This article explains why Lean Management keeps failing outside Toyota. The explanation is a flaw that exists at Toyota itself, and in every organisation that runs Hierarchical Decision-Making (HDM) and wants to embrace Lean Management. The same flaw explains why only one of the Toyota Way’s two pillars travels. The first, Continuous Improvement, has spread through Lean to every industry on earth. The second, Respect for People, has not: its values can be embraced, but its scope is almost never reached, and fifty years of teaching, training and culture work have not changed that. The literature calls it the forgotten principle, and studies link the low success rate of Lean implementations to its absence. Lean Management, as this article uses the term, covers the whole of how an organisation is run: its work, its strategy, its staff and everyone its choices reach. That takes both pillars.

Continuous Improvement works on the floor, meaning wherever the work is done, in every department. On the production line, what one worker sees can stop the work: whoever spots a problem pulls the andon cord, and the line stops if it is not sorted out in time. Respect for People names a scope — customers, employees, business partners, whole communities — that only the choices made by senior management reach: the target, the price, the volume, which plant closes. Those choices are made above the floor. The people who must carry them out can see what those choices will do to customers, suppliers or a town, but what they see never has to reach whoever made them. That is the flaw: the Toyota Way names who the second pillar is for, and leaves it to whoever makes those choices whether they come out well or badly.

What is called a flaw here, in Toyota’s context, is the decision-making architecture almost all organisations run: decisions from hierarchical positions that others are expected to act on, with nothing requiring the reasoning behind them to be put to those people and answered before they must act. This is how Decision Free Solutions, the framework this article uses, defines HDM. And wherever the second pillar is adopted, HDM becomes the flaw. In short: Lean Management is incompatible with HDM.

At Toyota, which also runs HDM above the floor, its culture makes up for the flaw to a large degree. It draws out the collective perceptiveness of its people, and most choices above the floor have their reasoning questioned and answered before anyone has to act on them. Toyota grows its leaders over whole careers, and they can see what a choice will do to those it reaches. But a culture does not require every choice that others are expected to act on to be fully substantiated, with every question answered. Where one is not, people are expected to act on it all the same.

One such choice was a growth target set by Toyota’s senior management: about fifteen per cent of the world market in the 2010s. Between 2000 and 2010, customers in the United States were misled over a recall, as Toyota admitted in 2014; temporary workers taken on for the growth were the first to go when demand fell; and suppliers were told to cut their prices by thirty per cent. At Fremont, California, a plant where Toyota’s way of working had run for twenty-five years closed in 2010, and its workers lost their jobs. What Toyota did for them came once the closure was announced, and it was fully aligned with Respect for People. The growth target that set conditions for the closure, however, never had to take in what it would do to them.

Organisations that take up Lean Management to follow Toyota start without that culture. Their senior managers have not spent a lifetime at Toyota, so by definition they do not see what Toyota’s lifelong leaders see. Nor, under HDM, can they reliably tell who among their own people has the expertise to see what a choice will do. So they cannot draw on the collective perceptiveness of their people the way Toyota’s culture does. With HDM in place, such a culture is almost impossible to grow.

Continuous Improvement can be taken up without touching how senior management makes its choices. The second pillar rests on those choices, and needs what people across the organisation can perceive to reach them. That takes a requirement, built into how the organisation works rather than written as a formal rule: every choice that others are expected to act on is fully substantiated before they have to act on it, or named as the decision to be considered for risk management where it cannot. Without it, the second pillar will never reach those it names. With it, reaching them depends far less on a culture like Toyota’s. That requirement, the fix for the flaw, is an andon cord above the floor.

1. Introduction

Half of the Toyota Way does not travel. Toyota’s own account of what it does rests on two pillars, Continuous Improvement and Respect for People, and Lean, the West’s name for what it saw at Toyota, includes both. The first has reached every industry on earth. The second has not reached those it is for, and not for want of trying. It has been taught, trained for, written into strategies and turned into culture programmes, and it has had a name in the literature for over a decade: the forgotten principle. Nothing has moved. A diagnosis that is accurate, widely accepted and produces no change is naming a symptom rather than a cause.

Toyota wrote the two pillars down in 2001. The first pillar works on the floor, which in this article means wherever the work is done, in every department. The second pillar has two parts. It sets out values — understanding one another, taking responsibility, building trust, developing people — and it sets a scope. Toyota’s own words are that it respects all Toyota stakeholders, which its Guiding Principles spell out as customers, employees, business partners and the communities it operates in. The values alone do not reach that scope. What reaches it is settled above the floor, that is, wherever choices that set the work on the floor are made, at any level: the target, the budget, the volume, the price, which plant closes.

This article uses Decision Free Solutions (DFS) to explain why the second pillar does not reach that scope. DFS is a deductive organisational framework, built to achieve goals at minimal risk by bringing an organisation’s available expertise to bear, which it sums up as Making Expertise Matter. It starts from two principles. Reality is causal: outcomes come from the conditions present and the rules that operate on them, and from nothing else. And people perceive those conditions and rules unequally: everyone perceives something, and some perceive more (Verweij, 2026a; 2026b).

DFS defines Hierarchical Decision-Making (HDM) precisely. HDM is an architecture for how directive decisions are made: others are expected to act on them because of the hierarchical position they come from, with nothing requiring the reasoning behind them to be put to those people and answered before they must act. Almost every hierarchical organisation runs it. The word ‘decision’ is precise here: a choice that has not been fully substantiated. A choice is fully substantiated once its reasoning has been put to the people expected to act on it and every question they raise has been answered, which draws what they can see, their perceptiveness, into the choice. HDM runs within a hierarchy, but not every hierarchy runs HDM: one in which every directive choice is fully substantiated first is not running it, so nothing here argues against hierarchy.

The argument runs in eleven steps, one to each of the sections that follow.

2. The pillar that does not travel. Continuous Improvement has spread to every industry. Respect for People travels only in part: its values can be embraced, but its scope is almost never reached.

3. What the Respect for People-values cannot carry. The second pillar’s values require nothing of the directive choices that reach those its scope names.

4. What is missing. Substantiation of directive choices, with the conditions of each group the scope names inside their reasoning. Making it happen for every such choice has to become part of how an organisation works; it is not a formal rule, and this article calls it the requirement, for short.

5. How Toyota makes its choices above the floor. Hoshin kanri and nemawashi invite in the perceptiveness of people across the company, to a degree few other organisations match. They end in agreement, but agreement is not substantiation, so the second pillar’s scope may still not be reached.

6. When other organisations take up the second pillar. Without the requirement, directive choices are made under HDM, the ordinary architecture running almost everywhere. It becomes a flaw once an organisation adopts the second pillar, because that pillar cannot be achieved under it: nothing has to carry what the floor sees into a choice before anyone has to act on it, and whether an improvement comes out win-win turns on exactly that.

7. What Toyota’s culture makes up for. Above the floor, Toyota’s culture makes up for the flaw to a large degree: leaders grown over whole careers can see what a choice will do to those it reaches.

8. Trust in place of the requirement. Where the requirement belongs, Toyota’s own values ask the people expected to act on a choice for their confidence: mutual trust, as its executive training defines it.

9. Where the choices above the floor landed. From 2000 to 2010 the culture was fully in place, and no choice above the floor had to take in what it would do to the people it reached. A growth target, the prices asked of suppliers and the choices made about a recall set the conditions for every group the second pillar names, and people in each were hurt. What Toyota did for the people at Fremont came after those conditions had landed.

10. Why Toyota’s culture cannot be copied. That culture does important work, but other organisations cannot copy it. No teaching or training gives their senior managers what a lifetime at Toyota gives its leaders. Under HDM, little shows who among their people can see what a choice will do. Without that culture, whether a choice takes in what it will do to people is luck. The way out is to fix the flaw: once directive choices are substantiated, they draw on what everyone in the organisation can see.

11. So one pillar travels, the other does not. Continuous Improvement can be taken up without changing how directive choices are made; Respect for People cannot reach its scope until those choices are fully substantiated. Under HDM that does not start by itself: a question about a choice that carries no reasoning can only land on whoever made it, so perceptiveness stays on the floor, pointed at the work. A culture that draws out perceptiveness, like Toyota’s, changes how long it takes before those the scope names are hurt, not whether they are.

12. Fixing the Toyota Way’s own flaw. What would hold the second pillar up is the requirement. What would hold the requirement up is a different responsibility for the roles that carry authority. DFS proposes that HDM above the floor goes and the hierarchy stays, meaning that people are no longer expected to act on a directive choice because of position alone, and that a position sees that every directive choice gets substantiated. The requirement fixes the flaw, so the second pillar can reach its scope without anyone having to copy Toyota’s culture.

Toyota is the case with a public record, not what the argument is aimed at. Toyota’s engineering and its people are not the subject here, and nor is its performance, in itself or next to its competitors.

The explanation rests on three things. Two are on the record: the scope Toyota wrote down, which names its customers, and a recall in which, as Toyota admitted in 2014, it misled its customers in the United States. The third is whether the choices that reach that scope have to be fully substantiated before anyone has to act on them. Toyota’s 2014 admission answers that as well: staff could be told to keep a cancelled design change out of writing. This can only happen where people are expected to act on a choice without its reasoning having been put to anyone. Knowing more about how Toyota works adds detail, but it does not move the argument. Only one finding would: that Toyota does not run HDM above the floor, because some part of it makes a target, a budget or a volume wait until it has been fully substantiated.

The question was never how to implement Lean better. It is whether an organisation that makes its directive choices the ordinary way, running HDM, can hold the second pillar at all. It cannot.

2. The pillar that does not travel

Toyota’s own statement of the Toyota Way, set down in 2001, rests on two pillars: Continuous Improvement and Respect for People (Liker, 2004; Liker & Hoseus, 2008). The literature more often calls them principles.

Until 2001 Toyota’s principles were not set down for the company at large; they were absorbed through the job itself, and taught in person where the system had begun travelling. Toyota ran a seminar on its beliefs and values for its North American plant managers in November 1990 (Husar, 1991). They were written down because the company was expanding worldwide: Fujio Cho, Toyota’s president at the time, said the booklet was compiled so that Toyota’s philosophy could reach employees across different languages and cultures (Toyota Motor Corporation, 2003). Four members of Toyota’s production control department, Fujio Cho among them, had named a ‘respect-for-human’ system in print by 1977, one in which workers take an active part in running and improving their own workshops (Sugimori et al., 1977). The 2001 booklet made respect for people a pillar of the whole company’s way of working, and widened it to all Toyota stakeholders.

What happened to it in transit has been studied repeatedly. A review of implementation strategies against the Toyota Way principles found that the principles least present in those strategies are the ones linked to the second pillar (Coetzee, Van der Merwe & Van Dyk, 2016). The Lean researcher Emiliani reports that the principle has been almost entirely ignored in Western implementations (Emiliani, 2008), and it is his account that the later literature cites for the point. A study of the principle outside Toyota reports that most of the articles it reviewed attribute the low success rate of Lean implementation to the absence of the second pillar (Ljungblom & Lennerfors, 2021). The pattern has a name in the literature: the forgotten principle (Cardon & Bribiescas, 2015).

Writing in 2021, Ljungblom and Lennerfors observe that the missing component has been known for more than a decade, and that there is still more failure than success in implementing Lean outside Toyota.

3. What the Respect for People-values cannot carry

The second pillar is easy to read too narrowly. Respect for People is not politeness and it is not a wellbeing programme. Toyota’s 2001 statement names five values: challenge, kaizen and genchi genbutsu under the first pillar, and respect and teamwork under the second (Toyota Motor Corporation, 2001). Respect means understanding one another, taking responsibility and building mutual trust. Teamwork means growing people professionally and personally, sharing the chances to develop, and getting the most from individuals and teams (Liker, 2004). The English name of the pillar renders ningensei soncho, closer to respect for humanity than to respect for persons.

Alongside the values, the pillar sets a scope. Toyota’s own statement of the two pillars ends on it: the company respects all Toyota stakeholders (Toyota Motor Corporation, 2001). It puts no limit on that, and the Guiding Principles the Toyota Way exists to carry out say who is in it: customers, employees, business partners, and the communities Toyota operates in (Toyota Motor Corporation, 1997). The values and the scope are different kinds of thing. The values describe conduct between people who work alongside each other. The scope names who the pillar is for, and most of them do not work alongside anyone at Toyota.

Between the two stands senior management: the people who set strategy, agree prices, fix volumes, decide where a plant opens or closes, and conduct the company’s dealings with everyone outside it. A supplier negotiates with them. A town feels their choices through the plant it hosts. An investor hears from them. None of that is work on the floor, and nothing done on the floor settles what reaches a supplier or a town. A choice made above the floor does. Whether Toyota’s scope is met at all therefore rests on people whose work the values barely describe.

Two of the five values do speak to how a choice gets made: genchi genbutsu, under the first pillar — go to the source, find the facts, choose from those — and respect, under the second, which starts with the effort to understand one another. Both point the same way. They tell senior management how to inform itself about its own operation, which is the first pillar’s business. They say nothing about the choices only senior management makes: the price target, the plant closure, the production volume, the choices that land on the stakeholders the scope names.

And they are values. A value describes what a person does when they are doing it well. It does not require them to do it, and nothing follows when they do not.

The second pillar therefore has two jobs. The first is sized to the floor, where people work alongside each other: it is the job Toyota named in 1977, for workers running and improving their own workshops. The scope names those no work on the floor reaches, so it hands the pillar a second job, and only senior management can carry it out. That job arrived when the pillar was widened in 2001 to all Toyota stakeholders, with values to describe it and nothing to require it.

On the floor, Toyota built one thing that obliges rather than describes. A worker who sees a problem pulls the andon cord or presses a button, and the station’s number lights up on a board above the line. A team leader has to come, then and there. The line keeps moving only if the team leader can sort the problem out before the car reaches the end of the station; otherwise it stops there (Balle, 2010; Lean Enterprise Institute, n.d.). Starting with its Kyushu plant in 1992, Toyota has split lines into short segments with a small buffer of cars between them, so most stops halt only their own segment (Roser, 2014). Nobody needs persuading and nobody’s good intentions are involved: what the worker saw is what starts it. The cord does not draw out what someone sees; it makes what they see something that has to be answered, and keeps the work from moving on until it has been. Toyota applies jidoka, the principle the cord comes from, to its office work (Toyota Motor Europe, n.d.).

Toyota required a great deal above the floor as well, and required it seriously; what it required there was that people follow a process. For the choices that set the work, the ones that reach every other stakeholder Toyota names, there is nothing to pull.

Both of the pillar’s jobs matter for Lean Management, the subject of this article. Lean is the name the West gave, from 1988 onward, to what it saw of Toyota’s systems for production, product development and management. Lean Management, as this article uses the term, is the whole of it: how an organisation sets its strategy, runs its work, treats its staff, and deals with everyone else its choices land on — suppliers, customers, investors, the communities it sits in. It is a scope as much as a set of values. That is what separates it from the tools, which travel anywhere and ask nothing of the choices above the floor, and it is why Lean Management needs the second pillar, both jobs of it.

That gives a line between two outcomes. Embedded Lean is both pillars reaching the whole of the stated scope: the pillars at work on the floor, and the choices made above it substantiated before anyone has to act on them. Fenced Lean is everything short of that. Running the tools alone is fenced. So is a great deal of real second-pillar practice, if what is in place stops at the floor, because the choices outside the fence are the ones that settle whether a stakeholder comes out ahead or behind.

So Lean Management is what puts Embedded Lean in place. Where Lean is fenced, Lean Management is not in place, whatever is written down and however real the practice inside the fence. That makes it a different way of running an organisation, not a better-managed version of the usual one. Which raises the question the rest of this article answers: why almost nobody has it.

4. What is missing

Lean goes in, the tools work and the floor improves, but the second pillar does not change how senior management makes its choices. At Toyota, as almost everywhere else, HDM runs above the floor: directive choices do not have to be substantiated before others are expected to act on them. Toyota makes far better use of the perceptiveness available across the company, with nothing requiring it, than almost any other organisation. There as elsewhere, the position can still decide without drawing on it. So Lean stays fenced. On the floor, what somebody sees can hold the work until it has been answered; above the floor, nothing does.

The pattern of senior management not changing how it makes decisions when Lean goes in has held for decades. Two explanations for it are on the table. The first is Emiliani’s, and it turns on what people in senior management want: they gain from deciding without having to explain themselves, so they hold on to it, and whatever threatens that gets watered down. He calls the arrangement that carries this the Institution of Leadership (Emiliani, 2018, 2026).

The second is the one this article arrives at, and it turns on how the organisation is built: nothing in it requires a choice made above the floor to be substantiated. So what the floor can see may never bear on those choices, and nobody has to want anything for it to stay that way. On this account, naming the missing pillar more accurately cannot move it. What would have to change is the decision-making architecture, and nothing in an organisation puts a question to its own architecture.

What the architecture of HDM leaves out shows most plainly in a choice that anyone who has worked on Lean outside Toyota will recognise. Some choices only the person making them has to act on. Others are choices that other people are expected to act on, whether or not they can see why. Those carry what DFS calls directive authority (Verweij, 2026c), and senior management holds a great deal of it: the target, the budget, the headcount, the price, the volume. In most organisations an improvement target arrives as a number — twenty per cent, this year — and nothing has to arrive with it: not where the number came from, not what it was weighed against, not what happens if the work will not carry it. The team takes the figure and starts working out how to hit it. At Toyota a good deal usually does arrive with a target, through hoshin kanri, which the next section takes up. But there too, nothing has to.

What the team has been handed is a desired outcome: a result to bring about, which may or may not be achievable (Verweij, 2026a). Reaching it with as little risk as possible means seeing every condition that would have to be in place for it. The difference between what has to be perceived and what the people involved actually see is what DFS calls the perceptiveness gap (Verweij, 2026c), and risk is what that gap does to the outcome. Toyota’s problem-solving works on a gap too: the one an A3 sets out between the current condition and the target condition. The perceptiveness gap, however, sits one step earlier. It is about what the people pursuing the target can see, not about how far they still have to go.

What makes the gap visible is substantiation: setting out how a choice gets from what can be perceived to the result wanted, plainly enough that others can follow it and question it. That does not narrow the gap by itself; it shows where the gap is, which is what anyone would need before they could do anything about it. A substantiated choice can be examined. An unsubstantiated one, coming from management, can only be obeyed. The A3 comes close to substantiation. It puts the situation, the causes, the proposal and the reasoning on one sheet, where others can read and question them. The difference is that the A3 is a format people choose to use: nothing requires a target, a budget or a headcount to arrive on one before the people who have to deliver it are expected to act on it.

Narrowing the gap takes what DFS calls the full substantiation process (Verweij, 2026a). The reasoning is put to the people expected to act on the choice and to those best placed to see what it will do, whether or not they will carry it out. Their questions are drawn out and answered, and the process ends only when nobody taking part has a question left that would bring more into view. That is how what they can perceive gets into the choice. Nemawashi is the closest thing at Toyota: a proposal is taken round the people it affects and questioned before anything is settled. It differs in who is asked and in what ends it. Nemawashi goes to the people the proposal affects, and ends when they agree. Agreement can come before the questions run out. People can agree with a proposal they do not fully understand, because they are unsure, feel too junior, or are not comfortable speaking up. And perspectives that bear on what the proposal will do to those the second pillar’s scope names may never have been brought in.

Earlier, a decision was described as a choice that has not been fully substantiated. With the desired outcome in view, the definition can be given in full: a decision is a choice that is not fully substantiated to contribute to achieving a desired outcome (Verweij, 2026c). Defined precisely, decisions are to be minimised rather than optimised, and what minimises them is bringing expertise to bear. Toyota’s own principle points part of the way: decide slowly, by consensus, and implement quickly (Liker, 2004). DFS treats a decision as a source of risk. Where perceptiveness that could still be obtained has not been brought in, that risk is unnecessary, and the work is to bring it in. What cannot be brought in is named: the choice is flagged as a decision, and its risk is stated and managed like any other known risk.

Substantiation runs from none to full. A fully substantiated choice is not a decision at all. The gap has closed, and the choice becomes the logical and obvious thing to do (Verweij, 2026c). Standard work is the same: once it has been worked out, it is simply how the job is done. All that remains is monitoring, in case conditions change. Where substantiation stops short of full but all the obtainable perceptiveness has been brought to bear, DFS calls the decision complete (Verweij, 2026c). It adds no unnecessary risk, but it still carries identified risk, which keeps the gap open and keeps it a decision. That risk is best kept in view the way the floor keeps a known abnormality in view. At the far end sits the bare authoritative decision (BAD): a directive choice with no substantiation at all, and with nothing having drawn out what anyone else perceives (Verweij, 2026d). The target in the example above is one. It carries the most unnecessary risk, and the people who have to deliver it act on it without ever seeing why.

What is missing above the floor is a requirement. Under HDM, a directive choice can come down as a BAD, and nothing requires it to be fully substantiated first.

Seen this way, Lean’s toolkit splits in two. One kind of tool draws out what people perceive about the work: kaizen, value-stream mapping, the five whys, the A3. The other kind settles why a choice is one that people are expected to act on, and the andon, cord or button, is the one example. Whether a tool travels depends on which kind it is. The first kind can be picked up anywhere, because it asks nothing about who gets to settle what.

5. How Toyota makes its choices above the floor

A substantiation requirement would do for the choices that set the work what the cord does on the floor: a directive choice would wait until it had been fully substantiated. Toyota built more for those choices than almost any other company.

Two of Toyota’s practices come near the same thing above the floor without being of the same kind. Genchi genbutsu sends a manager to go and look, and leaves the choice exactly where it was. Nemawashi, set beside the full substantiation process in the previous section, differs in one more way: it is a custom, and a custom is done when people choose to do it.

The largest practice Toyota runs above the floor is hoshin kanri. Since 1963 it has come in three layers — basic policy, a long-term plan, the annual plan — with the year’s objectives sent down through every level. Each level presses on its target against what it can see, the target adjusts on the way, and progress is reviewed through the year (Smalley, 2026). The passing back and forth has a name of its own: catchball (Tennant & Roberts, 2001).

What ends it is agreement, as with nemawashi. The exchange runs until the levels agree on the target and the means, and a number passed back and forth until both sides agree has been negotiated. Whether it was ever fully substantiated is not what the process runs on.

What Toyota built here is real. Hoshin and nemawashi draw out more of what an organisation perceives, before anyone has to act on a choice, than most organisations manage anywhere at all. And where the largest choices cannot be settled in full beforehand, consensus of that quality can look like the most anyone could do, and therefore enough.

The exchange is expected to converge, and it cannot run for ever (Tennant & Roberts, 2001). Where a review during the year moves the target again, that round ends in agreement too. A level can press on its target and work out how to meet it: the plan, the steps, what it will take, what has to be in place. What it cannot do is hand the target back as the wrong target, or hold it open until it has been fully substantiated. And what one level settles as its means becomes the target for the level below it (Witcher & Butterworth, 2001), where the same thing happens again. How to meet a target is open at every level; whether it stands at all is not. Each level gets a turn. None gets a cord.

For the levels to agree, only one of the two sides has to give way. The level with the authority can hold. The level that has to deliver is the one that gives way. Where the levels disagree, what ends the exchange is therefore where the authority sits, not whether anything is left to bring out.

What the agreement settles is that the parties agree. Whether the target can be met, and whether meeting it will cost anybody in the scope the second pillar names, stays open. Tennant and Roberts document an implementation in which managers set goals that could not be met. What the method monitors afterwards, in several written accounts of it, is progress against the agreed target and the countermeasure when it slips (Akao, 1991). That tracks whether the target is being hit; what the agreement left open goes unwatched.

So what hoshin kanri produces sits between the bare authoritative decision and the complete one. It is a long way from a bare decision: a great deal has been drawn out, and the number has moved. It is not a complete decision either, because the pressing stops where the authority sits. What would still have come out was obtainable, and the decision goes ahead without it, carrying unnecessary risk that nobody has looked for. What no amount of pressing could have brought out is never named as unsettled, so neither the identified risk nor the risk nobody could see coming is written down, and nothing is set to manage the first or watch for the second.

A process runs before anyone has to act on the choice, and a process is not a requirement. A requirement would say: nobody is expected to act on this choice until it has been fully substantiated. What is there instead says: people can be expected to act on this choice once the process has been run. In the end, the hierarchical position may still decide, and running the process does not change why people are expected to act on the choice. Which is why the whole of it can run without anything in the architecture changing.

In the terms of the previous section, all of this belongs to the first kind of tool: it draws out a great deal, and none of it settles why people are expected to act on a choice. On the floor, Toyota’s answer to that question is what somebody perceived. Above the floor, the ordinary answer stands.

Two answers to one question, running in the same building, with a line between them.

Tools of the first kind sit at what DFS calls the operational layer, where drawing out what people perceive is one of the main tasks, and Lean has already built that layer well. One layer up is the architectural layer, where the andon cord works and where DFS does its own building. It specifies what has to be in place for expertise to bear on an organisation’s outcomes, and the operational layer gets its job from that.

HDM, defined earlier, has three parts, and all three are now in view. The choices are not fully substantiated: decisions. Other people are expected to act on them: directive authority. And nothing requires them to be fully substantiated before those people have to act: the requirement is missing. Almost all hierarchical organisations run this architecture (Verweij, 2026d), and BADs are its signature. How authority is arranged, by level, division, region or function, is not what the argument is about: the mechanism can be taken out and the hierarchy left standing.

Of the three, the third is the one Lean Management would have to change. As things stand, a choice lands on the people who have to act on it, and nothing gives them standing to ask what it rests on. What the Toyota Way does not have is a cord above the floor: one that could be pulled whenever a directive choice comes down before it has been fully substantiated. In most organisations such a choice is often a BAD. At Toyota, where the culture and practices like hoshin kanri see to it that a good deal usually comes with a choice, BADs are likely to be rare. The cord would still be needed there, whenever the people expected to act on a choice had no chance to question what came down, or were left with questions nobody answered. Toyota’s own record from 2000 to 2010 shows whether what it built instead is enough.

6. When other organisations take up the second pillar

The Toyota Way set down a scope in 2001, all Toyota stakeholders with no limit put on that, and built one mechanism to hold a choice to what somebody perceived: the cord, on the work itself. For the choices that set the work, the values were the whole of it. That is the Toyota Way’s own flaw: a document that names who a principle is for, and requires nothing structurally of the choices that reach them.

In organisations that take up the second pillar, the missing requirement is part of the ordinary architecture, HDM, which was there long before anyone had heard of Lean. It becomes a flaw once an organisation commits to something HDM cannot deliver, and adopting the second pillar is exactly that commitment.

Take one such organisation, implementing Lean Management with both pillars adopted, and an improvement target set by its senior management. Once it is set above the floor, everyone below is expected to act on it. Employees are stakeholders, so the second pillar applies to it: make every effort to understand the people this lands on. What is there to understand?

Not how they feel about the number. Why this number and not another, and what reaching it is in service of. Whether the work can carry it at all. Where the slack actually is, which lines are already running at their limit, what the last round of improvements took out and what it left behind, which of the changes on the list are real and which are wishful. And what would have to give if the number is to be met anyway, and who would absorb it. All of that is perceived where the work is done, by the people doing it, and it is not available anywhere else. It is not in the reporting, because the reporting shows what the plant produced rather than what producing it cost.

So the effort the pillar asks for has nowhere to happen. Not because nobody wants to make it, but because there is no point in the process at which what the floor perceives has to reach the choice before anyone has to act on it. A manager may ask. Nothing requires it, and nothing depends on the answer. In an organisation where nobody has had to say their reasoning out loud, little tells them whom to ask, because perceptiveness becomes visible only when somebody expresses it.

The cord that would supply that point is the thing HDM is least able to take, because of what it does to the authority it would hang over. Under HDM people are expected to act on a choice because of where it came from. A question put to the choice is therefore a question put to the position, and pulling a cord on a directive choice is an act against whoever made it.

The conditions the second pillar needs are therefore ones the HDM architecture cannot produce. The second pillar needs that point in the process; HDM is the architecture in which no such point is required. That is one condition described twice, and no arrangement holds both.

What the missing requirement costs shows up in the improvements themselves. An improvement comes out win-lose when the business gains at the expense of employees, suppliers or a community. Whether it comes out that way turns on one thing: whether the conditions of everyone it lands on were in the reasoning. What losing a job means to the person who holds it. What a price cut does to a supplier’s books. What closing a line does to the place it sits in.

The interests were never separate. A supplier that goes under is a supply problem. A town that hosted a plant supplied the workforce that ran it and the small suppliers that fed it, and closing the plant takes both apart. Employees who stop bringing things forward are the first pillar running dry. An organisation that cannot draw out what its own people perceive cannot perceive that either, and pays for it twice.

Win-win is not a preference held by decent people. It is how an improvement comes out once the conditions of everyone it lands on have been perceived, and an organisation with the requirement in place gets it wherever it can be had.

7. What Toyota’s culture makes up for

Organisations that take up the second pillar under HDM meet the problem the previous section set out, and most are left with whatever senior management happens to perceive. Toyota meets it with a culture. That is also the answer waiting for any failed Lean implementation: they did not have Toyota’s culture. The answer fits every failure, whatever produced it, so no failure tests it.

Toyota’s record can test it. Toyota has that culture, has run the Toyota Way longer than anyone and with more riding on it, and has nearly everything else the second pillar needs. So whatever is still missing at Toyota cannot be put down to a missing culture. The record also holds a plant in California, where Toyota’s way of working went into a workforce that had none of Toyota’s culture.

On the floor, Toyota built more of what the second pillar needs than any company since. On choices about the work, frontline expertise held real authority, and the results set the global benchmark. Raising a problem cost nobody anything: pull the cord and colleagues come, people were applauded for admitting a mistake, and the standing assumption was that a problem lay with the system rather than with the person who hit it (Liker & Hoseus, 2008). Perceiving was rewarded. The Creative Idea Suggestion System has run since 1951, took 810,000 suggestions in 2023, and since 2019 has given separate credit for identifying a problem before anything is done about it (Toyota Motor Corporation, 1973; Toyota Times, 2024). What one plant learned was carried to the others.

Improvement work on the floor runs into one question before any other: whether an improvement will cost somebody their job. Toyota took that question off the table. Layoffs were a last resort as a matter of standing practice, so nobody stood to lose their job over an improvement. That made raising a problem safe as far as anyone’s job went: the part improvement work leans on hardest, because improvement is what frees the capacity in the first place.

The plant in California shows what that floor is made of. GM had shut its plant in Fremont; in 1984 Toyota and GM reopened it as a joint venture, New United Motor Manufacturing, Inc. (NUMMI). The workforce was American, largely rehired from the people GM had let go, and it had been thought the worst in GM: strikes, grievances, sabotaged quality, absenteeism over twenty per cent. Within a year absenteeism was around two per cent and quality was the best in GM, with the same people in it, the old troublemakers included (Shook, 2010).

Shook, who helped build the training that carried the Toyota system into Fremont, sets out what produced that. The order runs the opposite way to the usual one: people were given the means to do their jobs successfully, and the behaviour the first pillar needs followed. His example is the andon. When GM colleagues asked whether workers were really to be given the right to stop the line, the answer was that they were to be given the obligation to stop it, whenever they found a problem (Shook, 2010). A cord is not a permission. It is a condition on the work.

The no-layoff commitment went in the same way, and NUMMI had no tradition to lean on. Toyota’s position is that laying people off comes last. That was tested in the late 1980s, when the product was not selling and several hundred workers were not needed to run the plant. NUMMI’s answer was to write into the contract the steps that would come first, among them cutting plant operating hours and cutting management bonuses (Shook, 2010).

What NUMMI shows is that the floor runs on conditions, and conditions can be installed: an obligation to stop the line, and steps written into a contract before anyone can lose a job. The way of working the first pillar needs followed them, inside a year, in a workforce with no tradition of it. None of it waited for the culture Toyota had grown over decades: NUMMI delivered the first pillar to a standard GM had not reached anywhere else.

Above the floor it is different. At Toyota a problem travels up until it reaches someone who can deal with it, and that person has to answer (Balle, 2010). A decision travelling down meets nothing comparable, because Toyota’s conditions covered the work and not the choices that set it. Masao Nemoto, a senior managing director who led much of the company’s quality work, set down that a command without a deadline is not a command, and had employees ignore any request that arrived without one (Nemoto, 1987). That is a requirement placed on a directive choice, with standing to refuse attached. It asks for a schedule and leaves the reasoning out. At Toyota, hoshin kanri runs on catchball and the A3, not on the apparatus the Western literature later built around it, and it ends in agreement (Reich, 2025; Smalley, 2026).

Toyota drew out what its people perceived further than almost anyone, and it could afford to, because none of it put the authority in question. Management decided whom to ask and what to ask about: how to meet a target, not whether the target should stand. Whatever came back was advice. It could change a decision; it could not stop one. So all three parts of HDM stayed in place: the choices above the floor stopped short of full substantiation, people were expected to act on them because of the position they came from, and a process was all that had to come first. A cord works the other way: whoever pulls it does not need to be invited.

Toyota’s record above the floor also holds outcomes that reached the scope. Emiliani (2008) records a customer-first rule, and supplier relations built since 1939 on joint problem-solving rather than price bargaining. They were built by people who could see what such relationships had to be for both sides to come out ahead, which is the perceptiveness the second pillar’s scope asks for. Outcomes like these show what Toyota’s senior management perceived. They cannot show what happens when a next generation perceives less, or when perceiving starts to cost more: a rule can be set aside, and a way of working holds only while the people involved keep working that way.

What makes up for the missing requirement is the culture at the top. Toyota grows its own leaders over long careers and carries the way of working from one generation to the next (Liker & Hoseus, 2008). So people who can see what a choice will do keep arriving at the top, and nothing has to ask for them.

Two things a culture cannot do. It cannot reach past the authority of the people who hold it: the best-run plant has no hold on a target set above it. And it cannot show when it has stopped working. A condition makes its absence visible: until it is met, the choice cannot go ahead. A culture only makes the right choice likely, and when it stops doing that, people are expected to act on the choice as they always were, and nothing marks the difference.

What Toyota reached is the ceiling: nearly every condition the second pillar needs, and no requirement on the choices that set targets, budgets and headcount. That is Fenced Lean built higher than anyone has built it: the conditions and the no-layoff protection inside the fence, and the choices that settle win-win or win-lose outside.

Toyota’s culture is not an argument against the flaw. It is why the flaw could stay out of sight for so long. The better the substitute for the requirement, the longer its absence goes unnoticed, and nobody has built a better substitute than Toyota. The culture buys time, and it cannot say when the time is up.

8. Trust in place of the requirement

The piece missing from the Toyota Way is the requirement. Something sits where that piece would go, and Toyota’s own words say what it is.

Mutual trust is one of the Toyota Way’s values (Toyota Motor Corporation, 2001). Toyota’s executive training for its North American plant managers set out what it asks of each side. Employees are to be confident that management will make the right decisions for the prosperity of the company, while management listens to them and rewards them fairly. Management is to be confident that employees will respect company policy and work hard (Husar, 1991, pp. 10-11).

For the choices above the floor, that is the whole arrangement. The people expected to act on a choice supply their confidence. In return, management is to listen and to reward fairly, and nothing requires either. The reasoning does not have to be put in front of them, and nothing lets their question hold the choice until it is answered. Where the requirement belongs, the training says: trust management.

In DFS, trust is not a way of running choices. Trust in a person rests on a record: past performance, carried into a new setting of the same kind and complexity. This record can justify acting where the reasoning cannot be put out and answered. Research on expert judgement draws a similar line: judgement earns trust only where the setting is regular enough to be learned and the person has had the chance to learn it (Kahneman & Klein, 2009). DFS takes the question a level deeper, connecting Klein’s and Kahneman’s work in one account of what makes someone expert (Verweij, 2026e). Where the reasoning can be put out and answered, confidence follows from that, and it rests on something the people expected to act on the choice have seen. The training asks for confidence in advance, from everyone, in every decision, because of who is making it.

The requirement would put the reasoning in front of the people expected to act on a choice, and see their questions answered. Toyota asks them for their trust instead.

9. Where the choices above the floor landed

At Toyota the culture was fully in place for everything that follows, and trust still stood where the requirement would go. Senior management set the targets and decided how fast to expand, what to demand of suppliers and where to cut cost. Between 2000 and 2010 both culture and trust came under load. The record of what happened is unusually complete, and the account here keeps to it: Toyota’s statements, official findings, published research and reporting at the time.

The largest of those choices was the growth target. Toyota’s push for faster growth began when Hiroshi Okuda became president in 1995 (Camuffo & Wilhelm, 2016). Its 2010 Global Vision, adopted in 2002, aimed at about fifteen per cent of the world market in the 2010s (Toyota Motor Corporation, 2012). The target was set at the top, and everything underneath it was expected to act on it. Toyota’s own account of what went wrong begins with the pace of that growth: its president, Akio Toyoda, told Congress in February 2010 that the company had grown faster than it could develop its people, and that its stated order — safety, then quality, then volume — had become confused (Toyoda, 2010).

The growth target needed capacity, and capacity meant people. Between 2000 and 2007 Toyota opened eighteen plants, and within a decade its headcount went from 183,000 to 321,000. The shortfall was filled with temporary workers, outside the training and career system and not equipped to spot irregularities or pull the cord. The team leader position had been eliminated in the late 1990s and was restored in 2008, and, by one Toyota executive’s account, some of the company’s plants around the world did not let workers stop the line when they spotted defects. At Toyota’s UK plant, team members hesitated more and more to pull the cord, to avoid the overtime that came with falling behind the daily production target. Defects per vehicle doubled on one model and quadrupled on another (Camuffo & Wilhelm, 2016).

No policy cut the cord. It kept the authority it had always had, and it stopped being pulled. What changed was the conditions it runs on, and those were set by a choice made where the cord does not reach. This is the first pillar giving way: Continuous Improvement runs on people trained to see and equipped to act, and the expansion put a growing share of the workforce outside both. Below, the problems accumulated. Above, there was nothing to pull.

Car markets shrink as well as grow, and by mid-2008 Toyota had significant overcapacity in North America (Toyota Motor North America, 2009). When demand fell, Toyota let its temporary workers go first. By August 2008 all its temporary assembly-line workers in the United States were gone (WardsAuto, 2008). In Japan, 3,000 temporary workers were to go at the end of March 2009, while its regular employees there still had their jobs (WardsAuto, 2009). In the United States it laid off none of its regular manufacturing team members, who trained and worked on improvements while the lines slowed (Cable, 2011). It offset part of the cost through what it called shared sacrifice, cutting bonuses and executive pay (Toyota Motor North America, 2009). The no-layoff practice held for everyone it covered, and Toyota had staffed much of its growth with people it did not cover.

Fremont came next. In 2009 General Motors, working through bankruptcy, withdrew from the plant. Toyota then reviewed its contract with the plant, and its own capacity and production needs, in the light of the market. In August 2009 it announced that production at Fremont would stop at the end of March 2010 (Toyota Motor Corporation, 2009; Toyota Motor North America, 2010a). The Tacoma pickup went to Toyota’s Texas plant, which had opened in 2006 and been cut back to one shift as truck sales fell (Roberts, 2009). More than four thousand people at Fremont lost their jobs (NBC Bay Area, 2010). The plant closed where three things met: capacity built at the pace of the target, a market that had fallen, and a partner that had gone.

Respect for People did what it was supposed to do. Rather than stop its orders at once, Toyota took seven more months of production from the plant, so that NUMMI could keep its people employed. Although NUMMI was an independent company, Toyota committed $250 million to support its workers through the transition. It also expected to keep using some of NUMMI’s suppliers for its own American plants (Toyota Motor North America, 2010a). When Tesla agreed to buy the plant in 2010, Toyota put $50 million into Tesla (Tesla Motors, 2010). Each step Toyota took reached people the second pillar names, and nothing required any of them.

Everything Toyota did at Fremont came after the conditions had landed. The conditions the growth had set traced back to a choice made years earlier, the growth target, and that choice was never held to the requirement. Everyone it reached included the workers in plants that would be surplus if demand fell, and the people hired to staff the growth outside the no-layoff practice. Holding the target to the requirement need not have stopped the growth or kept Fremont open. It would have weighed those people’s conditions before anyone had to act on the target, and named in advance what could not be avoided, to be managed as a known risk. Both the growth target and the choices Toyota made at Fremont were made without being held to the requirement. The Fremont choices took in what they would do to the people they reached; the growth target did not. The difference was what the person choosing happened to see, and nobody on the receiving end could tell in advance which kind was coming.

Toyota’s own principles name business partners among those the pillar is for, and its executive training states why: if suppliers and dealers cannot make a profit or provide job security, Toyota eventually goes out of business (Husar, 1991). A price target set inside Toyota reaches a supplier as a condition it has to live with, and whether the target’s reasoning took in what it would do there was left to whoever set it. Toyota’s cost programme CCC21, announced in 2000, required a thirty per cent cut in parts prices over five years, and a further round followed from 2005 (Camuffo & Wilhelm, 2016). One account from the bottom of that chain, reported in 2008, is of a supplier employing fewer than a hundred people who met his target by releasing a fifth of his workforce (Miller, 2008).

Customers are in the scope too. Around nine million vehicles were recalled worldwide for unintended acceleration, from two causes: floor mats that could trap a depressed accelerator, and pedals that could stick part-way. The US regulator, working with NASA engineers, found no electronic cause (NHTSA, 2011). The defects themselves were engineering problems, and the requirement this article argues for would not have prevented them. What was done once they were found is a different matter: those choices were made above the floor.

At Toyota, safety and recall matters were treated as engineering questions and settled by a group centred in Japan, with the American arm reporting to it (Liker & Ogden, 2011; United States Department of Justice, 2014). That was a disconnect between where a problem sat and where it was being solved, and it ran against Toyota’s own tenet that decisions belong close to the work (Liker & Ogden, 2011). What had to be perceived to answer well was not engineering. It was the state of the American regulator, the courts, the press and the public.

On the floor, what has to be perceived and the choice it bears on sit together: the problem stands in front of the person doing the work, and the cord puts it into the choice before the line moves. Above the floor the two sat apart: what had to be perceived was in America, and what settled it was in Japan. Perceptiveness of an environment is built by standing in it, and Toyota had thousands of people standing in this one. A requirement would have brought them in by itself, because a choice that cannot be substantiated has to find the expertise that can. There was no requirement, so nothing sent the choice looking for them.

What was settled is on the record. On the pedal, what the people nearest it could see was drawn out and went up: the American side flagged it to Japan as critical. That part is the culture doing what it is for. In October 2009 the group in Japan cancelled a design change, and staff were told to put nothing about the cancellation in writing, contrary to Toyota’s own standard procedures, so that the regulator would not learn of it. That autumn Toyota also told the public and the regulator that recalling models for floor-mat entrapment had addressed the “root cause” of unintended acceleration, while it knew other models were just as susceptible by design. It later gave the regulator and Congress a timeline that put its knowledge of the pedal fault later than it was. Toyota admitted these facts in a deferred prosecution agreement in 2014 and paid a criminal penalty of $1.2 billion (United States Department of Justice, 2014).

Telling staff to keep a choice off paper is possible only where its reasoning never had to be anywhere. Under the requirement, withholding the reasoning would mean nobody could be expected to act on the choice. Concealment is one of the things the missing requirement permits.

Toyota’s response changed how recall choices were made. In March 2010 it set up a Special Committee for Global Quality, chaired by Akio Toyoda. Each region gained a chief quality officer to represent its customers’ concerns, regional safety executives would take part in deciding whether and how to recall, and Toyota commissioned outside experts to review the measures (Toyota Motor Corporation, 2010). The group in Japan had been settling recalls without the people who could see the regional conditions, and this brought them in. What it gave them was a part in the choice, not a hold on it. The choices that set the other conditions, the growth target and the prices asked of suppliers, were outside the response altogether.

Those outside experts were the North American Quality Advisory Panel, led by Rodney Slater, a former US Secretary of Transportation (Toyota Motor North America, 2010b). It reported in 2011, and it had looked at how Toyota ran quality and safety across North America rather than at the recalls alone. It found that Toyota had leaned too far towards central control and should give its regions more authority. That went further than the 2010 change, which had given the regions a voice in recall choices and left the authority where it was. It also found that Toyota applied its problem-solving less rigorously to management decisions than to manufacturing: when complaints came from outside, it had often been slow to pull “a metaphorical andon cord” (Toyota North American Quality Advisory Panel, 2011). That is the line between the floor and the choices above it, described from outside.

Its main remedies were to find and correct decision errors as rigorously as faults on the line, and to move authority closer to the market. Neither is the requirement. A decision corrected faster has still been acted on first. A choice made nearer the market is still one people are expected to act on when the person making it decides, and so is a choice a regional officer took part in. What none of the three does is hold a choice until its reasoning has been put to the people expected to act on it and answered.

Customers, business partners, employees and a community: the choices in this record reached all four, and all four are the scope Toyota’s Guiding Principles name. The conditions each group met were set by the growth target, the price targets given to suppliers, and the recall choices made in Japan. None of those choices was held to the requirement, so none had to set out what it would do to the people it reached. The flaw sat there, in each of them, before any outcome showed it.

10. Why Toyota’s culture cannot be copied

The absence of Toyota’s culture is the explanation waiting for any failed implementation. It stops the question being asked: whether anything in the organisation requires a directive choice to be substantiated before anyone has to act on it. A culture of that kind takes decades and cannot be ordered in, so naming it as the missing part leaves nothing to look into. Teaching and training are what is left to do.

At Toyota, perceptiveness keeps arriving at the top, and growing that takes people already there who can tell who else has it. Under the HDM architecture almost every organisation runs, few occasions ask for perceptiveness to be shown, and in any case selection works against the people who perceive most. Senior roles go to people who are comfortable choosing without explaining and equally comfortable on the receiving end, taking a directive without asking what is behind it (Verweij, 2026d). The habits that go with this: sounding certain whether or not they can see what is going on, not naming the limits of what they can see, not going looking for people who see more, treating expertise as advice to weigh rather than as the thing that settles it, hearing a question about a choice as a challenge to the authority behind it.

Those people choose the next ones the same way, and an organisation that has run HDM for a while has staffed its senior positions with the people least likely to hang a cord. Toyota’s culture makes up for the flaw to a large degree, and the flaw is why that culture is almost impossible to grow anywhere else. Where the requirement is in place, substantiation makes perceptiveness visible. The place to start is the flaw, not the culture.

Without that culture and without the requirement, good outcomes will still happen, because good managers are common. Most people have worked for one: a manager who draws on what the team can see, absorbs pressure from above, makes sure what is asked for can be done, and produces win-win with no requirement anywhere in sight. Such a manager has perceptiveness nobody asked for and nobody checked. The outcomes happen by the luck of who is holding the choice. What cannot happen is the pillar. A pillar is what a building stands on, and it holds when the person in the chair moves on.

Luck of that kind can run for years, which is what makes it hard to see. It runs out when something arrives that a disposition will not carry: a bad quarter, a target from above, a choice where seeing what it does to people costs money. When it comes, it reaches the whole team at once. What such a manager builds is Fenced Lean, said plainly: improvement work standing on a pillar with nothing underneath it, inside a boundary drawn by the choices it cannot reach.

In HDM organisations, the people holding the choices above the floor are not refusing the second pillar. To refuse something you have to see it and turn it down. The second pillar may not read to them as a matter of performance at all, but as the soft half, the bit you get to once the numbers are safe.

Seen from outside, an organisation where the second pillar never takes hold looks like one with a people problem, and that is where the field’s usual explanations come from: resistance to change, lack of leadership commitment, wrong culture, poor communication, lack of training. With no requirement in place, the perceptiveness gap seldom becomes visible, so most of the risk is never found, and explanations like these are what is left over.

All five offered explanations say something about what the people are or want, and nothing about the conditions they work under. That makes them unusable by the person they describe, since a manager told he lacks commitment has nothing to go and change. It also makes them uncheckable. A manager may not see what a choice will do to people, or may see it and go ahead anyway. From outside the two look the same, and an explanation built on what people are or want can only guess which it is.

The usual fixes run into the same conditions. Teaching adds knowledge to people the structure picked for other reasons, and changes nothing about what the structure asks of them the next morning. Its long record is consistent with that. Plenty of managers do have something to learn, and understanding the processes better helps any of them. What knowledge cannot do is place a condition on a choice, and that is what is missing. When nothing changes, the result looks like resistance. Replacing the individual brings the next one through the same selection, into the same position, under the same architecture. And the culture fixes reach for what the flaw keeps from growing.

Nothing in the Toyota Way asks for the requirement, of itself or of anyone copying it, and an organisation copies what its model asks for. Across the whole of an organisation, the requirement would have to be brought in by the people it would apply to, because nobody else’s authority reaches that far. That is why fifty years of teaching, training and culture work have not brought it in, and why fifty more would not. Every one of those efforts is added to the structure, and a structure that asks nothing of its own choices takes additions without changing.

11. So one pillar travels, the other does not

The Toyota Way names a scope that only the choices above the floor reach, and requires nothing of those choices before people are expected to act on them. At Toyota a culture makes up for much of that flaw. Between 2000 and 2010 the choices that set the conditions did not take in what they would do to the people they reached, and so carried unnecessary risk for them. The choices Toyota made at Fremont afterwards are aligned with Respect for People. Neither kind was held to the requirement, so the people on the receiving end had no way to tell in advance which kind was coming. Everywhere else, the same flaw makes that culture almost impossible to grow. So the first pillar travels and the second does not.

The tools come through all of this untouched. They work below the level where directive choices are made, and they ask nothing of that level. Continuous Improvement can therefore be taken up at no structural cost. Of the second pillar, only the part on the floor travels: NUMMI wrote its no-layoff steps into a contract. The pillar’s scope cannot be reached until the choices above the floor are held to the requirement, because those are the choices that reach it.

What differs between organisations running HDM is how much perceptiveness turns up at the choices above the floor anyway, with nothing asking it to. Toyota sat at the top of that range, and its record from 2000 to 2010 shows that even the top falls short. Most organisations sit a long way below it, and not because they are more timid. Drawing perceptiveness out produces questions, and a question about a choice that carries no reasoning can only land on the person who made it. So the drawing out gets rationed. What is left of it stays on the floor, pointed at the work, where improving it puts no one’s authority in question.

That is why, inside an HDM organisation, Fenced Lean does not drift towards Embedded Lean. Of the two outcomes, Fenced Lean is the stable one: it never puts a question to HDM itself. Embedded Lean has to put questions to the choices above the floor, so under HDM every attempt at it falls back to Fenced Lean. Nobody has to want that for it to hold.

Toyota, at the top of that range, has stayed Fenced for decades. Its culture draws perceptiveness out further than anyone’s, and none of it becomes a question the choices above the floor have to answer. Improvement runs continuously, and it runs where it always has: on the floor, out of reach of the choices that set it. Meanwhile the people making those choices change, as they do, and so does what they treat as important. What finally put questions to them came from outside, in a congressional hearing and a criminal case. Where nothing inside holds a choice open to questions, outside is the only place they can come from, and that is the size of event it takes.

A culture that draws out perceptiveness, like Toyota’s, buys time before the flaw shows, but does not remove it. Where there is less of such a culture, HDM runs more forcefully, and the second pillar’s scope is rarely met. What the literature has named, for more than a decade, is the second pillar going missing from implementations. What DFS adds is that it never reached that scope as a pillar, in the implementations or at Toyota.

The account offered here is derived, and Lean’s history illustrates it. The pattern the derivation predicts was written down in detail, over decades and across an industry, by people who had never heard of the argument. They recorded which half of a management system survives being moved, that teaching does not fix it, and that the field reaches for culture, commitment and communication to explain it. At the company that built it best, they recorded what a crisis exposed: the consequences of a growth target set years earlier and pursued faster than the company could develop its people.

12. Fixing the Toyota Way’s own flaw

What would hold the second pillar up, so that it reaches its scope and not only its values? The requirement. Lean’s tools already bring the perceptiveness spread across an organisation to bear on the work, and at Toyota a good deal of it reaches the choices above the floor as well. The requirement brings it to bear on every directive choice, and goes on doing so when the people in the chairs change. That is what DFS means by Making Expertise Matter, and what it was built to specify.

That leaves the harder question. The requirement applies to the choices made from hierarchical positions, and only the people in those positions can bring it in. Where HDM has run for years, those are the positions it has filled with the people least likely to hang a cord over their own choices. So what would hold the requirement up?

DFS proposes that HDM above the floor goes and the hierarchy stays. For that to happen, the responsibility of the roles that carry authority has to move from decision-making to creating, sustaining and communicating the conditions for the available expertise to be brought to bear. That is how DFS defines the leadership role (Verweij, 2026f). Anyone can take that role on wherever perceptiveness has to be drawn out, from a project leader to whoever chairs a meeting. A hierarchical position is still a leadership role, but people are no longer expected to act on its directive choices because of position alone, and it is responsible for seeing that every directive choice gets substantiated.

The perceptiveness a directive choice needs usually sits with the people doing the work, and for a target or a price it sits in several disciplines at once. So the full substantiation process runs among them. DFS leaves open how an organisation arranges that. A hierarchical organisation may, for instance, shift from management by decision-making to management by approval, with the position approving what comes out of the process, including where the seeing ran out. Where the position holds the perceptiveness itself, it sets out the reasoning and is questioned like anyone else.

The requirement is the cord above the floor, and it meets two kinds of directive choice. Where the perceptiveness a choice needs can be got, as for a production problem or a sourcing question, drawing it out closes the gap, and the choice becomes the obvious thing to do, the way a problem raised on the line gets settled. Where not all of it can be got, as for a strategic direction or a plant closure, or where there is no time to draw it out, a decision cannot be avoided. The cord then asks that what can be drawn out is drawn out first, and that the choice is named as the decision it is, with its assumptions said out loud (Verweij, 2026c).

Naming the decision is not the end of it. DFS asks that its risk be watched once people are acting on it, and the position that named it sees to that as well. Whoever is expected to act on the decision has the standing to flag that it is not working, as a worker flags a problem on the line, and to ask for the decision to travel back up the hierarchy, with what has been learned attached, so that it can be changed. It no longer has to cascade down.

At first, while an organisation changes over, the cord may be pulled often. Over time it may just hang there, because directive choices arrive already substantiated. It stays all the same: a cord nobody pulls is still what makes them arrive that way. Drawing perceptiveness out takes time at the point of choosing, and the time comes back when the work goes ahead, which is Toyota’s own principle of deciding slowly and implementing quickly (Liker, 2004).

The requirement needs practices to run on, and at Toyota most of them are already there: nemawashi, catchball, the A3, and Nemoto’s rule that a request without a deadline could be ignored. Toyota has also found the place. The trust its executive training asks of employees sits where the requirement would sit. For Toyota, the repair is to put substantiation where mutual trust now stands. An organisation with none of Toyota’s practices has far more to build and no such marker to work from, but what has to be built is the same.

The requirement takes back the one job a culture should never have been holding: whether a directive choice took in what it would do to people before they were expected to act on it. And once the requirement draws out the collective perceptiveness, what Toyota’s culture does at the top becomes something other organisations can grow.

It can start on Monday.

References

Akao, Y. (Ed.) (1991). Hoshin Kanri: Policy Deployment for Successful TQM. Productivity Press. (Originally published by the Japanese Standards Association; translated by G. H. Mazur.)

Balle, M. (2010). Pulling the andon cord: Toyota responds to challenge and change. The Systems Thinker, 21(1), 9-10 (February 2010). https://thesystemsthinker.com/pulling-the-andon-cord-toyota-responds-to-challenge-and-change/

Cable, J. (2011, April 6). Staying true to the Toyota Way during the recession. IndustryWeek. https://www.industryweek.com/the-economy/article/21960972/staying-true-to-the-toyota-way-during-the-recession

Camuffo, A., & Wilhelm, M. (2016). Complementarities and organizational (mis)fit: A retrospective analysis of the Toyota recall crisis. Journal of Organization Design, 5, 4.

Cardon, N., & Bribiescas, F. (2015). Respect for people: The forgotten principle in lean manufacturing implementation. European Scientific Journal, 11(13), 45-61.

Coetzee, R., Van der Merwe, K. R., & Van Dyk, L. (2016). Lean implementation strategies: How are the Toyota Way principles addressed? South African Journal of Industrial Engineering, 27(3), 79-91.

Emiliani, M. L. (2008). The equally important ‘respect for people’ principle. In Real Lean: The Keys to Sustaining Lean Management (Vol. 3). The CLBM, LLC, Wethersfield, CT.

Emiliani, M. L. (2018). The Triumph of Classical Management Over Lean Management: How Tradition Prevails and What to Do About It. Cubic, LLC.

Emiliani, M. L. (2026, August 13). Deriving the IoL and SoPP from first principles. The Trouble With Business Leadership [Substack]. https://bobemiliani.substack.com/p/deriving-the-iol-and-sopp-from-first

Husar, M. A. (1991). Corporate Culture: Toyota’s Secret, Competitive Advantage. General Motors internal white paper, 15 May 1991. Content drawn from a Toyota executive training seminar held in November 1990 for Toyota’s key North American personnel, plant manager level and above. Cited by Emiliani in Lean Is Not Mean (2015, Lesson 1, note 20); a retyped copy was posted at https://bobemiliani.com/wp-content/uploads/2016/08/husar_nummi.pdf, offline when checked in September 2026.

Kahneman, D., & Klein, G. (2009). Conditions for intuitive expertise: A failure to disagree. American Psychologist, 64(6), 515-526. https://doi.org/10.1037/a0016755

Lean Enterprise Institute (n.d.). Andon. Lean Lexicon. https://www.lean.org/lexicon-terms/andon/

Liker, J. K. (2004). The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill.

Liker, J. K., & Hoseus, M. (2008). Toyota Culture: The Heart and Soul of the Toyota Way. McGraw-Hill.

Liker, J. K., & Ogden, T. N. (2011). Toyota Under Fire: Lessons for Turning Crisis into Opportunity. McGraw-Hill.

Ljungblom, M., & Lennerfors, T. T. (2021). The Lean principle respect for people as respect for craftsmanship. International Journal of Lean Six Sigma, 12(6), 1209-1230.

Miller, J. (2008, August 15). Toyota’s mom & pop suppliers feel CCC21 squeeze. The Manufacturer. Reporting the Chunichi Shimbun series ‘At the Feet of Toyota’, May-June 2008. https://www.themanufacturer.com/articles/toyotas-mom-pop-suppliers-feel-ccc21-squeeze/

NBC Bay Area (2010, March 18). NUMMI workers accept shutdown terms. https://www.nbcbayarea.com/news/local/nummi-workers-accept-shutdown-terms/1864905/

Nemoto, M. (1987). Total Quality Control for Management: Strategies and Techniques from Toyota and Toyoda Gosei (D. Lu, Trans. & Ed.). Prentice Hall. Original Japanese title: TQC to toppu bu-kacho no yakuwari. The leadership principles cited here are as reproduced from Lu’s translation in Hiam, A. (1990), The Vest-Pocket CEO, Prentice Hall, pp. 55-58.

NHTSA (2011). Technical Assessment of Toyota Electronic Throttle Control (ETC) Systems. National Highway Traffic Safety Administration, February 2011. Prepared with the NASA Engineering and Safety Center. https://www.nhtsa.gov/sites/nhtsa.gov/files/nhtsa-ua_report.pdf

Reich, M. (2025). Managing on Purpose: Using Hoshin Kanri to Develop Strategy, Align Teams, Grow Leaders, and Innovate Your Enterprise. Lean Enterprise Institute. Reich spent twenty-three years at Toyota, six of them in Japan, and managed the hoshin kanri process for its North American operations. His statement that he never encountered the X-matrix at Toyota is from the book, as quoted in Smalley (2026). See also Lean Blog Interviews, Episode 528, 11 June 2025, https://www.leanblog.org/2025/06/hoshin-kanri-managing-on-purpose-mark-reich/

Roberts, M. (2009, August 28). Toyota moving Tacoma production to Texas plant. Manufacturing.net. Associated Press report. https://www.manufacturing.net/operations/news/13073191/toyota-moving-tacoma-production-to-texas-plant

Roser, C. (2014, January 26). Evolution of Toyota assembly line layout: A visit to the Motomachi plant. AllAboutLean.com. https://www.allaboutlean.com/toyota-line-layout/

Shook, J. (2010). How to change a culture: Lessons from NUMMI. MIT Sloan Management Review, 51(2), 63-68.

Smalley, A. (2026, July 4). Does Toyota use hoshin kanri? Gemba Coach, Art of Lean. Draws on the first-hand account of Mikio Sugiura, responsible for hoshin kanri in Toyota’s Corporate Planning Department in the late 1970s, published in Japanese in 2017 as Toyota Global 10. https://artoflean.com/gemba-coach/does-toyota-use-hoshin-kanri/

Sugimori, Y., Kusunoki, K., Cho, F., & Uchikawa, S. (1977). Toyota production system and Kanban system: Materialization of just-in-time and respect-for-human system. International Journal of Production Research, 15(6), 553-564.

Tennant, C., & Roberts, P. (2001). Hoshin Kanri: Implementing the catchball process. Long Range Planning, 34(3), 287-308.

Tesla Motors (2010). Form 10-Q for the quarterly period ended June 30, 2010. U.S. Securities and Exchange Commission. https://www.sec.gov/Archives/edgar/data/0001318605/000119312510188792/d10q.htm

Toyoda, A. (2010). Prepared testimony before the US House Committee on Oversight and Government Reform, 24 February 2010.

Toyota Motor Corporation (1973). Toyota company suggestion program vigorously supported by employees. https://global.toyota/en/detail/7756353

Toyota Motor Corporation (1997). Guiding Principles at Toyota. Established January 1992, revised April 1997. Corporate philosophy pages, 75 Years of Toyota. https://www.toyota-global.com/company/history_of_toyota/75years/data/conditions/philosophy/guiding_principles.html

Toyota Motor Corporation (2001). Toyota Way 2001. Corporate philosophy pages, 75 Years of Toyota. https://www.toyota-global.com/company/history_of_toyota/75years/data/conditions/philosophy/toyotaway2001.html

Toyota Motor Corporation (2003). Annual Report 2003, p. 19. Statement by Fujio Cho on why the Toyota Way was compiled. Extracts assembled in Rother, M. (n.d.), The Toyota Way 2001. http://www-personal.umich.edu/~mrother/KATA_Files/The_Toyota_Way_2001.pdf

Toyota Motor Corporation (2009). TMC to end production contract with NUMMI in March 2010. https://newsroom.toyota.eu/2009-tmc-to-end-production-contract-with-nummi-in-march-2010/

Toyota Motor Corporation (2010, March 30). Toyota begins radically reshaping operations to meet customer expectations: Company holds inaugural meeting of Special Committee for Global Quality. https://global.toyota/en/detail/337250

Toyota Motor Corporation (2012). Rapid growth. History through text, Part 3, Chapter 5, 75 Years of Toyota. https://www.toyota-global.com/company/history_of_toyota/75years/text/leaping_forward_as_a_global_corporation/chapter5/section1/item1.html

Toyota Motor Europe (n.d.). Toyota Production System. https://www.toyota-europe.com/about-us/toyota-vision-and-philosophy/toyota-production-system

Toyota Motor North America (2009). Toyota’s Indiana plant turns the corner, preps for new Sienna. Toyota USA Newsroom. https://pressroom.toyota.com/toyotas-indiana-plant-turns-the-corner-preps-for-new-sienna/

Toyota Motor North America (2010a). Toyota commits $250 million to support NUMMI transition. Toyota USA Newsroom. https://pressroom.toyota.com/toyota-commits-250-million-to-support-nummi-transition/

Toyota Motor North America (2010b). Leading safety and quality experts named to Toyota North American Quality Advisory Panel. Toyota USA Newsroom. https://pressroom.toyota.com/leading-safety-and-quality-experts-named-to-toyota-north-american-quality-advisory-panel/

Toyota North American Quality Advisory Panel (2011). A road forward: The report of the Toyota North American Quality Advisory Panel. https://www.safetyresearch.net/Library/Toyota_Quality_Report.pdf

Toyota Times (2024, October 31). Making work more fun and fruitful — Toyota’s Creative Idea Suggestion System. https://toyotatimes.jp/en/series/Imaginative_and_creative/001_1.html

United States Department of Justice (2014, March 19). Justice Department announces criminal charge against Toyota Motor Corporation and deferred prosecution agreement with $1.2 billion financial penalty. https://www.justice.gov/archives/opa/pr/justice-department-announces-criminal-charge-against-toyota-motor-corporation-and-deferred

Verweij, J. (2026a). Outcomes are caused, not decided: How a framework built on causality redefines risk and expertise and lays the groundwork for organisational design. Decision Free Solutions, Theoretical Positioning Series (in preparation).

Verweij, J. (2026b). Decision Free Solutions: A causal-perceptive paradigm unifying organisational science: How a deductive framework built on first principles reorganises what organisational science already knows. Decision Free Solutions, Theoretical Positioning Series (in preparation).

Verweij, J. (2026c). Decisions add unnecessary risk: What a precise definition of decision reveals — after seven decades without one. Decision Free Solutions, Theoretical Positioning Series (in preparation).

Verweij, J. (2026d). Hierarchical decision-making as structural pathology: What happens when authority needs no explanation. Decision Free Solutions, Theoretical Positioning Series (in preparation).

Verweij, J. (2026e, July 9). When to trust the gut when making a decision? Klein and Kahneman derived conditions for trusting ‘intuitive expertise’, DFS proposes the big picture. LinkedIn. https://www.linkedin.com/pulse/when-trust-gut-making-decision-klein-kahneman-derived-jorn-6ckje/

Verweij, J. (2026f). Leadership that stays when the leader leaves: How a functional definition of the leadership role reveals the architecture organisations need to utilise expertise. Decision Free Solutions, Theoretical Positioning Series (in preparation).

WardsAuto (2008, August 5). Toyota Japan lays off temporary workers. Ward’s staff from wires. https://www.wardsauto.com/news/archive-wards-toyota-japan-lays-off-temporary-workers/772587/

WardsAuto (2009, January 8). Toyota asking workers to take pay cut. Ward’s staff. https://www.wardsauto.com/news/archive-wards-toyota-asking-workers-to-take-pay-cut/773296/

Witcher, B. J., & Butterworth, R. (2001). Hoshin Kanri: Policy management in Japanese-owned UK subsidiaries. Journal of Management Studies, 38(5), 651-674.