The most common problem observed in companies adopting lean practices is focused projects led by lean experts that produce stunning results on paper, however, within three to six months self-destruct. It is natural for managers and executives to ask the simple question: “How do we sustain the gains?” The question is logical, but the underlying assumption is the reason these efforts will continue to self-destruct.

Let’s consider the underlying assumption of someone who is trying to figure out how to sustain the gains from their great lean projects. We lay out an image of how the process looks over time from this person’s viewpoint in Figure 1. We take the time to hire a consultant and apply the tools and learn how they work, internal experts spread the tools to lean out the operation and now we want to sustain the gains. What methods should be used for this sustainment phase? The underlying assumption is that the organisation is a type of machine. Like a machine we want to upgrade it so it works faster and more efficiently with more power and once upgraded it requires maintenance. Sustaining the gains is a search for the correct maintenance procedures. This usually involves a lot of auditing, finger pointing, and policing of the new lean methods.

This model of lean management is only a revamping of Frederick Taylor’s scientific management, or the 20th century command and control paradigm.

Figure 1

Perhaps it is kinder and gentler for managers to ask opinions of employees, but the result is still the same. The top wants results. The experts work with middle management to blast through the organisation with many changes using lean methods. And the top has their hands out waiting for the results.

What actually occurs is quite disheartening. As shown in Figure 2 sustaining the gains turns into major slippage back toward the prior operating condition (Liker and Franz, 2011). Then we want to know “why are we slipping”? In systems theory terms any system taken out of equilibrium will over time try to revert back to its prior steady state. This is particularly true in human systems when the change has brought us to a higher level of performance.

Figure 2

Higher performance means we have new systems and procedures that require more human discipline, focus and therefore energy to maintain. Systems theory uses the concept of dynamic homeostasis, which can be viewed as a hamster on a wheel running like crazy to stay in place. What appears to be a steady state is actually an awful lot of work improving things because the world is constantly changing and throwing challenges our way.

Figure 3

The world simply does not like order, as it takes work to maintain order, everything, including machines, naturally degrade unless there is constant energy to maintain the new higher level of performance (see Figure 3). In fact what really happens in any dynamic environment is that the solutions (countermeasures) of yesterday become inadequate for the situation today. As Gorbachev noted when he was arguing for nuclear disarmament “we need to be constantly improving or we will slip backwards”.


So what is the answer? We must first ask what we are in search of. Is it making a lot of quick changes and then sustaining the gains or is it continuous improvement? Almost all organisations that adopt lean claim they want continuous improvement.

Anyone who is unhappy with their body image or health, or is concerned about their future health, understands there are two things they can deliberately do to reduce the effects of entropy on their bodies. They can eat healthily and exercise regularly. Of course there are many approaches that shift your weight up and down like a yo-yo, such as diet pills or the diet fad of the month. You get addicted to the pills, which cause their own health problems and eventually that crazy diet is ruining your lifestyle so you regress back to old habits. The issue is not sustaining the gains, but fundamentally changing lifestyle.

In his book Change or Die, Alan Deutschman tells us a great deal about the difficulty of this change process. Patients approach doctors and are told they need to change their eating and dieting habits fast, or die. They can provide endless statistics on the likelihood of death and an estimate of when they will die if they do not change. There is a crisis, which we all know motivates change, except 90% of those people do not change and die soon after. Why? Because what the doctor is trying to change is decades of habit – routines that are deeply rooted in our brain. Unlearning deep-seated routines, and replacing them with new ones is difficult. Deutschman concludes that the process the doctors are using, the 3Fs of facts, fear and force, are too blunt an instrument to make the fundamental changes needed in our way of life. Force could work, for example, if the doctor followed the person constantly holding a loaded gun, but who has the time and bandwidth to do that 24 hours a day?

What Deutschman then experimented with was a different approach that actually had a 90% success rate. He calls it the 3Rs and it fits the model used by Alcoholics Anonymous:

  • Relate: to a mentor and peer group;
  • Repeat: deliberately practice the new habits in small bits every day;
  • Reframe: begin to develop new ways of defining a good life.

The conventional approach to selling people so they will sustain the gains looks a lot like the 3Fs. Tell them why it is necessary to change, threaten them if they do not, and constantly audit using rewards and punishments to force people to change. It works if management is constantly monitoring people and taking immediate action any time they deviate from the standard. But even then, do not expect any innovation, and continuous improvement quickly becomes an empty motto.

When we look at Toyota’s approach to developing people it looks much more like the 3Rs. A sensei or coach is assigned to develop the person. They put the person in situations to practice the new way of working. They challenge them with stretch targets and, perhaps most importantly, they provide corrective inputs regarding procedure. They are relentless about asking them what problem they are working on, what they will try next and by when, what they learned from what they just tried (Deming’s plan-do-check-act cycle). They may ask them to generate 40 ideas to improve a specific process or to stand for hours and observe a process and describe what they see. These are all exercises – ways to deliberately practice a specific skill.

To formalise what Toyota leaders were trying to teach they developed Toyota Business Practices (TBP), an eight-step method for improvement. What Toyota is trying to achieve conceptually looks like Figure 4. The starting point is always a clear picture of the purpose. Any Toyota manager can tell you about how tired they got of being asked over and over what the purpose of the activity was. The purpose should not only be to achieve some result for the business, but also to develop people in the process.

Figure 4

To understand the purpose and the general direction of where you need to go requires spending some time at the gemba, observing and understanding. Then TBP asks you to define the ideal state of the process before jumping into any specific problem area. The gap between the ideal state and the current condition defines the scope of improvement issues you might undertake, but there are too many issues and the gaps are too large. So you must “break down the problem” to define nearer term targets to begin concrete activity. This may be a one year target, but even that must be broken down to shorter term projects, perhaps month by month, to practice PDCA.

The shorter the loops of PDCA, the more rapidly we can learn. Learning our way to a target through small experiments is a fundamentally different proposition from planning our way to a major system change. As Dwight D. Eisenhower aptly put it: “In preparing for battle I have always found that plans are useless, but planning is indispensable”. The process of planning and trying to predict and anticipate the future is very worthwhile, as long as you realise that any predictions are simply guesses and you must adapt as you experience and learn.

While some people learn alone, more commonly people at Toyota are parts of teams. For example, the Toyota work group in the factory has a group leader (supervisor), team leaders (hourly peers who have advanced farther in their development of improvement skills), and team members. Within this group of other team members, team members who have coaching ability and have been appointed team leaders, and the group leader as chief coach, there are daily opportunities to relate, repeat and reframe. Daily meetings in front of visual management boards, called the Floor Management Development System (FMDS) at Toyota, are one place to see the challenges we face as a team, target versus actual, the progress we are making, and what specific actions we are testing along with the results compared to our predicted results.

Unfortunately few organisations outside of Toyota have the breadth and depth of sensei and experienced coaches to make such a system a natural daily occurrence. And even in Toyota, sustaining this process is constant work with leaders regularly lamenting “we have to go back to the basics because we have lost our way”.


If the only effective model for changing deep-seated beliefs, assumptions and behavior patterns is deliberate practice in a coach-learner model, with peer support, we need a way to create that environment on a daily basis. Obviously no company could afford an army of “lean experts” to spend every day with work groups constantly watching them to find teaching moments. The alternative is to develop the managers already there to act as coaches of new routines for improvement that can be deliberately practiced daily. Of course a supervisor is not likely to learn these new patterns or practice them in a disciplined way unless their boss is doing the same. And this continues up the hierarchy. Thus, continuous improvement, a result of continuous learning, cannot simply be delegated. It must be learned from the top down.

Mike Rother in Toyota Kata (2009) took on the challenge of trying to develop teachable routines that can be effective in cultures that do not already have in their DNA effective improvement routines. He used the term “kata” from the martial arts where there is remarkable skill attainment by breaking complex mind-body feats into individual steps that are deliberately practiced. The novice practices them exactly until they have mastered an individual kata – perhaps one body motion at a time. They systematically progress and as they learn the body movements, they are also learning a new way of thinking about how to control their body, how to respond to what the opponent does and how to anticipate their opponent’s moves evolving to mind-body integration.

Mike dissected what the Toyota sensei do to develop people and concluded there should be at least four characteristics to the improvement kata:

  • Used system-wide in daily work
  • Suitable for any goal or problem (a content-neutral habit: the way we improve)
  • Based on a scientific model that relies on facts and data
  • Include structured practice routines for beginners (proficient users can vary the routines).

The resulting improvement kata has four steps (see Figure 5). With a clear direction, we grasp the current condition, which allows us to define our next target condition, and test ideas for improvement through small experiments following rapid PDCA cycles.

Figure 5

The concept of a “target condition” is an important part of the method. It is not a distinction that Toyota makes, but it is a distinction that naturally comes out of their coaching process. A target condition can include an outcome such as reduced inventory but it also must include a desired characteristic of the process itself. For example, build to a leveled schedule with an order delivered to production every 15 minutes. The 15 minute cadence will reveal obstacles, parts may be produced early or late, which will lead to further PDCA learning cycles until the process and people can reliably produce to the 15 minute signal. The hypothesis, which can be tested by trying things, is that building just what the customer needs, on average, to that cadence will reduce inventory. In fact, a broader systems view would predict the inventory reduction will be greatest as we move upstream to earlier processes in the supply chain. Level schedules mean predictable demand patterns and less need for large mountains of safety stock.

Like the Toyota process visual management in a team setting is part of the kata. The “learner’s story board” mimics the improvement kata (see Figure 6). This provides focus for the group. It is not a historical record, but a constantly updated document of what we tried, what we learned, and what we will try next. Small PDCA experiments are each recorded as learning cycles and we write in what we tried and when, what we expected to happen, what actually happened, and what we learned. Those teaching this process repeatedly find it is not real until it is seen on the board. The story board becomes a living representation of the improvement process.

Figure 6

Note that it would be impossible to teach someone to repeat this daily routine until it becomes an embedded way of thinking and acting by dropping in for a one week kaizen event, or even doing a four month project with only periodic coaching. The daily improvement routine and coaching routines to get the level of repetition needed require dedication, focus and discipline. But mere mortals can do it, even progressing rapidly with as little as 15 minutes dedicated to practicing each day. That is, they can do it if they have an experienced coach guiding them, who injects procedural course corrections.

The coach should be their manager, not a lean expert who drops in occasionally. Therefore, their manager-coach needs some degree of mastery of the improvement kata that they are regularly practicing. They also need to develop coaching skills by mastering each step of the coaching kata. A set of standard questions have been developed for the coach to practice as s/he walks the learner through the story board (Rother, 2009). This is reinforced by going to see what the learner has worked on. The learner reflects on what they predicted would happen, what actually happened and what they learned.

None of this would be surprising to a Toyota master sensei. They understand each step of the process and it is what they naturally teach. They might even ask: why do we need all these new words? But what we do not find at Toyota is a teachable set of routines that can be deliberately practiced with this level of granularity. On the other hand, they already have a critical mass of well-trained coaches through investments over decades.


I have argued that sustaining the gains from lean projects is a hopeless endeavor. Use of the new procedures or tools will require skill and dedication by the people who work in the process. Moreover, conditions will change so someone has to constantly update the tools and procedures, for example, rewriting the standard worksheets. When the expertise to create them has left, the lean expert documents will become stagnant, outdated, and empty artifacts.

Even if you can teach people the technical skills to rewrite standard work, it is not likely anything will improve. It is more likely that rewriting the standard work will be an attempt to document what we are doing today. Improvement requires stretching to achieve a higher goal – striving for something. When striving for something we have not yet achieved, we must overcome our biases about what we are certain will work and be willing to test what we think might work. If those tests are rapid and not costly we can rapidly learn what works and what does not.

In the end, if we want continuous improvement that is focused on meeting the true needs of the organisation we need a tool for sustainment. But what we need to sustain are not gains, but the continuous improvement process itself. That will only be sustained when people learn a set of routines that become routine. These are the routines for how to improve, how to strive, how to learn our way to any goal.