The Management System Your Organization Doesn’t Know It Needs 

Here’s an uncomfortable thought: most organizations are addressing only a fraction of the problems they actually have. By problem, I simply mean a gap between what is and what should be happening. And worse, leaders in those organizations don’t even know it. 

I’m not saying they’re ignoring problems. Most people do what they can to fix known problems. I’m talking about something more fundamental. Their systems for detecting, surfacing, and responding to problems are dramatically under-built. Problems that should be caught in minutes, if not seconds, persist for hours, days, weeks, or longer. Patterns that should trigger root-cause analysis go unnoticed. Deeper issues that require new processes and capabilities — a matter of strategy — fester for quarters, years, or until it’s too late and the business is lost. 

I believe the answer lies in understanding what a management system actually produces — and how it relates to other systems it serves. For this piece, I’ll focus on the production system. I’d like to share with you my current thinking about useful ways to understand these two systems. 

A production system is how an organization makes and delivers its products and services. It includes the flow of material and information, the standards and methods that define the work, and the means of interaction when multiple people (and now agents) are involved. The Toyota Production System (TPS) is the most studied example, which implies it’s the most effective. I certainly think so. 

A management system includes how an organization monitors performance against standards and targets, understands when things are going off-track, mobilizes people to respond, and facilitates the ensuing actions. It includes things like detection devices and signals, visual boards and other sources of key information, paths of dissemination and escalation, huddles and reviews, and, of course, the means of response. 

Put together, a Toyota-style production and management system integrates all these elements for ongoing cycles of PDCA. While most people are familiar with this concept, familiarity isn’t the same as lived experience. A vast majority of us have not worked in effective, much less integrated, versions of lean systems for production and management. 

For nearly 30 years, LEI has been experimenting with ways to grasp and explain how these systems work, individually and together. Most recently, my colleague Mark Reich offered a management system framework in his outstanding book, Managing on Purpose, drawn from his 23-year experience at Toyota as well as working with dozens of organizations on behalf of LEI. What follows is my attempt to describe the relationship between production and management systems more clearly and usefully. 

Consider this parallel. TPS is a system for making and delivering products and services that simultaneously develops craftsmen through continuously improving their work. What if a lean management system is simply seen as a system for making and delivering countermeasures to abnormalities and delays — anything disrupting the flow — that simultaneously develops problem solvers through continuously solving problems? 

If that framing holds, it opens up some interesting design questions. 

Start with demand. In TPS, demand comes from customers. It’s what they’re asking for; in what condition (price and quality), at what rate. That rate, called the takt time, becomes the heartbeat of the system. At the pinnacle of TPS execution, everything that’s moved or made is synchronized to it. It enables continuous flow. Then, when something disrupts that flow — a deviation from the standard — the system’s problem-detecting capability kicks in with a signal: an andon. In this sense, the system is designed to see problems. Lots. In a Toyota factory, hundreds if not thousands of problems are detected every day. 

Now apply that thinking to a lean management system. Who’s the customer and what’s their demand? At least one customer is the production system, and what it’s asking for are countermeasures to its myriad problems. Not in the abstract, but specific abnormalities from safety and quality standards, as well as delays from timing expectations. The vast and detailed precision of a lean production or operating system makes them easy to detect, bringing me back to the management system. Its mechanisms for monitoring performance connect the two systems. 

When the mechanisms show “green,” indicating the absence of abnormalities and delays, the management system is running but it’s not producing. You could say there is no demand. When things go “red,” however, the value-creating work of the management system engages: signaling, sorting, disseminating, perhaps escalating, organizing, diagnosing, facilitating, counter-measuring. 

This is where the capability of each system, as well as their relationship, becomes critical. The more developed your production system — the more rigorously it establishes standards and the more sensitively it detects deviations — the more demand it generates for your management system. A well-tuned production system is constantly surfacing problems. It’s fragile by design. And that’s a feature, not a bug. But only if you have a management system with the capacity to respond. 

A well-tuned production system is constantly surfacing problems. It’s fragile by design. And that’s a feature, not a bug.

Here’s the trouble: many organizations build management systems without production systems capable of generating meaningful demand. The result can be management activity organized around random things rather than genuine problem-solving demand. Rituals in front of boards make people busy, but they’re often not connected to the work, much less its reality. Moreover, board reviewed once a day are as useless as yesterday’s newspaper. A lean management system worthy of the name has the capacity to detect and respond to problems as close to real time as the work demands. 

Even organizations with strong production systems can have a gap. Think of it like a coffee shop: actual customer demand includes the person who looks through the window, sees a long queue, and walks away. You never knew they were there. In any operating environment, actual demand for problem solving will always exceed what the system detects. The question is: by how much? And what are you doing to close that gap? 

Art Smalley’s Four Types of Problems framework is useful here. Type 1 problems (troubleshooting deviations from standard) are numerous and should be handled by people closest to the work. Art has said, perhaps provocatively, that 90% of all problems are Type 1. This explains why Toyota has one team leader for every four to six team members. The company has learned this ratio provides enough capacity given the demand. Higher ratios would mean accepting injuries, defects, and late deliveries. But Toyota pursues perfection. 

At the other end, Type 4 problems requiring genuine innovation in the form of new products, processes, and capabilities are vastly less numerous but much higher stakes, demanding higher orders of leadership and collaboration on research, design, experimentation, and engineering. Think hoshin kanri and lean product and process development. 

It’s worth noting that the demand logic I’ve described applies most naturally to Type 1 and Type 2 (gap from standard), and arguably Type 3 (target condition), where the production system surfaces problems and the management system responds. Type 4 operates on different logic. The best Type 4 work, creating something that doesn’t yet exist, isn’t triggered by a deviation from standard. It’s triggered by vision and strategic intent with a long-range mindset. Think of the Prius: Toyota didn’t detect a gap in customer demand for hybrid powertrains. They imagined a future and built toward it. A complete management system needs capacity for both: responding to what the system reveals and pursuing what it can’t yet see. 

A well-tuned production system is constantly surfacing problems. It’s fragile by design. And that’s a feature, not a bug. 

A well-designed management system has differentiated capacity for each type. Standardized work and team leaders create rhythm and capacity for Type 1. Hour-by-hour boards and daily, weekly, monthly, quarterly, even ad hoc huddles with cross-functional participation address Type 2 and 3 problems. A high-performing management system puts all this demand for problem solving, continuous improvement, and innovation into relationship with organizational capacity and capabilities throughout the organization: top-down, bottom-up, end-to-end. Combined with a high-performing production system, it creates the outcomes engine every leader desperately wants. 

I don’t claim to have this explanation fully worked out. But instead of asking “Do we have the right practices in place?” this concept suggests we might ask: “Is our management system engineered to detect the full range of deviations we face and produce countermeasures at the rate those deviations demand? And is our production system developed enough to generate the demand signals that make the management system’s work meaningful?” 

This week, about 500 people will gather at the Lean Summit in Houston to explore what it takes to build lean enterprises. I’m talking about complete business systems, not partial toolkits. As I look at this year’s speakers, I’m struck by how many are addressing multiple facets of this very question, from frontline team leader systems to executive-level management systems, while simultaneously integrating AI and other emerging technologies. How do we build organizations that see more, respond faster, and develop people as required in the process? 

If this perspective provokes something in you, I want to hear it. Perhaps you vehemently disagree, or agree with caveats, or have alternative frameworks for understanding management systems. The best ideas in lean have always emerged from the collision of theory and practice. Let’s keep colliding.