On the floor · For operations leadership
You cannot improve a process you have not standardized.
Lean settled this on the plant floor forty years ago. The same rule applies to the data your plant produces, and almost nobody has applied it. This page is the version of that argument for the person signing off on the spend.
What it costs
The bill arrives as a meeting, so nobody reads it as a bill.
When two systems disagree about a number, the organization does not stop. It schedules. Somebody pulls both reports, somebody reconciles them by hand, and a recurring meeting appears to decide which one is right this week.
That meeting is the most visible cost and the smallest one. The expensive part is every decision made in the gap between those meetings, on a number that had not been settled yet.
None of this shows up as a line item, which is exactly why it survives budget season.
There is a worked example of this on the home page: one recurring meeting about ship dates, priced at roughly $34,000 a year in salary alone. See the arithmetic.
The decision
Four ways to respond. Three spend on the wrong layer.
This is the choice in front of most operations leaders, whether or not it has been named out loud.
Fix how work gets done
Process change, no system work
What you keep: Everything. No system changes at all.
What it costs you: Fast, then it drifts back, because nothing wrote the standard down.
Add tools or modules
More software, same foundation
What you keep: Your stack, plus one more license to maintain.
What it costs you: The disagreement gets copied forward into a new place.
Start over
New system, same definitions
What you keep: Very little. Replacing is the point.
What it costs you: Quarters, at best, and the same fight on newer software.
Settle the definitions first
Standard work before AI work
What you keep: The ERP, the MES, the historian. All of it.
What goes: The disagreement is what goes. Nothing else has to.
You do not have a software problem. Your ERP already holds the numbers, and replacing it moves the disagreement rather than settling it. The cheapest option on this page is the one nobody budgets for.
What it takes
Three things, and none of them are a new system
A vague ask reads as an open-ended one, so here is the specific version.
A value stream worth settling
One, to start. Order-to-cash and plan-to-produce are the usual first choices, because they are where a disagreement reaches a customer.
Someone who can settle an argument
Definitions fail on authority, not on analysis. The work needs a person who can say which number is the number and have it stick.
Access to what you already report
The spreadsheets, the SQL, the reports people actually open. We compare them against each other. Nothing new has to be built to start.
What you end up with
Artifacts, not a slide deck.
- One written definition per metric, with one owner against each.
- A value stream map that shows where the definitions were diverging, and at which handoff.
- At least one divergence priced: traced to a real customer consequence, with a number on it.
- A way to tell when something drifts back, so the standard outlives the project.
The definitions and the value stream map live together, versioned, so the standard outlives the project and the person who wrote it. See the six steps.
Evidence
The same failure, in three industries
None of these were manufacturers, and we are not going to pretend they were. What repeats is the failure, not the industry.
Start by finding out where you actually stand.
Twelve questions a COO can walk into a room and ask, with a note on what each bad answer costs downstream. Score it yourself.
Get the Definition Audit→