Steven Pereira, lead consultant at Visible Value Stream Consulting, shows how simple maps expose where work actually flows and where it stalls. Drawing on his time as a CTO stuck in finger-pointing and assumptions, he explains why the act of mapping delivers most of the value even if you throw the map away. He argues that AI mostly accelerates the small share of time teams already spend being productive, while the rest still disappears into meetings, approvals, and deliberation. Starting from a target outcome, he moves from value stream networks down to operational learning loops, uses a large regulated bank to show several constraints surfacing at once, and explains how a map becomes a roadmap for the next three to six months.

Hi everyone. I'm very pleased and humbled to be here. This is an incredible space and I've been kind of introduced very graciously by Brian. A lot of people here talking about mapping, modeling, visualization, which are all things that I really believe in. I have found that they're really the only ways that I go from not knowing what to do to having a clear picture of what's going to make a difference, especially in groups of people. So my introduction to value stream mapping was out of failure, out of trying many, many different things and never having that answer about what is going on? Where should we invest? Where should we focus?

Initially, I had no idea that it was even a thing. I'd heard about it a couple times, but as a CTO, I ran into multiple situations where we just couldn't get anything done. We couldn't get out of our own way and we had no idea what was really happening. Lots of finger pointing, lots of assumptions, lots of bias, no clear picture about what was actually happening in the environment. And mapping really made a difference between struggling through that and getting everybody on the same page and everybody speaking the same language. The one thing that I want to leave you with out of this, I'm going to show off a lot of maps. Tomorrow we're going to build maps, but really what the value of any mapping exercise is the mapping itself. I always say that the entire value of the exercise is just mentally building a model in your mind that matches, to some extent, everybody else's model so that you can have better conversations and you can get on the same page with people.

So if you map, you can throw the map away and still gain almost all of the benefit. What we do with flow engineering and the mapping that I do is very, very simple maps. I want to show you very, very simple maps that give you a lot of benefit without a very expensive time-consuming artifact that just gets stale and becomes a liability over time. But I want to zoom in all the way because everyone's talking about AI. Everyone knows that AI is both the cure to everything and also a gigantic liability. But what I see in a lot of organizations is that their flow at the individual contributor level looks a lot like this. Lots of activity, very little value added time, very little time actually spent on getting work done. Anybody familiar with this? Anybody look at their calendars recently and see all that builder time blocked out or is it all eaten up in meetings and syncs and everything else that gets in the way, right?

So what do we do with AI? We just accelerated that time that's already fast, the time that we're already productive. We're just amplifying that as much as possible while we still have this 92% just eaten up in meetings and approvals and endless deliberations and consensus and analysis where we really could be accelerating entire systems of value creation, entire outcome and mission factories. I want to talk first about sub-optimization. This natural tendency that we have to zero in on areas that we know really well and fixes that we think are definitely going to move the needle because they will, unless we zoom out and think about the big picture. What is the default view in organizations that leads us to this, doesn't make it any easier, is we have lots of artifacts, lots of spreadsheets, lots of org charts, lots of these artifacts that give us a static view, this component view of organizations, which naturally lends us to suboptimization. Pick a component, optimize it, move on to the next one. But even that tendency to zoom in on one component means that to do that effectively, we'd have to understand what's the most important component that's holding back the entire system. Delivery actually looks more like this. It looks like an interdependent network of activity. Some activity is actually positively or negatively affecting other activity. But what we see with delivery is that we're actually flowing through this network all the time in different ways. Some of those flows are stable, some are highly volatile, some change every single time we have something to do, and some of them are 100% stable and persistent and predictable. But we have to be able to zoom out to that network level to really understand where is the biggest opportunity here? What's going to make the biggest difference overall or else we might end up working against each other? So finding the constraint in that entire interdependent network is absolutely critical. Now, is it easy? It is not easy. We have to really boil that down into something simple, digestible, and easy to communicate with people because we don't work in silos. We should not work in silos. We need to be collaborating effectively with people, which means we need simple ways of clearly depicting what's going on, addressing that situation of what's happening. So this is an example of a value stream map that I commonly create. Doesn't get any more complicated than this until you add the actual data, which is where a lot of the value comes from. When we start talking about data, we start challenging our assumptions, challenging our biases. We stop talking about it depends and we get more specific. We start talking about real actual delivery that happened in the past and we think about what does that mean for focus?

What we're able to do with this is clearly identify a focal point. Whether it's the ultimate focal point or not, it gets us all focused on something that is not individual pursuits of various assumptions. We can all come together, clearly decide to focus on something measurable, improvable, and move forward rather than constant deliberation or chasing 50 things at the same time. I'm going to talk a little bit about this zoom in, zoom out aspect of mapping that I think is really, really important. Bryon has showed the Wardley view, which is kind of the network view that I like to think about at the top level, the strategic level, and then all the way down to operational. But I'll go back to operational at the IC level, the inner loop to talk about that first. We're seeing a lot of investment at this level because it's easy, because it's local, because it's well established, it's pretty stable.

We have to build things before we test them effectively. We've been here for a long time. We've been doing continuous delivery for a long time. But what's really important here is the parts that people are not investing their AI resources in, the learning aspect of this. When we loop through things, if we're just publishing output, just putting out results, we're missing 98% of the value, which is learning consistently how to do it better every single time or making sure that we're actually making forward progress. But even if we optimize this, what I've been saying is that that's a suboptimization if it's in isolation, if it means that it's pulling us in different directions. A lot of what's happening at the operational level is now we have a lot of ICs. We have basically everyone in this room can have their own team of agents who is reinforcing their biases, creating an echo chamber, creating a whole bunch of resources that now has to be reconciled with everyone else's silos of agents and we bump up against each other.

Everything that comes to us, we filter through an agent. Everything that goes to someone else gets filtered through the same agent, but with their skills and their context. It's a huge mess and it could potentially get a lot worse. It sounds to me a lot like the early days of DevOps where we had developers, we had operators, two different silos, people throwing work back and forth, and what they really needed to do was just have lunch together and talk about what's it really going to take for us to get results, to get positive outcomes. For me, mapping is a huge asset for that scenario where we need to come together and we need to break silos and really get on the same page about what's happening and where should we focus. We zoom out. Now we're getting a little bit more consequential. We're talking more at the cross silo level, the cross team, cross organizational, cross portfolio level at tactical views. Really, that's where we actually get to outcomes. That's where we get past the iterations to shipping results to users, shipping results to the field, getting our outcomes, getting our measurable impacts, moving past the assumptions of I think this is going to be good to yes, we actually hit our targets so we're actually seeing things trending in the right direction. But if we zoom even further out to the strategic level, that value stream network view, if we're focused on individual innovation, if we keep enabling individual contributors and even managers and leaders to be operating in their own silos, iterating rapidly, we're really creating this siloization once again that we're going to have to break down. What's really needed is this end-to-end flow where we say, despite the fact that we're all able to work in isolation, it all needs to be aligned to the same view and really focus on the constraint that's holding us all back collectively.

If we put that together in a view that's kind of top to bottom, you can think of strategic at the top, we have value stream network. If you zoom into any of those nodes in that network, you get a value stream that is delivering results and then even further down you have your operational loops, which should be OODA loop, learning loop, Toyota, Kata, whatever you want, hopefully resulting in continuous improvement on a regular basis, not just publishing more output. But there's a lot of measurement here. This is probably one for the pause on the YouTube, but none of this really matters. I just want to express that all of this needs to be a loop, not just the inner loop. There's loops everywhere. The thing about a loop is that it closes and moves back on itself. We actually think about what are the results of what we just did? Did we hit our target? What's the gap? What are we going to do to close that gap or address that gap in the future? Okay, so we've all heard this, if you want to go fast, go alone, if you want to go far to go together. AI is actually making it possible for us now to go faster alone than ever before and together is going to get harder and harder. As we double down, everybody has their own harness, everybody has their own echo chamber, their own bubble, their own infrastructure. This is going to be really challenging. I think that mapping really gives us a sense-making exercise that can tie us back together and make sure that we don't go off in different directions as we evolve these individual capabilities. Okay, so if we simplify this and we think about what is a typical current state, obviously it's a mess in almost everywhere that I walk into, people don't call me when they want to show off their perfect value streams, but I walk into a lot of environments where it is a very complicated situation.

It can be difficult to actually map it out, which is why we always start from outcomes. We actually start with what are we trying to do? Because I don't care about everything in the current state. I care about what matters to the outcome that we're trying to achieve. So that helps me filter this down and helps me eliminate all the things that we could possibly map and all the things that will really make the mapping process take forever to what's actually most relevant to achieving the target outcome, whether it's improved quality, faster lead times, fewer defects in production, whatever it is, it will help us constrain this to a useful degree so we can do it quickly. I wonder how many people here have mapped a value stream before? Anybody feel like it took a long time and was really exhausting? Only a few. Okay, good. I go into a lot of environments that say, I've tried this before, it took forever and we really didn't act on the results. We got a map, we didn't act on anything that it told us. Another piece of the puzzle is downstream of the mapping, which is like acting on what it tells you. But what happens when we do this mapping and we get our simplified picture is that a simplified picture makes it easy for us to point to something and have a short conversation. We can share it effectively with people who were not part of the process and we can make it very clear, this is what I'm talking about. Here's the part that's outside of my purview. Here's the part that's downstream of me. Sometimes we don't even have visibility into those things. We just map them as little clouds and say, "I don't know what happens here, but some stuff happens here and I wish I did know." I can take the map to those people and say, "Can you help me fill in the gaps?" That's very helpful.

But in a more realistic scenario, here's what it kind of looks like in a large complex organization. This is a bank, but highly regulated, lots of legacy, lots of silos, very complicated. By mapping, you get these clear focal points. We're able to clearly draw attention. We don't have to pick a single constraint. There's multiple things that we could pursue in parallel, that's totally fine, but what shows up is all these handoffs and these negative conditions, these wasteful activities that of course we know about, but making a map is going to draw it to our attention and make it a lot more obvious what to do about it and get everybody on the same page about not just dealing with toil and pain all the time, but things that we can actually address in an area that's going to have a measurable impact and meaning to the company or the organization, meaning to the mission.

So what happens afterwards in a situation like this is that we don't even have to actually redesign any process. We could just get the things that we don't want to happen out of the way when they can happen in parallel. We don't have to redesign every process. We just start to see a lot of things that don't need to happen or they can happen slightly differently or we can make them happen in parallel and we see this almost overnight improvement, which is really powerful because not everything requires tooling or a massive investment. Sometimes it's just getting people together and saying, "Oh, I didn't know that." I've been a part of a lot of mapping exercises where one party in the working group says, "This is the first time we've ever been in the same room together." Or the other one is, "I've been at this company for over 15 years and this is the first time I've ever seen our start to finish flow of activity." That's really powerful. Those people leave, they have new relationships, they have new perspectives, they feel empowered to actually change the system of work, whereas they would usually just be dealing with all this pain and frustration until a breaking point, right? I want to leave you with actionable, tangible result. If you can't make the workshop tomorrow where we're going to go through everything from start to finish and I'll show you exactly how I map and how the entire process results in actual results and actions, it's really this simple. We really can do this in 90 minutes with a group of people. It doesn't have to be perfect. It just has to give us a common focal point, start a new conversation about making a difference. People always say, "Well, okay, that's great. That's one value stream. We have dozens. We have hundreds of value streams." But what happens when you start mapping is you start to see that there are these common constraints, common dependencies, right?

I don't need to map every value stream in the organization to find out that everybody struggles with the review process, with procurement, with access to environments, with environment stability. All of these things really show up quickly once you map a couple of value streams. That's how this scales and how it really doesn't have to be a massive investment of time or anxiety.

Dan talked about making failures visible, treating failures as these opportunities and looking at a value stream is a great way to really lean right into that. These failures are happening today. They're going to affect us unless we do something about them. A map is a great way to start that conversation and not talk about individual contributors, but talk about the system and scrutinize the system rather than people doing the work. Just to give you a quick overview of what flow engineering is, which is just a flavor of value stream mapping that gives you outcome focus at the beginning and then a roadmap at the end of the process, which are two big missing pieces that I find pretty common. We start with this outcome mapping to really identify what are we really trying to do that helps us filter out a lot of the noise and make sure that we don't get distracted by every detail, all the meandering paths that work can take in the environment and we stay focused on what do we need to know to understand our current state relative to the target outcome.

But the value stream mapping is of course a critical piece of this. Again, it doesn't have to be complicated. It doesn't have to be fancy. It can be post-it notes on a whiteboard. It can be scribbled on a piece of paper and you can get your AI to clean it up and make it look super fancy, right?

This has never been possible before and a lot of people had mapping anxiety where they didn't like their handwriting or whatever, and then they'd have to spend a ton of time making a slide deck for somebody, but luckily that's one of the benefits of AI that we have today. The dependency piece, I didn't talk a lot about aside from the ROI slide, but that's where we really dig deep into the constraint. Where we see a hotspot, where we see a lot of time being spent, that's the only place that we look for detail. We do not try to boil the ocean and try to map everything that we could possibly map. We care about where things are going wrong the most, where we see the most opportunity, and that's where we look for detail. Then after this point, we get into this stage of actually designing a better system, right?

So what does better look like? What can we achieve in the next three to six months? So that gives us the visual asset that people are going to use to actually go back to their jobs and put all this in context. But what's really going to make this actionable is the roadmap piece, and that's the piece that I find a lot of people fall off at the end of mapping. They don't take what they've learned and the opportunities that have showed up and break that into here's what we're going to do today. We're going to do these things later. Here's who's going to own those actions and here's how we're going to measure the progress towards those action items that are going to bring us to the target outcome. This is really meant to be start to finish everything that you need to do value stream mapping effectively.

I hope you're interested in coming along on a journey tomorrow where I'll take you through the entire thing. But if you are just curious about this, if you struggled and you've got questions, I would love to talk to you about mapping or value streams or flow or measurement. Anything in that domain is my favorite stuff. Thanks again for having me. It's an absolute pleasure and I hope to see you tomorrow at the workshop.