. It might be no good improvements ever come out of them, no one actually comes to the show and tell. We’re not doing anything right. If a team was working well, you’d expect to see these things happening. Let’s talk about, are you getting value out of them? If you’re not, can we improve it? The last one, I mentioned before about trying to pick the perfect answers, going back and forth about, should we do this and should we do that? It doesn’t really matter. We should try something. If it doesn’t work, we can try something else. If we get this right, we could try anything.
How It’s Going
How is that going? Initially, this was a bit tricky. Like I said, that first, here’s a phone book of things to do, I got away from that. The seven things were interesting because we went around different teams, and if you talk it through like this, people see, this is useful, this is helpful. If there’s one of them your team doesn’t have the skills or time or capability to do, you’ll get help with it. It was good. It’s to help you reflect and surface issues. You can go around teams and get someone external to try and do the Fist of Five rating, which gives you really eye-opening things. If somebody says, five, we have retros all the time. Someone else says, two, nobody actually goes, and it’s always the same person who runs them, and we never get any improvements out of it. That was good. There was also the bad. Just a few steps away from me as we tried to scale this out, I found somebody calling it the seven things audit. Someone would come out with quite a formal forum and audit you. I was like, that’s not what we talked about at all.
I think that was eye-opening for me, how hard it is to have influence wider. I could understand if a different company picks something up and they misunderstand what you’re talking about. This is just over there. It was quite humbling to think, if I’m not with people all the time talking about what we mean, is that how quickly it goes apart? Over time, banging on about this has been super useful. As Co-op Digital ways of working and long-running teams have spread wider to a department that’s now well over 700 people, it’s been a super-useful, short, clear introduction to what we think is important. There’s lots of people who still use this today to run training based around it and to check in with teams like, are these things going ok? You’d be astonished how often it’s one of those, it’s something to pick and work on. It’s also reassuring for people. If it feels like you’ve got 100 problems, let’s rank these and pick one. The rest aren’t forgotten, we’ll come back to it. It has been a super-useful tool.
The Cynefin Framework, and Wardley Mapping
My next example of trying to help things was when I spent some time with the project managers. Outside of software, the Co-op does like, preparing for new rules around Brexit, opening new warehouses, doing things about distribution, all kinds of things. They’ve got projects and program managers and a PMO. They’ve got a waterfall process with just what you’d imagine if you’ve seen PRINCE2 or other things like that, so stage gates and lots of formalism around it. There is some confusion about sometimes when work comes in, if it goes to that part, we do some agile thing. If it goes over there, we do something else. With software or with other things, it’s just your luck about which part of the business it comes into. They were interested in helping the Co-op understand, what are the options and how would you pick? When would you choose different methods of delivery? Because sometimes it just feels like, I like agile, we should do agile. Someone else would say something different.
The best tool I found for helping talk people through was Cynefin, which if you’ve not come across it, definitely look it up. If you have come across it, I recommend this book hugely. There’s so much depth and understanding in it. Tianyi mentioned bounded applicability. That is definitely something that Cynefin talks about a lot. The idea, I think Dave Snowden invented it, to get away from new management fads. The idea that this is what you should do now. It turns out we’re all stupid for doing that thing before. It’s this. It’s this new thing. The fairer answer, the thing that people are happier hearing, and actually what’s true, is probably everything we do has value, but only within a set context. That project management stuff is useful for some kind of work, and this agile stuff is useful for some other kind of work. The difficulty we have is identifying what situation we’re in. That bounded applicability is definitely something that’s talked about a lot.
This is a diagram from the book. There’s different versions of it. It looks quite simple, Cynefin, as it talks about one of these four square things. What kind of thing are you in? When things are clear, you can write down the rules. You can write a recipe to follow. This is the only area where best practice is a sensible idea to talk about. The problem we have is when you’re doing something complex, where you can only experiment, you’re not sure what’s going to change or what the right answer is in this kind of work. We try and apply the same thing everywhere, is where you get in trouble. I’ve seen it go the other way as well. I absolutely love trying things out, seeing what’s wrong. Let’s not assume anything. Maybe we shouldn’t even sell sandwiches, which is fabulous and needs to be done more often in some cases.
Another time, that’s just me wasting time when we know we need to replace network switches in stores. Let’s just get on with it, make a plan and do it. This is good, and it was useful to talk people through. Something that you don’t see in this version is this gap right here. Some of these you can move between. You can understand more about your work and try and move it around there. Or if you want to put more process into something, if you’re sure the way you’re approaching something doesn’t change, you can try and haul something around, put more formalism onto it, try and bound things up. This is the way you do it. Always do it like this. What you don’t see in this one, but you do in a fabulous diagram from somebody else, it’s Martin Berg and Rob England, is that gap, that move between clear and chaotic is a cliff.
The fact that you can’t see it too well in the 2D version is on purpose, because you actually can’t. This is a good example that really made project managers sit up and take notice. If you’ve ever felt like the wheels are coming off the bus, something’s changed in my project, and I don’t actually have tools to put this back together, and I don’t know what to do, that’s what’s happened here. You have fallen into chaos, and the only way back is through experimenting the long way around and hauling back up to there if it makes sense. With that and a few other things, it really got their attention. Rather than asking them to read books and do lots of lessons on it, I think that made people feel, I see what you’re talking about here, I get this, and I’ve been in situations like it. What do you do?
Another really good way to learn more about Cynefin is Liz Keogh, has written loads of accessible blog posts that simplify it. This is something we actually used to rate different types of work, if you think about it. On this scale, people can really relate to that and can understand then, if no one has done this before, then we absolutely cannot do a year-long Gantt chart and tell you exactly what’s going to happen over time. We’re going to have to try and work it out. That idea about, for us, how complex and unknown is this, was important to do.
Another thing I like, and similarly to Cynefin, I think some of the writing or things about it can be quite daunting and off-putting. Even stronger than that is Wardley Mapping. That is another tool that is fabulous to use, but it feels like you have to do a PhD before you properly grasp what’s going on. Then, when you try and walk out to somebody else and be like, can I tell you about this? They’re like, what’s going on? I do not have time to read through all this. Similar to Cynefin, there’s some ideas that can really catch people. Simon Wardley talks about where things are new and uncertain and just coming into being, and where things are just commodities, they’re utilities and plug-in. At one point, it was a sensible decision for businesses to think, should I make my own power for my factories? Now it’s, no, just plug it in. Why would you do that? That shift’s happening all the time with all different kinds of work.
Rather than getting people to draw maps, although there are some folks in the Co-op who have had great success with doing that with stakeholders and getting views on things, we talked about this. The other thing that helped, there’s a list of different qualities that would help you identify what kind of work or what kind of thing you’re going to be doing. If the work you’re doing is to gain a competitive advantage, that’s more to that side. That’s more to the parts of Cynefin where best practice doesn’t really apply. Or maybe it’s more cost of doing business and we want to be more efficient. People are like, I get you. I could identify that.
The interesting thing was, in the same project or program of work, some things might be the key competitive advantage that are uncertain and you want to use something different there. Whereas others, you don’t have to do everything the same way. Others you can plan out and there is no need to have workshops and think about it. There’s a list of different ones of these. For different people, different ones catch them out. It’s, yes, volume. That’s what we’re doing with some of our work. Whereas other people, other types of work. It’s like, yes, we do need to experiment. What does that mean for which tools we should use? That really helped us get away from the idea that if project management isn’t working for you, you just need to plan harder, is what some people think. Or on the agile side, we can’t assume anything. We should always start with discovery and we should work it out from first principles. I think there’s right answers for both.
I put together a decision guide and it was interesting doing this as I got to talk to people in lots of parts of the Co-op. We’ve got such depth of experience and understanding. There is no way I would have been able to write this up and put some of the ideas we’ve talked about together myself. Apart from writing a guide and helping people navigate through how to do different things, so many people got in touch with people and had such useful conversations. I’d recommend doing that. Some interesting things that came out of it was things like, shouldn’t we have discovery, alpha, beta, live everywhere? Project management side are very familiar with, you’ve got defined processes and everyone goes through them. I would say, no, that’s designed to slow you down. That’s designed to make you stop and think. Sometimes that’s useful. Give an example of the operational innovation store. It does lots of work about store processes, like how can we change them and make them more efficient? They’ve got a version of that discovery, alpha, beta, live.
When they look at something like leakage, where’s our stock going? Is that getting stolen, or spilled, or misrecorded? They have no idea what the problem is. There’s no way you could get funding and start off saying, this is how we’re going to solve it. There’s an example here of a smart gap. At one point you had paper, you had to go around and see what the gaps in the shelves were and do that. That’s part of the inventory process, knowing what you have to reorder. It took ages and it wasted so many trees. Through an exploratory process of what would work and actually going into a store and spending lots of time with the people there, they came up with this technical solution that saves 15 minutes per worker, per store, per day across thousands of stores, and something like 5 million sheets of paper a year as we don’t have to keep shoeboxes full of paper in the back anymore. That started out very exploratory. Then their alpha phase would be trying things out in that one store.
Then you would move on to, let’s go to a couple of stores but let’s put a lot of attention in and be prepared to tweak it and change it as we realize it’s not working. Then later that just flips to quite unknown. It’s still tricky to implement, but how do we buy thousands of these? What order do we visit the stores in? How do we set up training? I can absolutely see that flipping from exploratory agile through to trying to optimize it, through to something plan driven, and waterfall seems like a sensible answer at the end. Should everyone do that? No. If you’ve got something like our online e-commerce site, when we want to try something, you can just go and try it.
If it works, you leave it on, and if it doesn’t work, you turn it off. I think right now if we all went to shop.coop.co.uk, I think 1% of us would see the different version of this site that they’re trying out now and some new technology in a different setup. If that goes well, that will ramp up to 10%, 50%, as everybody. When we’re making those tweaks and incremental improvements, there’s no need to put that formal process around it.
How is it going? What I would have loved, that idea of Cynefin and what kind of work is this, I would have loved people to formally go through each new bit of work, how can we split it down, answer the questions, and build up like a case library. It’s not just saying it has these qualities, it’s like, do you remember this? Is this one of those? That would be so helpful. Record, when we’ve suggested working in this way, did it actually help? Come back to the projects later and look at them. That’s absolutely not happening. It’s a big read. People are busy, and they’re not going for it.
One thing I like from Simon Wardley’s work, he says, quite often it’s the act of making the map that is more valuable than what you get out at the end. Again, I might be kidding myself to make myself feel better but those conversations and that understanding, the light bulb moments I had as I talked to people, felt super valuable. The fact that the process as described in the 20-page document I’ve got about the decision process isn’t getting used, maybe it doesn’t matter and maybe it will get used one day.
Building Community
I did mention briefly earlier that the Co-op Digital was separate to lots of parts of the Co-op. In different areas, they had some agile teams but they also had some software written in quite a traditional project way, so you might never meet the testers, you might be 20% on this and 20% on something else. Possibly due to time with the project managers, possibly due to all those other conversations and really qualified people around the Co-op, that’s changed. Digital technology and data isn’t scattered anymore, it’s now something over 700 people. By and large, every single team, if you’ve got a software system, that’s not just for Christmas. Everything’s owned, and it should be in a sensible size of team. That team has responsibility for keeping it maintained, for keeping the tech debt down, for helping navigate their own roadmaps and stuff. For loads of people this was brand new. To help, I tried building a community. It was called the Ways of Working community. I suggested, let’s just get together.
A lot of people don’t actually know what agile looks like, what do those teams do? I quite like that since COVID times Co-ops stayed pretty much, mostly remote. You can go in the office if you want. In the old days, if I thought, I don’t know who will come to the community, might be 5, might be 200, I don’t know where you’d find a room for that. I really do like the idea that you can just put a calendar invite in and put a few messages online and see who turns up. Loads of people did. I think when something’s new and exciting people come along to it, but it can die away if you’re not clear about what it’s for.
In building community, if you’re interested in doing that at all, I thoroughly encourage you to check out Emily Webber’s work. We started talking about purpose. A lot of people hadn’t seen a community of practice yet before. There’s these posters on Emily Webber’s website about like, what’s it for? This was useful. That’s the “Communities of Practice” book that I also recommend getting. Getting people together and some of them using Miro for the first time, we all encouraged each other to talk in small groups and apart. I think there was something like 90 people in the first session.
We managed to collaboratively put together, what is this community for? Because I wanted to come back and we put things up, and if you agree with it, you can add some emojis, if you disagree, you can add different post-its. What if we came back together in three months, and how many people would you want to get together? What statements would they agree with? How would we measure that? I think this was good because it was something we kept checking back in on. Lots of things start off with loads of excitement and then fizzle. What keeps it going? This is something we came back to every couple of months. The other innovation I super liked was these little profiles. Write down why you’re here, what do you want to learn? Help us choose topics. Also, what could you share? Because I’ll come and nudge you, and try and get you to do things. We made a souvenir badge for each topic you went on. You can see the kickoff was there.
Some of the topics, the things people were willing to share was really impressive to me. We had modern agile practices and how close I’ve seen Co-op teams get to them, which was interesting. It was a tour. People like, that’s agile stuff, could it ever work at the Co-op? It does. It has. That’s got practical examples from all kinds of places. Technical excellence and why it’s important. Some of the ones I really liked were from teams brand new to working as a long-running agile team. Cloud journey and the identity team were talking about, we’re trying this. In many ways that’s more useful to people. If somebody has been doing something for absolutely years, you’re like, I don’t know when I’ll learn. How do you know how to work these ways? If it’s somebody who’s just tried it this week and they’re really happy to share what they’re trying out and whether it’s working or not, often, that’s what people want to hear. It was really good. Potentially more successful than some of the talks and write-ups we shared was the badges.
If you’re interested in making your own, here’s a detailed guide to setting them up in Miro. When people agree to speak, they really like the chance to design their own badge and see what’s going to be on people’s profiles afterwards.
One other thing that grew out of that was another experiment I tried out, Lean Agile University. I’d come across this at one point, so Jez Humble co-wrote “Accelerate,” and “Lean Enterprise,” and loads of other stuff. He teaches at Berkeley University in the States, and he’s put his short lectures and the activities you can do online. He teaches a course in product management to people learning software engineering. Why does any of this stuff matter? I could not describe how excited I was to find this. I said at the start of running the training course, if you’re doing like a fantasy football style, who’s your best team for teaching you this stuff? I think Jez Humble might be my first one. I cannot believe he’s put this all here. We tried. We made Miro boards about all the topics, and worked through it. I think it’s all Creative Commons, so you can remix this, you can use that, you can do things like that.
We all watched the short videos separately, but you make notes and comments on it together on the Miro board, and then you come back together to do some of the different activities, try out user story mapping. It was a really good two-day course. There were loads of people who had never met each other in the Co-op. There were some fabulous eye-opening moments, so somebody said, so a team that has metrics it wants to move, and you choose your work depending on what happens every week. That could never happen here. The way teams get funded, you get set up with a project, you have to do a thing that the stakeholder has decided. Someone else in the course said, no, we do that. We’re doing that right now. I think the legal services team actually shared some screenshots and we kept it for future versions of the course, of here’s an A/B test we’re running right now, here’s the metrics we’re trying to move, and if that doesn’t work, we’re going to do something different.
Just that eye-opening, this isn’t just some Silicon Valley dream story, this is happening in a team right next to me, was fabulous. Other things that came out of that that were really useful was, I was learning absolutely loads as well. Jez Humble talks about personas and how important that is. I got people to share, did we do that at the Co-op? There are some examples there of what’s going on. At one point a senior designer said, actually, X, Y, and Z are reasons why personas don’t really work for us as much, we’re moving more towards these missions that people are going on. That matters more than the individual. There’s a Co-op Digital blog post about that. I was like, I had no idea. That was something that we shifted. I think that designer actually came and co-hosted a two-day training course to help share the load.
It was just interesting to get different perspectives. We moved towards the rare and coveted university facilitator badge as people in different roles. There was people quite new to their engineering career talking about their perspective of it. There was designers talking about what they see, as well as senior engineering leaders. Everyone put their own different focus on what they thought was important in the course. That was called Operation Not Just New, and it handed over quite well.
How’s it going? Lean Agile University is still going. You can get bespoke versions for your own team. Because I think that that go wide and meet people you’ve never met before is fascinating. Doing it for a specific team is really useful. There’s also a self-paced version. It’s only internal at the moment. It’s been useful. It’s been eye-opening. It’s still something that people refer to now about, do you know like we did on the course? It warms my heart when I hear about somebody using a technique, years later, that they first saw on that course.
What’s Next?
What’s next? There’s other things going on. There’s too much to talk about. It’s always hard. It’s never perfect. There’s always more things to try. I think if I was to sum up what I’ve learned from putting this talk together and reflecting on it, is, be boring. Keep banging the drum. I’ve written advice for people about, with the seven things or other things that are important, your job is to be a broken record. Because you’d be astonished how fast things get forgotten about. The kind of help people need is I think they can do anything. They can copy the outer forms of whatever you tell them is important. That’s not what I want. Be super clear on that. There’s something here that is meant to solve a problem. It will feel like this problem is getting solved if you’re doing this thing, and you can’t see it solving a problem. You should either stop or you should ask somebody about it. If you’ve ever seen a user story starting with, as a database, I want to be updated. You know what I mean by people can copy the outer forms of things. That’s not what I want you to do.
The last one is, make it fun. Make it welcoming. The Ways of Working community and the lean agile unit in particular is some of the most fun I’ve had at work. Those rare and coveted facilitator badges are still up for grabs, I think, as more people hopefully will co-facilitate that in future.
Resources
If you want to know more about the Co-op, I’ve mentioned the digital blog a few times, digitalblog.coop.co.uk. There are years of what teams have done, how the Co-op works, and things like that.
See more presentations with transcripts
