A Rate Nobody Wrote Down Belongs to One Crew

The agenda was three days of line observations

A two week assessment has one scarce resource, which is time in front of the operation. So the onsite plan gets built the way most onsite plans get built, out of the data room and the production schedule. Pick the long runs. Sit on the line that has historically given trouble. Catch a changeover on the evening shift. Leave a block for the warehouse and for staging.

That plan went to the person who runs the floor for review, and they stopped it on the first pass. There were no conversations with personnel anywhere on it. The objection was not that observation is useless. It was sharper than that. The plant does not have disciplined standard work, so most of what a walk of the line surfaces is already known inside the building. The team would spend three days confirming a list the plant could have written in an afternoon.

Then came what the plan had actually missed. A fairly new VP of operations. A plant manager whose real scope is production management, with the larger title granted at hire. A slim supervisory team under both. How the place is run is a bigger observation than how the lines run.

That was correct, and the plan was rebuilt before the visit.

A rate history describes the lines, not the capability

Every plant carries two different things that look identical in a spreadsheet. One is what the equipment can do. The other is what the organization can repeat.

When standard work is written, trained, and enforced, those two converge. The rate belongs to the process. Any trained crew under any competent supervisor reproduces it inside a band, and that band is the number you can schedule against.

When standard work is not written, the two diverge, and nothing in the data tells you which one you are holding. A line that hit a strong rate on a Tuesday hit it because one operator sequenced the changeover a particular way, because a supervisor who has run that product for years was standing where they needed to be, because the crew that day had done it before. That is a real rate. It happened. It is also a personal performance, and a personal performance cannot be staffed, scheduled, or financed.

This is where it costs money. A growth plan takes the demonstrated rate and multiplies it. A capital case takes the demonstrated rate as the baseline and prices the increment above it. Both are pricing something that exists only when specific people are on the floor. Then the equipment lands, the volume lands, and the average shift delivers the average result, which was always the real capability. The shortfall gets diagnosed as a labor problem or an equipment problem, because those have owners and budgets. A method nobody wrote down has neither.

The supervisory layer decides whether any of this is visible. Standard work travels through supervision. A thin supervisory team across three shifts means the method lives in the heads of whoever has been there longest, which means the plant's capability is a roster of specific individuals rather than a documented process. That is the capacity the investment case assumed it was buying and never had.

Interview before you observe, then observe where the method is contested

Three changes, in order.

Get the narrative before the walk. The people running the floor already know where the operation loses. What they have rarely been asked for is the reason. A data room tells you a line ran at a given rate. An operator tells you what they do when the film tracks off, and whether that answer is written anywhere. Interview first, then use observation to test what you were told instead of to discover it.

Spend observation time where the method is contested, not where the runs are long. Two of the lines on that original agenda were nearly stepless: product goes into a hopper, through metal detection, into a tote or a case. Another was a manual pack operation. Hours had been budgeted to watch them because the schedule showed long runs. There is very little method to observe on a line with three steps, and the engineer cut the planned time the moment someone said so. The lines worth the hours are the ones where two crews would do the job differently.

Separate what the line cannot do from what the plant did not do. In that same record, a launch had lost about three hours waiting for product to move from the warehouse into production. Nothing about the line caused that. Capacity work that starts at the line will find the line, price the line, and leave the constraint sitting in the warehouse.

My rule after that session: when the person who owns the floor tells you your plan will only confirm what they already know, they are not being difficult. They are naming the constraint, and the right response is to rebuild the plan rather than defend it. The people who have to live with a finding after the team leaves are the ones who can tell you whether a rate is repeatable. Ask them directly. If your best crew moved to another line next month, what would this rate be?

What schedulable capacity looks like

Written standard work exists for every running line, and a supervisor who has never run that line can set it up, run it, and change it over from the document. Attainment on the same product varies inside a stated band across shifts, and someone owns that band. The gap between the best shift and the average shift is explained by something written down, not by a name. New supervisors are trained against the document rather than by following someone around. And when a rate enters a capital case or a growth plan, the person signing it can say which crew produced it and whether any other crew has reproduced it.

Before the next capacity number goes into a plan, ask who holds the method that produces it. If the answer is a person rather than a document, you have found the first project, and it is not equipment.