Systems Thinking For Leaders Designing Solutions That Work

TITLE: Systems Thinking for Leaders: Designing Solutions That Work CHANNEL: MIT Sloan Executive Education DATE: 2026-05-15 ---TRANSCRIPT--- Welcome everyone. It’s great to see this very global group joining us from all over the map and all times of the day. We’re really delighted to be with you and it is my pleasure to introduce our speaker today Professor John Sterman. Professor Sterman is the JW Forrester Professor of Management here at MIT Sloan and a professor in MIT’s Institute for data systems and society. He also serves as director of MIT Sloan’s sustainability initiative. Professor Sterman’s work focuses on improving decision-making in complex systems spanning corporate strategy, sustainability, public policy and climate change. We’ll touch on some of these areas today. He’s widely known for pioneering management flight simulators, tools that are now used around the world to help leaders better understand how real systems behave over time and we’ll also get a glimpse of those in today’s webinar. Professor Sterman is the author of the award-winning business dynamics, one of the most influential text in systems thinking and he’s also faculty director of our long-running executive education program by the same name, business dynamics, which we’ll learn more about at the end of the session. Please join me in welcoming Professor John Sterman. John, thank you so much for joining us today. Oh great. Thank you Molly. Thank you Charlotte and welcome everybody. Delighted to be with you all and from all over the world. Amazing. So let me tell you what we’re going to do today very quickly. I’m going to just try to give a little introduction as to why systems thinking is needed and then we’ll do a couple of examples of using some of the different tools, not all of them. We don’t have time for all of that and we’ll take your questions and at the end I if time permits talk a little bit about how these tools in system dynamics have been used successfully in a wide range of organizations. The key thing here is please put your questions and comments into the chat. Don’t hesitate. Uh you can also post questions through Q&A. Uh but use the chat and uh our our uh terrific team here is going to monitor that as well. I can see them as well. And we’ll try to make this as much about you as possible. So, the first question is why do we need systems thinking? Well, it’s become very popular lately. Everybody’s talking about we need to become systems thinkers, systems thinking is a core capability for success. And why? Well, a lot of times people say, “Well, listen, we live in this overwhelmingly complex world and you can’t do just one thing and everything is connected to everything else.” All of that is true, but it isn’t very helpful. So, what we need to do is to move beyond slogans about everything being connected to everything else towards operationally useful tools that you, even with uh say a one-week executive course such as the one that we’re offering this June, can actually begin to start to use and develop your capabilities to make a difference. So, why then is systems thinking needed? Well, I think you’ve all probably experienced the phenomenon that we call policy resistance. Now, policy resistance is not the idea that you have a great idea and you couldn’t get it implemented. It’s more subtle. It’s the idea that you’ve got a pressing problem. You need to find solutions. And you do all the research, you consult all the experts, you you do everything you need to do, and you come up with what you and all your advisers agree is a terrific solution that’s going to work. And you actually get it implemented. But then it either doesn’t work or more commonly and insidiously it works in the near term over here, but the problem comes back a little later over there. That’s the idea of policy resistance. It’s extremely common and it’s extremely discouraging. Let me give you just a couple of examples. So, almost every big city in the world suffers from terrible traffic problems. What do we do when there’s terrible traffic? Well, you all know the answer. We build more roads. What happens then? Well, in the near term you’ve made driving a whole lot more attractive. So, people drive a lot more and they don’t take public transit and they drive more during rush hour and even move farther away from the city core where their job might be and order to have the ability to buy a house for less money and then they’re locked in. So, what we often see is you build more roads, traffic congestion not only doesn’t get better, it often can get worse. Another example is for those of you who are in the United States, you’ll be familiar with this. If you’re not, the United States health care system is the most expensive in the world measured any way you want to measure it. Cost per person per year, fraction of the GDP. It’s about 18% of the world’s largest GDP and despite decades and decades of efforts to contain costs and slow down the cost escalation, none of this is really working. This is a major example of policy resistance over at least 70 years now. Mergers and acquisitions. Many mergers and acquisitions fail, particularly destroying value for the acquiring firm. So, mergers are promoted as we’re going to realize cost savings by eliminating duplicative back office operations. We’re going to have synergies in our market position and our product line, etc. But often they end up destroying value for the acquiring firm. Many of you out there have probably participated in process improvement programs, whether it’s Six Sigma or re-engineering or going back total quality management. There’s a wide range of these and the tools work. Many organizations have succeeded, but vastly more find that their attempts to implement a successful process improvement program have failed and that breeds cynicism about the next one that comes along and makes it harder to succeed. Most of you out there, if you’re in a business organization or an NGO or government or any organization, you’re probably involved in projects, project management, whether it’s new product development, software, construction, whatever it might be. And typically projects are late, expensive and wrong, meaning they don’t meet customer requirements. This is just a very, very short list and I’m sure you can easily add to it from your own personal experience. So, why does this happen? Well, it’s not just that the world is complex, it’s that there’s a mismatch between the mental models that we use to come up with our solutions for difficult problems and the complexity of the real world. One of those mental models is what I call the open-loop mental model. So, just ask yourself the following question. How many times in a project or any activity in your organization have you seen this kind of a diagram? Here’s how our project is going to go. We’re going to identify the issue, gather the data, evaluate our alternatives, select the optimal choice and then implement it. Problem solved. I imagine many of you have seen kind of diagram in various guises. Well, I don’t know about you, but no project that I’ve ever done from co-chairing the project to develop the new building for the Sloan School of Management to putting together and shopping for the grocery list has ever gone that way. Instead, we find out as we go through the process uh we didn’t really know what the real problem was. So, we have to go back. Uh we get the data, we find out key data aren’t available or where we we were looking at the wrong thing. We have to go back and so on. There’s constant iteration, planned and unplanned iteration, where we discover problems, rework that has to be done, and we have to go back. So, although this is especially popular and common in organizations of all types, it is not how the world works. Instead, it’s a world of feedback. This is the idea. You make decisions, your decisions change the world in various ways, and that creates new information that conditions your next decisions in a never-ending feedback cycle. So, the Escher drawing here is magnificent, and it’s a great metaphor for one of the core ideas in systems thinking. But, let me make it a little more precise. So, here’s what we call a causal diagram, and mapping your system issues, as we’re going to see in a minute, into this framework is a very powerful tool to help develop your systems thinking and avoid policy resistance. So, how does it work? Well, I’ve got goals. Maybe uh in my company we want to gain market share. And uh and then there’s an actual market share, an actual state of the system right here. And what I need to do is I need to make decisions that are going to help move the system towards my goal. So, I could um I could cut prices. I could hire more and better sales people. I could advertise more. I could try to buy market share through a merger or an acquisition uh and develop better products. There’s a lot of things I could do. So, what’s the problem? Well, let me give you an example. I’m a bicycle commuter. I live about 12 miles, what, 17 km, 18 km from campus. And when I ride my bike to work, I need to use this idea of feedback in order to do it safely. At any instant, I need to know right now, which way do I turn the handlebars? Do I turn them a little left or do I turn them a little right? I can’t do that. I can’t know which way to turn them unless I have feedback. So, here’s Here’s a little picture of me going down the bike path that I use every morning. And I know I have to stay on the right. That’s my goal. But too far to the left, I’m going to hit those folks. Too far to the right, I’m going to hit a tree. So, I need constant feedback to tell me where I am relative to the goal. Stay on my side of the path. And if I don’t, I’m going to have a serious problem. I can’t know what a good decision is unless I have that feedback telling me where I am relative to my goal. Now, listen. If uh riding uh if if controlling your company, your organization was as simple as riding a bike, we wouldn’t need systems thinking. But it’s not for a bunch of reasons. And the main reason is we are embedded in more complex systems. So, think of this diagram as your mental model of how the world works. And it’s not wrong. It’s just incomplete. So, you’re going to make decisions that you sincerely believe will move the world closer to your goal. I’m going to cut those prices, hire better sales people, develop new products, uh advertise more, et cetera. All of those things in isolation ought to succeed in helping me build market share for my company. But, wait a minute, there’s all the other consequences of my decisions. So, you can’t do just one thing. You make decisions, some of them are going to do what you hoped, but others are going to trigger so-called side effects. Now, there’s no such thing as a side effect in reality. There’s no such thing. People talk about it all the time, but there’s no such thing. There’s just effects. You make decisions, and your decisions have multiple effects. The ones that worked out well for you, that you planned, yeah, that was my plan all along. I’m a fantastic manager, you give me a promotion. And everything else, especially the things that feed back and undermine your goals, oh, that was a side effect. That wasn’t my fault. That came from out there. When people say, “Oh, it was a side effect. It was an unintended consequence. Nobody could have anticipated that.” They’re trying to persuade you that they are terrific managers, leaders, but what they’re reeling really communicating to you is how narrow and inadequate is the mental model they’re using to make those decisions. Now, of course, it’s even worse than that, because you’re not the only actor out there. In the business setting, you have your employees, your customers, the financial markets, your suppliers, the people in the communities in which you operate, and they all have their own goals. But, since we all live in the same world, when you pull the state of the system closer to your goals, you’re almost certainly pulling it away from theirs. And they aren’t just going to sit there. They’re going to take their own actions and try to pull the system back. Of course, they also have limited mental models, which means their actions are going to generate unintended so-called side effects. And now we have a very complex system. And even the goals are part of the system. Think about a salesperson who has quotas. You do too well, your quota is going to go up. And there could be time delays in any one of these causal linkages. This is now too complex for people to just kind of figure out without the aid of systems thinking tools. And the result is unintended consequences. And this is an ancient idea. Here’s an example going back to Sir Thomas More 500 years ago in his masterpiece Utopia, where he said, “It will fall out as in a complication of disease that by applying a remedy to one sore, you provoke another. And what removes one ill symptom produces others.” Perfect description of policy resistance from 500 years ago. So, let’s do an example, and I’m going to draw one from healthcare. Those of you in the United States, you experience this. And those of you in countries with your own national healthcare systems, you probably experience something similar. So, to try to combat cost escalation, folks have been working on how to avoid overuse of expensive tests and drugs and procedures. And there’s three main ways this is done in the US. The one main way is called prior approval, which requires clinicians, your doctor, to receive prior approval from the payer, the health insurance company, uh in order to be able to do any of the tests or many of the tests and uh prescribe many of the drugs and procedures that you, in their judgment as your doctor, think you need. That’s prior approval. Preferred drug lists is similar. It says clinicians cannot just prescribe what they think is the most appropriate drug for you. They have to use generics or cheaper alternatives that may not be perfect substitutes from a list of of drugs that are approved by the payer. That’s designed to keep doctors from prescribing expensive things when something else in the minds of the payer would work just as well. And step therapy is similar. It says doctors have to initiate treatment with a cheaper substitute from an approved list and only switch you to the drug they think is most appropriate if the if that initial treatment fails. And usually in order to escalate, they have to get prior approval. So, I hope that’s clear. Many of you out there have personally experienced this and maybe you’ve actually had your doctor say, “This is what I think you need.” But then the prior approval process resulted in a denial. So, is it working? Well, this is the estimated cost of prior authorization in the United States. It’s now 35 as of 2024, 35 billion a year just on the administrative costs. That’s 11,000 per clinician per doctor, if you like, per year. And it costs about 20 to 30 per request for prior authorization. Um So, that would be fine if in fact it cut health care costs by more. But in fact, that’s not true. And it’s not only not true, we get exactly the opposite. And we’ve known this for 30 plus years. So, here’s a study from 1996, 30 years ago. The conclusion was limiting the drugs that can be prescribed with prior approval, preferred drug lists or formularies, etc., which is intended to prevent the unnecessary use of expensive drugs, actually has the unintended effect of raising medical costs. Now, that’s just not one isolated study. A brand new study just came out earlier this year was a um a meta-review of 25 different evaluated studies, and that concluded that in addition to delays in getting care, authorization requirements were associated with disease exacerbation, that is people got worse, preventable hospitalizations, prolonged hospital stays, and lower rates of disease-free survival. That’s medical jargon for more people died as a result of prior approval. So, how can this happen? We have very smart people running our health care system. They sincerely want you to get better when you need care. Uh so, how can we actually produce the an outcome that nobody wants? So, let’s use that diagrammatic technique, causal mapping, to get a handle on this. So, here’s my idea for how prior approval, preferred drug lists, and step therapy might grow beyond the point where they’re actually helpful. So, here what you’ve got is a health plan. It has revenue. It has total costs, and the difference between them drives pressure to cut costs. So, if for example, total costs go up, revenue hasn’t changed, let’s say. So, that means there’s more pressure on the managers to cut costs. One of the things they do, have done, continue to do, is they take even more drugs off the formulary, they tighten up the requirements for step therapy, and they put more drugs and tests and procedures on the list of things that require prior approval. The goal of this is to keep clinicians, keep your doctor from prescribing things you don’t really need, or prescribing things that are where there’s a cheaper, pretty close substitute. Uh that should lower the unit cost of the drugs and procedures, and that will lower the total drug and procedure costs, and that lowers total costs. Fantastic. That’s what we call a balancing feedback loop. This is a feedback loop that helps control costs, and it is the motivation for all three of these uh policies, PA, PDLs, and step therapy. So, the question then is, that’s not the whole system. Think of it as an iceberg. This is the part managers of health plans see. It’s what’s above the waterline. And what’s below the waterline, well, that’s what I want you to tell me. So, use the chat and just quickly drop in any unintended, so-called side effects that might arise from the use of prior approval. So, go ahead. What would happen? Prior approval means it’s going to take a while before you can get what you need, if you can get it at all. Preferred drug list means you’re not getting exactly what the doctor thinks can be best for you, and so forth. So, yeah. Uh And some of you may be providers, meaning in health care as a doctor or a nurse practitioner. And some of you may be on the payer side, meaning you work for an insurance company. So, yeah, it’s there’s some fantastic ideas here. Extra visits to the doctor, administrative costs go up. Person dies waiting. Hopefully, that doesn’t happen very often, but the data show it does some. Um Uh And uh oh, litigation. Somebody mentioned litigation. Interesting. Uh staff don’t care and deny with good reason, without good reason. Wow. Uh that sounds like somebody who’s experienced this. More time on paperwork instead of patient care. These are fantastic ideas. Keep going. There’s a lot here. Let me show you just a few cuz we have limited time here uh how that might work. So, um So, what’s below the waterline here? Well, one thing, and you many of you pointed to this is the appropriateness and the timeliness of care is going to drop. So, think about how this works. Costs go up, pressure to cut costs, we tighten up these requirements, more PA is required. It does have the intended effect here of lowering unit cost of drugs and so forth. But, then because not everybody gets care soon enough or the right care, the timeliness and appropriateness of care declines, health outcomes deteriorate. People stay sick longer, some deteriorate, etc. Well, that means they have to go back to the doctor uh and that’s going to raise total costs. And when they go back to the doctor, they’re going to need more tests, more treatments, and more procedures. That raises the total drug and procedure costs. So, the unit cost went down, but the total number went up. All of those things raise total costs further. Now, [clears throat] you can see how that undermines the intended effect, but it’s worse than that because total cost just went up more for reasons that are typically not perceived. And so, pressure to cut costs increases further, and the reaction is, well, we need more of what we know is working. Well, what we think is working is PA, PDL, and step therapy. So, we’ll tighten up those requirements even more. Care appropriateness and timeliness falls further, people get worse, more doctor visits, more tests and treatments, higher costs, and around and around we go in a vicious cycle, a reinforcing feedback loop. Well, what else? Well, you had some fantastic other ideas in here. Let me show you a couple that I thought uh and many of you mentioned this. So, total costs go up, we tighten everything up, care quality and appropriateness declines, people get worse, doctors can then appeal. Doctors can go to the payer, the insurance company, and say, “My patient really needs this drug, this procedure.” And that they might win, often they don’t, but whatever the outcome, it raises administrative costs. That raises total costs. There’s another vicious cycle. In addition, and some of you mentioned this, people who are seeing their health deteriorate, they’re going to end up in the emergency department more. Some of them are going to be hospitalized more. Some of them will have to be readmitted to the hospital after initially uh making a little progress. All of that raises costs some more. Another vicious cycle. Uh and you could have more litigation, more malpractice suits, and somebody, at least one person, actually mentioned this in the chat, same thing, it creates another vicious cycle, raising costs. Now, one of the things about systems thinking is you always want to challenge the boundary of your model. So, here is revenue, and so far revenue is just out there, it’s just happening. That’s almost never true that there’s something that just happens to you, but whatever you do doesn’t influence it. So, maybe you can think of how what’s happening under the waterline here will feed back to revenue. Here’s one idea. So, care and health outcomes deteriorate, people are unhappy, and they can switch health plans. And if oops, if that’s an error, the number of subscribers should go down, and then revenue would go down. That’s another vicious cycle that’s going to undermine the financial position of the health health plan. Uh so, does any of this really happen? Of course it does. That’s what the data clearly show. And are health plans figuring this out? Well, it’s been at least 40 or 50 years since formularies and prior approval and so forth have been implemented. And some health plans that are better at systems thinking have actually moved away from them, but many are just now realizing that they need to move away from prior approval and so forth. Here’s two recent examples. Uh Blue Cross Blue Shield, one of the biggest uh payers in the United States, uh announced that they’re cutting back on prior authorization, and you can see why. Their their argument is it’s uh it’s too expensive, and it’s going to ease administrative burden and speeds speed access to evidence-based care. And it’s going to be better for patients and for them. Uh United Healthcare also announced just uh in May, this month, that um they’re cutting prior authorization requirements by a substantial amount. Um So, those are kind of late to the party, but many other health plans are better at systems thinking, were better at seeing what was under the waterline, and have already been cutting back on these three practices. Uh not to zero, but uh relaxing them uh significantly. So, doctors have the knowledge, they’ve always had that knowledge, uh but then can apply that knowledge to get you what you really need in order to recover in the most effective way. So, that’s one example. Let me take a few minutes now, and let’s talk about uh a completely different example. I mentioned project management. Um and uh how projects typically don’t work very well. But, let’s find out what you all think. So, you know, in just a few seconds, we’re going to launch a poll. And uh here’s what the poll is going to ask you. Projects in my organization are on time, under budget, deliver high quality, and delight our customers. And your job is to pick one of these options about that statement. Always happens in my organization, usually happens, sometimes happens, rarely happens, never happens, or no comment, we’re still in active litigation. So, go ahead. Let’s launch that poll. And everybody go ahead and select the option that is closest to the experience in your organization around projects. Any kind of project. Software, administrative change, new product development, construction projects, whatever it might be. So, give you a second. Oh, don’t put it in the chat. Put it in the poll, folks. I got a usually, I’ve got a rarely, but um let’s answer the poll. And Molly, Charlotte, whatever you think you’ve got a pretty fair fraction of everybody responding, go ahead and publish the results. Okay, looks like sometimes and rarely, which is not great, are the most popular answers. And almost nobody says always. And

[snorts] uh a fair number say never. And we did have one no comment. Okay. So, that’s pretty interesting. Sometimes, rarely, in our experience, that’s maybe uh even a little on the optimistic side. So, how do we help people learn how to manage projects better? Well, lectures don’t work. One of the things we know from many decades of research on policy change and communication is Well, let me put it this way. Research shows that showing people research doesn’t work. If I tell you what the right thing to do is, you’re not going to do it. It’s not going to change your mind. Those of you who are parents and have children who are or were teenagers, you know this very well. You can’t tell them anything. Certainly this was true with with our kids. So, the way we try to get at that to develop your systems thinking skills is to use management flight simulators. Put you in the system in a simulated environment where you can try whatever strategies you like costlessly. If you go bankrupt in a management flight simulator, nothing actually happens. If you do great, well, then you can try to experiment and see how robust your good strategy is. Flight simulators uh the analogy is really very good one. You don’t want to get on any commercial flight unless those pilots have had significant simulator training so that they know what to do if there’s an emergency without having to go, “Oh, I’ve never experienced this before. Now what do I do?” Because there isn’t going to be time for that. So, let’s actually try it out. Um I’m going to show you really quickly one of our simulators here uh on project management. Going to do it super fast. I’m going to give myself a screen ID. Uh and we’ll start the simulation. Now, when you start the simulation, and we’re going to do this in the exact class, you can manage a construction project, a hardware project, a software project. You can also, if you um uh get credentialed as an instructor, set up custom projects. Let’s do, oh, I don’t know. Let’s do a hardware project today. So, I’m not going to read this. You don’t have time to read it. It’s not that important for our demonstration right now, but there’s a description of the setting, and you play the role of the project manager for a project for a new consumer electronic product like this VR headset here. We’re going to enter a name for our project. We’ll call it um kill the competition. And then we’ll begin. Now, it’s pretty complicated and quite realistic, carefully grounded in many, many decades of system dynamics research on projects, proactive as well as forensic context. Um so, don’t worry about all the detail here, uh but here is a lot more information than you might actually have to manage your project. There’s two phases in this project. Real projects are of course more complex, but we have two phases. We have high-level design in blue here, uh scheduled right now to last 24 weeks, and then there’s detailed design. Uh and so, high-level design is meeting with the customer, what are the customer’s requirements, what are uh some of the main constraints on size, on weight, on power usage, uh heat dissipation, whatever it might be. Um and then those high-level product architecture decisions, uh drawings if you like, get handed over to the detail designers who have to code up the custom chips, who have to uh produce the plastic parts and the the molds for them for for injection molding, etc. And and produce uh a working product prototype that passes all the safety and other tests and can be handed off for manufacturing. So, you can change all this. One thing I’m going to do here is I’m going to say, as is common, uh let’s actually overlap. We’re going to start detailed design before the um high-level design is finished. That’s very common. Uh and then uh what else are we going to do? Well, over here we’ve got some project planning uh, help. A little spreadsheet that tells us, well, there’s 72,000 tasks. Now, a task here means what one experienced worker can do in one person-hour of effort. So, we’re going to need 72,000 person-hours of effort to do phase A, the detailed design, and 33,600 to do the uh, sorry, high-level design and detailed design over here. Given the current schedule, uh, and the productivity, uh, that means we’re going to need, um, 75 people upfront, and that’s what I’ve got here. Uh, but, you know, budgets are tight. I need to come in under the budget, so I’m going to change that to 70. Um, and right now there’s nobody in B because it hasn’t started yet. You have other decisions you can make. You can, for example, uh, respond when the marketing department comes to you and says, “Hey, our competition is launching a new feature that we haven’t thought about. We need to add that to the project scope.” So, marketing will propose, you get to decide, are you actually going to do that or not? You could also reduce the project scope. Some features you you might decide are bells and whistles you don’t really need in order to hit the timeline and the, um, and the, uh, budget. And you can also tell people to speed up work by overlapping tasks that really ought to be done in a certain sequence. In addition, you can apply what we call management pressure for progress and for quality. Right now, they’re neutral, but you could decide, you know, let’s pound on the table a little bit and tell people they have to work harder. Or, let’s tell them, “I don’t care what it takes, make sure you do it right.” All right. Well, that’s a very quick intro. There’s a lot of detail here, but, uh, let’s get started. And because it’s just a demo, I’m going to advance uh 1 week to get started. And now, let’s look at the dashboard. Where are we? Well, we’re just getting started on the high-level design. And let me advance more than 1 week here just for time. Uh so, let’s advance 4 weeks, and people are working. It’s going great. Everything’s good. Productivity is high. Work week is tolerable. Everything’s great. Marketing’s already telling us we need uh almost 2% more work to do to meet what the competitors are planning. Let’s go forward some more. Everything’s still looking great. More new projects. Okay, let’s let’s do that. Let’s accept all the all the scope. Uh and that’s estimated to boost our final market share when this product is released by 3.6%. Fantastic. Um Okay, things are going great. We’ve we’ve actually done some of the work. A lot more has been initially done, but it awaits testing. Um and things are going fine. So, let’s just keep going. Oh, now we’ve got to hire for for B. Well, what do we need here? We need about 28 people for detailed design, but we don’t get that many or we’re not going to hire that many because we’re trying to control costs. And uh and let’s go forward. Uh-oh, look, we’ve gotten some emails. Uh so, QA says, “Hey, we’re doing great work on our team. Uh we might get the best quality award.” And then you can respond or not. I’m going to respond saying, “Yeah, this is fantastic.” Uh what’s the other email? Let’s um let’s go back here. We’re doing great. We have a cohesive and collaborative team. Hey, thanks. That’s terrific, HR. We appreciate it. Uh now, let’s go back here, and let’s advance some more. And we’re starting to do the detailed design work, but look what’s happening. Um the work week is up to 60 hours now. We’re actually at the end of the scheduled high-level design phase. It should have been done by now, and it’s not. We’re behind. So, I’m going to apply some management pressure. You guys need to work harder. And quality is really important. Quality is job one. Don’t forget. So, do it right, but get it done. So, uh And we’re already up to 61 hours in the detail design phase. Let’s keep going. Oh, we have some new emails. Uh So, uh Engineering told us the preliminary design review did not go well. Senior management is not happy. Too many defects in the design requiring rework. So, please fix this. And now you get to decide how you’re going to respond. I’ll make changes to get back on track. Or hey, executives are never satisfied. I’m confident that you all in engineering can handle any rework. So, you can decide, and you don’t have to be honest about it. You can say, “Hey, we’re going to make changes, and uh and we’ll we’ll we’ll send it.” But we don’t have to do that. Um Here’s about the new product uh the uh features that we did uh accept already. And then here uh we’re very concerned about the amount of rework. This is QA telling us uh So, I’m going to take steps to correct it, but really I just resent your interference. Um Too many complaints about long hours, HR is telling us. So, I’ll stay on top of it, but you know what? We got to get the work done. So, let’s see that what else is happening here. So, we’re uh you know, we still have quite a few weeks to get this thing done. We’re at week 29. We have till 48. Let’s keep going. The work week is uh really high. We’ve now finished high-level design. That’s great, but we’re way behind on detailed design. And if we go here, instead of the 28 people, uh, that we thought we needed, now we need 51. Well, I’m not going to hire 51, but let’s let’s let’s boost the workforce here. Uh, we’ll go up to 45. And let’s go another few weeks. Is that helping us? Well, um, things are not looking too great. The work week is even higher now. 78-hour work week on average. People are getting burned out. That’s probably what we’re getting emails say. Urgent, competitors. Oh, there’s new new, uh, features they want. Design review is still not going well. I’m not answering my emails. I’m too busy managing this project. So, um, how are things going now? I think maybe we need even more people. Uh, oh, now we need 166 people in detailed design. Well, I’m going to go up to 120. Um, but, uh, that’s it. I can’t I can’t and, you know, and [clears throat] even more pressure for progress here. And we’ll advance some more. How are we doing? Uh, well, let’s look at the dashboard. We’re only about halfway done with the, um, detailed design and we’re almost at the deadline. Well, let’s keep going. Uh, I haven’t added the new features marketing wants except for the very beginning. You know, maybe we should reduce some project scope. Let’s let’s try to do that. Let’s reduce project scope by, uh, maybe 50% of the non-essential features. Whoops. Oh, no, can’t do that. It won’t let me. All right, got to keep all those bells and whistles in there. Everybody needs to overlap work, do it out of sequence if necessary, but just get it done. And, uh, now how are we doing? Well, we’re at the deadline. We’re over the deadline. And whoops. We uh This is the message from from the CEO. Uh How did our project come out? Well, we went over budget. We took way too long. And our product quality is horrible. The goal was less than 1% defects. Uh Which we can handle with warranty. It’s 25%. This is a product that’s going to fail. Does fail in the marketplace. And the net present value of our losses is about $50 million. Uh So, this is a pretty horrible project. Now, the way these things work is you would have more time to do this. And you would probably work with uh partner or two to think about what your strategy is. And then you could try it and see what happens. And then play again. You can go back and start the game again. And see if you can come up with a better strategy. What you just saw here is actually quite common. And uh a significant amount of research shows that this kind of outcome is typical. And often lead to outcomes that are far, far worse than what you even saw here. Including very commonly disputes between contractors and customers. Between contractors, subcontractors, and suppliers. That often escalate into legal action. And well, I’ve been a expert witness in half a dozen of these uh lawsuits or arbitrations over projects gone bad. And the lawyers are the only ones who really win at the end of the day. System dynamics has been used extensively to uh improve project management. And including the widespread use of simulators like this. So, let’s um let’s go back and talk for a second uh about um how system dynamics has been used. Those were two very quick examples. Um and uh let me just show you a couple of there’s a very small subset of some of the successful applications in the mostly the business world of uh of these tools. And it it spans all kinds of industries, uh high-tech, autos, biopharma, drug development, construction, chemicals, project management in every kind of industry you can imagine. Health care, many, many examples. One of my favorites is uh an alum of the Sloan School used system dynamics to develop uh a formal simulation model and a if you like a little bit of a flight simulator to help the front-line clinicians improve uh the care that kidney dialysis patients received, in particular to treat the anemia that is a very common consequence of being on dialysis. Saved millions of dollars and many, many lives. And many, many examples in energy policy and climate change, public health, and many, many others. Um In the program in the executive course that we run, we we devote the final day to uh presentations from folks in the business and other sectors who have actually used these tools successfully. Not to toot our horns, but to talk about the real issues that come up in implementation. How do you actually do this in organizations where people are very, very busy and don’t have uh a PhD from MIT in in system dynamics. So, that’s a quick tour with two different examples of how systems thinking works, just a couple of the tools, how they’re used, and um I think this is a great point to turn it over for questions. So, questions, comments, challenges, what have you got? I think Molly is going to curate the questions, is that right? Yes, great. Um if you have questions, please do put them in the Q&A. Um it should appear at the bottom of your Zoom screen, and we have a few already, John. Uh one specifically about the simulator. Does the simulator provide feedback in the form of recommendations on what decisions were suboptimal? Um and this is from Chris Sanford. Yeah, it’s a great question. And um our philosophy is uh as I mentioned, telling people the answer, telling people, you know, here’s the optimal strategy, here’s what you did, doesn’t work. The mental models people hold about complex systems, including projects but many, many others, are very, very powerful, and they’re often strongly reinforced by our everyday experience. You know, just go back to the prior approval healthcare case, in the short run, tightening up the rules for what needs to have prior approval, it does work in the short run. And so, you get very powerful feedback that that was a good idea. That helped. The unintended consequences, they flow through the system often with long delays, and they appear to be things that happened that weren’t connected to the initial decision to, for example, require more procedures to receive prior approval or uh more drugs taken out of the preferred drug list. And so, people don’t get good feedback. And what that does is it causes their mental model that’s damaging to be strongly reinforced. And and you know, if if an expert comes along or a simulator comes along and says, “Well, if this is better.” People just it’s too much dissonance. Most people are just going to say, “Well, that’s not right. Your model is wrong.” And listen, folks, all models are wrong. Some models are useful as George Box, the famous statistician said, “But all models are wrong. They’re not the real system.” So, it’s super easy for somebody whose intuition says what they just told me is best can’t be right. It’s super easy for them to say, “I don’t That’s wrong. And I know it’s wrong. And your model’s too simple. Your model’s too simple. Your model’s too complex. It’s a black box that I don’t understand.” Or “You don’t know my organization the way I do.” So, uh our philosophy is you can’t tell people they have to learn it by doing it in the safe space that the simulator provides. And that includes challenging the simulator’s assumptions. In many of the simulators, you get to change the assumptions. Um our climate policy simulator called En-ROADS developed with Climate Interactive is a great example. There’s a lot of uncertainty about the economy, the energy system, and the climate. And so, you get to change the assumptions. They’re based on the best peer-reviewed evidence, and we think we can defend those assumptions as grounded in the best scientific knowledge, but we can be wrong. And you might not believe that. So, you get to change the assumptions and see how that affects what we might do to try to reduce the harm from climate change. What else, Molly? Uh John, we have another question specifically on the simulator. This is from Mark Pellerine. How does the simulator factor in externalities that impact from outside the closed system of the corporation or organizations centric setting? For example, um, supply shocks, energy costs, regulatory changes, black swan events. Yeah. Does the simulator have a business continuity mode? Yeah, fantastic question. So, this one that I just demoed so quickly, uh, does not do that. It would make it even harder. I mean, and in many of our simulators, we do include surprises, sometimes positive, more often negative ones like, oh, the Straits of Hormuz are closed, and so your supply chain has been completely disrupted. And, oh, food prices are going to go up because fertilizer is less available and is becoming more expensive, so crop yields are going to go down. Um, we do include those in many of the simulators. Of course, nobody can predict the exact form that a black swan event or a uh, negative surprise might take, uh, but you can simulate many of what those would be. And in fact, that’s a very important use of simulators in particular context. You know, go back to the flight simulator. When a pilot is learning in a flight simulator, uh, you know, you don’t just simulate, oh, you know, a standard routine flight from Boston to Chicago where nothing much happens. What happens is the folks in the control room, uh, you know, partway through the flight, they’re going to give those pilots an emergency that they have to deal with. Engine failure, explosive decompression, um, cockpit instruments go dark, uh, bird strike, right? You all remember the Miracle on the Hudson? Well, Sully Sullenberger, fantastic, amazing pilot, uh, with a lot of experience, including a lot of experience in simulators. In fact, he was a certified simulator instructor. And after that amazing successful ditching in the river with nobody killed, uh one of his comments was “Well, you know, I didn’t I’ve never experienced a bird strike that knocked out both engines, but we practiced that in the simulator many, many times.” So, that’s one of the principal uses of um of these management flight simulators. Polly? Yeah. Um a specific question from Daniel Schneider, are there differences in applying this in the long-term versus short-term versus ultra-short-term situations, for example, trading? Yeah. Um also a great question. So, the simulators are developed to meet a particular need. And uh the time frame for those varies. Sometimes, as with our climate policy simulator, the time frame is how do we protect ourselves from the worst harms of climate that are coming at us under current policies um between now and the end of the century, because that’s the time scale under which these dynamics play out. Actually, it even takes longer. Others are very short-term oriented. Um I’m not aware of any publicly available simulators that would help traders in the commodity pit, say, trying to deal with the turmoil in the oil market, for example. Uh but I do know that uh there have been applications of system dynamics in high-frequency domains of that sort. So, it all depends on the problem you’re trying to solve. There’s no intrinsic limit on uh the time frame. Uh John Steve Hein and takes us beyond simulators. So, other than simulators, do you suggest organizations in developing their plans identify assumptions and other factors that might affect their basic planning and how they might modify or impact those plans? Yeah, absolutely. Uh, fantastic question. Not every project uh, requires uh, the development of a formal simulation model or a flight simulator. A lot of value can be and has been developed by uh, the qualitative use of the systems thinking tools like the mapping tool that I illustrated with the prior approval example. And often that delivers a lot of value for people and in two ways. One way is the substantive example, people gain deeper insight into what’s going on. And to do that, you need a what we call a group modeling process where you expand the possibly expand the number and diversity of the people who are brought into the process. The uh, the saying here is you want the system in the room. You want to have all the key people, all the actors from all the important subsystems, many of whom are not even perhaps part of your organization, some of whom might be your traditional adversaries, um, you know, consumer groups who are protesting that your products are not made in a sustainable way or with um, with a living wage for the uh, folks way upstream in your supply chain. You want those people in the room so that collectively they’re much more able to identify the so-called side effects that are leading to the persistent failure of some of the initiatives you’ve tried in the past. Those qualitative tools are extremely valuable. Uh, I don’t think we have time, but I have some stories from my personal experience working uh, with organizations uh, on exactly this point. Uh, and and so that’s the sort of direct impact. The The indirect or meta impact of the qualitative tools is people develop their systems thinking capabilities, including their ability, if they work at it, and it often requires coaching, to, um, to listen, um, to listen to others before jumping to conclusions, to, um, to work collaboratively, to, uh, to focus on activities, uh, instead of saying, well, I’m the expert on such and so, uh, taking a more humble personal stance. And, you know, that’s not easy to do. It’s not easy for a faculty member, you know, when you’re in the classroom, it’s a it’s a social setting that kind of puts me literally up on a pedestal, on a dais, where I can see everybody, and I’m supposed to be the expert. That’s a terrible thing. I’m not the expert. I’m here to help catalyze your learning. Uh, I’m you know, we we need to get away from what one famous paper in educational practice called, we need to get away from the sage on the stage, and instead the role of the facilitator, and your roles, is to be a guide on the side to catalyze the learning for the group, for other people. That’s an important systems thinking capability, and we have methods to help people develop those. Molly, do we have time for more, or do we need to wrap it? Perhaps one more question. Um, I want to answer. What do you Okay, great. Uh, so this is from Eric Kebshell. Eric, thanks for joining us today. If presenting the data alone is not enough to convince people to change, how does the simulation modeling change the mental models that tend to be entrenched in the status quo? Oh gosh, it’s a fantastic question. I don’t think there’s a real short answer. I will say data are essential. It’s critical to have data, quantitative data of all kinds, but also qualitative data. You need to walk around. You need to go what we call go see and assess. You need to be out there in the field. So, it isn’t just the numbers. Um but necessary but not sufficient. You also need those other processes that I alluded to. I don’t think I have time to give an example, unfortunately, but there are many and these are some of the themes and issues we talk about in our programs. Sorry, that’s unsatisfying, but Okay? Yes, we are. Um I think we’re ready for the closing slides and can share more about the course. let’s do that. Thank you. So, as we mentioned, Professor Sterman is faculty director of our long-running program Business Dynamics. This is a 5-day in-person program at MIT where you dig into all of these topics, multiple simulations with a group of peer executives from around the world. So, imagine spending a whole week on campus with John. It’s really lovely. Our next Business Dynamics is running in a few weeks. It will be June 8th through 12th on campus. We still have seats available, so if you’re interested in joining us, please scan the QR to see more about the course. We are also scheduling our Business Dynamics for next year. This will be June 2027. Please check the course site in the coming days and we’ll have those dates up. If you have any questions about the course, Uh there’s contact information on the page. We’re very happy to speak with you. Um thank you again for joining us, John. Thank you so much for today’s lecture. It’s a pleasure. Thank you, everybody, for participating.