Where the margin actually went
There is a meeting that happens at the end of every job or production run, and it goes the same way in almost every business I have worked in.
You sit down with your estimator and your head of production to work out why the job came in under the margin you priced. The estimator has his quote. The head of production has his timesheets. And within about ninety seconds, the two of them are pointing at each other.
The estimator says he priced the labour off historic data: this kind of work, this many hours, costed the way it has always been costed. Production says every one of those tasks took the hours the historic data said it would, and his timesheets back that up. Nobody is lying. Nobody has made an obvious mistake. The material estimate is usually close to spot on. And yet the margin is gone.
The tricky thing is, they are both right.
The estimator priced the work accurately. Production completed the work in the time the work takes. The margin did not disappear inside either function. It disappeared in the space between them, in the time that nobody owns and nobody records.
The first place it goes: the gaps between the workstations
If you only measure how long each task takes, you will never find this, because the loss is not in the tasks. It is in what happens between them.
Work in progress sits in a queue waiting for the next station to be free. A team finishes its part and the part waits. On paper this can look harmless. The line can only move as fast as its slowest process anyway, so what is the harm in a batch of work in progress sitting around for an extra half day?
Most of the time, nothing. But "most of the time" is doing a lot of work in that sentence. Occasionally it is something, and the something is expensive.
It can be a large batch built before the first quality check is done, so a defect that should have been caught on unit one is now sitting in all forty. It can be a congested floor where work in progress from one line is physically in the way of another, so everything needs moving twice. It can be damage from a forklift catching a stack that should not have been there. And every hour of it ties up cash: material and labour value sitting idle in a half-finished state that cannot be invoiced, when the same money could have been working much harder somewhere else in the business.
None of this shows up on a timesheet. The task hours are correct. The waiting hours are invisible, because waiting is not a task, so nobody times it and nobody records it. But the time is still real: it risks damage, it can mean rework, and properly accounted for it still carries the overhead. That is where the first slice of margin goes.
The second place it goes: the contingency nobody will name
The estimator does not want to price in contingency, because contingency pushes the number up and a higher number loses the job. Production does not want to ask for contingency, because asking looks like admitting you cannot hit the target.
So the plan gets priced clean. No allowance for the thing that went wrong last time, because we have made changes since then, and even if we have not, what are the chances of it happening again?
This is just how people are. We are optimistic about our own plans. Left to ourselves, we estimate the time a job will take and quietly discount the things we cannot predict: the machine that goes down, the drawing that comes back wrong, the person who is off the week we needed them. If we priced in every eventuality, every product would carry so much padding that nobody could justify buying it. A degree of optimism is the only reason anything gets quoted at all.
But there is a difference between optimism and pretending the disruption will not happen. The disruption will happen. Not every time, and not the same one twice, but something will go wrong on a high enough proportion of jobs that a business which never plans for it is structurally underpricing its own work.
And when something does go wrong, the crisis or the accident gets fixed fast and reactively, because it has to be. That firefighting almost never gets recorded: the hours, the chasing, the energy it took to put right. So it never feeds back into the record of past jobs the next quote is priced from. The cost was real, and the next price is built as though it never happened.
And it is not only what happens on your own floor. The price of your inputs can move between the day you quote and the day you build: steel, energy, the market rate for increasingly scarce labour. A quote that assumes today's prices hold for a job delivered in four months is carrying a risk nobody has named. It is one more reason a clean-priced job quietly underdelivers.
The answer is not a percentage slapped on the top of every quote. A blanket contingency percentage is just a tax on the customer that bears no relationship to the actual risk. The answer is a real, understood, agreed contingency for each product or project: backup stock, a backup machine route, a secondary supplier for key material, a staffing plan for when someone is off. These things have a cost to plan, arrange and manage. That cost can be projected, planned and accounted for properly, which is a different thing from an arbitrary uplift on the project value.
What good looks like
Two pieces of work, and they run in parallel.
First, measure the whole process end to end, not just the task times. Time the waiting. Time the movement. Time the work in progress sitting in a queue. It means being honest about how long the whole thing actually takes rather than how long the tasks on the timesheet take. Do it once, properly, and you have a true picture of your end-to-end production time. Then you attack it. Not all of it at once: find the biggest single source of lost time and go after that one. When it is down, move to the next.
Second, make the contingency internally explicit and review it consistently. Agree what each product or project actually needs to absorb the disruptions that genuinely occur, then watch it. If you are consistently not using it, you have headroom sitting in your price, and what to do with it is a board-level decision. If you are chasing revenue growth, you can hand it back as a lower price. If you are focused on profitability, you move it into margin. If you are burning through it, you dig into why and find the root cause. Either way it is a number you can see and adjust, not a fear you are managing by guessing.
Done this way, the estimator and the head of production stop being on opposite sides of the table. The price reflects the work, the risk is understood rather than hidden, and both of them are working the same data toward the same goal: a number that is as de-risked and as predictable as you can make it, with both of them incentivised to drive the waste and the contingency down over time.
The margin was never really lost in either of their functions. It was lost in the part of the business that belongs to neither of them, which is exactly why nobody was looking at it.
If you recognise this in your own operation, the hard part is usually seeing the gap clearly while you are standing inside it. That is the part I help with. If it is useful to you, drop me a line.
