A team spends six weeks on a line. They rebalance the work, kill three walking loops, move the parts presentation to the point of use, and cut the changeover in half. The line that needed six operators now needs four.
Two people are standing there.
Whatever happens in the next ten minutes becomes the plant's actual policy on improvement, and every person on that floor will know what it is by lunch. It does not matter what the banner in the cafeteria says. The workforce reads behavior, not intent, and this is the behavior they are reading.
Most organizations only see two options here. Both are bad, and one of them is bad in a way that takes three years to show up.
Option one: take them off the payroll.
This is the option that looks most like discipline. The improvement freed up two people, so the company banks the cost of two people. Finance is happy. The number is real.
It is also the last real number that plant will produce.
But before the practical objection, the principled one — because it is the more important of the two. Eliminating jobs as a result of improvement is not an aggressive application of the Toyota Production System. It is a violation of it.
Respect for People is not a value TPS holds alongside its operational methods, the way a company might hold a mission statement alongside its budget. It is one of the two pillars the entire system stands on. Remove it and what remains is a toolkit, not a system. An organization that improves a process and then removes the people that improvement exposed has not implemented TPS ruthlessly; it has stopped practicing TPS, whatever tools are still hanging on the wall.
This gets waved off as a cultural nicety, or as something Toyota can afford and ordinary manufacturers cannot. It is neither. It is a structural requirement, and the rest of this section is about why the machinery physically will not run without it.
The people who know where the waste is are the people doing the work. Not the engineers, not the analysts, not the consultant with the stopwatch — the operator who has been running that station for eleven years and could tell you, in about ninety seconds, four things about it that are stupid. That knowledge is the single most valuable asset in a continuous improvement program, and it is entirely voluntary. Nobody can compel it. It is offered, or it isn't.
Lay someone off as a direct result of an improvement and you have not made a staffing decision. You have made a disclosure decision, on behalf of every remaining employee, permanently. From that day forward, waste that would have surfaced in a team meeting stays quiet. Cycle times get padded. The suggestion box goes empty and everyone agrees the workforce just isn't engaged.
This is not sabotage and it is not sentimentality. It is a rational response to a clearly communicated incentive. The organization asked people to help find work that could be eliminated, and then demonstrated what elimination means. The message lands, and it does not un-land.
Toyota learned this at close range. In 1950 a financial crisis forced the company into large-scale layoffs and a labor dispute severe enough that Kiichiro Toyoda resigned the presidency, accepting responsibility for it. What was rebuilt afterward was a compact: the company would treat employment security as an obligation, and in return people would give the company their thinking and not merely their hands. Every part of TPS that depends on operator initiative — surfacing an abnormality, stopping the line, improving your own Standardized Work, offering a better method — sits directly on top of that compact.
Which is why it cannot be selectively suspended in a bad quarter. You cannot ask people to engineer themselves out of a task unless they are confident about what waits on the other side. Suspend the commitment once and you have not made a one-time exception; you have disclosed the real terms, and everyone adjusts to them permanently.
Option two: absorb them back into the line.
So most plants, sensibly, don't do the first thing. They do the humane thing instead: the two people stay on the line. Nobody loses a job, nobody gets angry, the improvement is celebrated, and the team moves on to the next project.
And the savings evaporate. Quietly, completely, and without anyone deciding to let them.
Work expands to fill the people available to do it. Six operators on a four-operator line don't stand around — they slow down, spread out, absorb the walking back into the cycle, and re-invent the very motion the improvement removed. Within a quarter the line is running at a comfortable pace with six people, and the only trace of the improvement is a laminated before-and-after photo by the time clock.
The damage here is subtler than a layoff but it compounds just as badly, because of what it does to the program's credibility. Improvement teams report savings. The savings are added up. At the end of the year the sum of all reported project savings is three times larger than the actual change in operating cost, and finance notices. The next time an improvement initiative asks for funding, the CFO has learned to discount the number by an unspecified but large factor — and they are right to.
A plant can survive one layoff. It struggles to survive becoming an organization whose improvement numbers nobody believes.
The third option: give the freed labor its own cost center.
There is a third move, and it is unglamorous enough that most organizations never consider it. It has four parts.
Move the freed labor out of the direct line and into a designated cost center. Not conceptually — actually. The positions come off the line's budget and land in an account of their own. The line is now staffed at four, and it must run at four.
Deploy that group on continuous improvement work. Model line development, changeover reduction, trials, standard rebuilds, 5S resets, supporting the next improvement on the next line. This is real work with a real leader and a real backlog, not a holding pen.
Backfill line vacancies out of that group as attrition occurs. Somebody retires, somebody moves, somebody leaves for a job across town. Instead of posting the req, you pull from the improvement group. The staffing reduction happens through the front door, at the plant's natural turnover rate, without a single involuntary separation.
Read the labor cost sitting in that cost center. That number is the savings. Not an estimate, not a projection, not a calculation someone did in a spreadsheet — a figure in the general ledger that finance can audit without taking anyone's word for it.
That last point deserves more weight than it usually gets, so here is what it buys.
Why the cost center is the part that matters.
It converts a claimed saving into a booked one. The chronic weakness of every improvement program is that its savings are an argument. Cycle time dropped eighteen percent, therefore we saved this much labor — except the labor is still on the payroll, in the same cost center, doing something. Move the people and the argument disappears. The cost is somewhere specific, it is measurable to the dollar, and it moves as the group empties out through attrition.
It is a forcing function for honesty. If a team claims four positions' worth of improvement and only two people can actually be moved out, then the improvement was worth two. The cost center says so immediately and without debate. Most savings-tracking systems will never tell you this, because they are measuring the claim rather than the consequence. This one check kills phantom savings at the source, and it does more for a program's standing with the CFO than any amount of reporting.
It funds improvement capacity out of the improvement itself. The most common reason lean programs stall is that there is nobody to do the work. Improvement is supposed to happen in addition to everyone's day job, so it happens for about four months. This structure makes the program self-financing: each round of improvement produces the labor that runs the next round. That is a compounding mechanism, and there are very few of those available in operations.
It puts the right people on the improvement work. The group is made of experienced operators from the actual processes, not staff analysts with certifications. They know where the waste is because they have been standing in it. And when they rotate back to the line to fill a vacancy, they take improvement capability with them — so over several years you are not running a program, you are seeding a workforce. This is roughly how a culture forms, as opposed to how one is announced.
It makes attrition do the work a layoff would have done, at lower cost. No severance, no notification requirements, no grievances, no damage to the employment brand in a labor market where you will be hiring again in eighteen months. And no loss of institutional knowledge, which is the cost nobody puts on the slide. It also avoids the whipsaw — plants that cut in a soft quarter and rehire nine months later pay recruiting, onboarding, and the productivity ramp of a green operator, and frequently end up spending more than they saved.
It gives you a flexible capacity buffer that is already paid for. The group can absorb absenteeism, staff a launch, cover a demand spike, or run a trial without overtime or temps. Its cost is known and it is doing useful work in the meantime.
It forces a decision instead of a drift. This is the quiet one. When freed capacity is left on the line, nobody ever decides what to do with it — it just gets reabsorbed. When it is parked in a visible account with a number attached, leadership is obliged to choose: hold it as capacity for growth, release it through attrition, or redeploy it to new work. Any of those is a legitimate answer. Drift is not.
It makes improvement velocity measurable. People released into the improvement group per quarter is a hard, non-gameable indicator of whether the plant is actually improving. It is considerably more honest than counting events run or projects closed.
What breaks it.
This practice fails in predictable ways, and all of them are avoidable if you know they're coming.
It becomes a parking lot for poor performers. This is the most common corruption and it is fatal. The moment supervisors start using the group as a place to send people they don't want, assignment to it becomes a punishment, being freed up becomes a threat, and the entire psychological foundation collapses. The group should get strong people. If assignment to it is understood as developmental — and if people visibly return to the floor as team leaders — the incentive runs the right direction.
The group has no structured work. Idle people in a cost center are highly visible waste, and they will be eliminated in the first downturn regardless of the strategy behind them. The group needs a leader, a backlog, standards for its own work, and a way of reporting what it produced.
Supervisors backfill the line anyway. If the line quietly hires back to six, the company is now paying twice and the savings are negative. This requires an actual control: the line's authorized staffing level changes when the improvement is validated, and it does not change back without a decision at a level above the supervisor.
Finance was not a partner from day one. If the cost center is set up as unallocated overhead, it will be read as bloat and cut on that basis. The account needs to be understood by finance as banked productivity in transit, with an agreed method for how it converts to reported savings and how long a position may sit in it.
Nobody rotates. People should pass through, not settle. A defined tenure — six to eighteen months is typical — keeps the group fresh, keeps knowledge flowing back to the floor, and keeps the account turning over rather than calcifying into a permanent department.
The uncomfortable implication.
There is a reason this practice sits at the center of the system rather than at the edge of it.
If an improvement did not let you take a person out of the process, you did not improve the process. You made it more pleasant. That is not nothing — a less exhausting job is worth having — but it is not a productivity gain, and it will not appear in any financial statement. The person you were able to move is the proof. It is the only proof that cannot be argued with.
Which means an organization has to be able to free people up, repeatedly, for years, if it wants to improve at all. And it can only do that if the people it frees up land somewhere good. The two things are not in tension. One is the precondition for the other.
Respect for People is usually presented as the softer of the two pillars, the human counterweight to the hard operational one. That reading has it backwards. It is the mechanism that makes the hard pillar repeatable. Take it out and you get exactly one round of improvement, followed by a plant full of people who have learned to keep what they know to themselves.
Which gives you a plain test for whether an organization is actually running this system or merely holding its tools. Not how many events it ran last year, or how many people it certified, or what it reported in savings. Ask what happened to the last group of people an improvement freed up. The answer settles the question.
Where to start if you don't have this today.
You do not need a new organization to begin. You need one improvement, validated, with the position actually moved — and a conversation with finance before you move it rather than after. One line, one cost center, one number that finance agrees is real.
The structure will not be the hard part. The hard part is the first supervisor who has to run at four when they have run at six for a decade, and that is a management system problem before it is a staffing problem. It requires Standardized Work that genuinely reflects the new method, a daily check of plan against actual, and a leader who treats the first bad day as a problem to solve rather than as evidence the improvement didn't work. If that daily discipline isn't there, the line will drift back to six no matter how the accounts are drawn.
Which is the same conclusion this work keeps arriving at from every direction: the management system is what holds a gain in place. The cost center is how you prove the gain was there.
More output from your floor — and a team that keeps finding it after we've left.
Fourtre Consulting has helped manufacturers install the authentic Toyota Production System since 2007 — the management system, not just the tools. If you're weighing where to start, let's walk your floor together.