Job margins in manufacturing: why you find out when the work is already done

The job is closed, the customer has paid, and only then does the back office put the numbers together: hours, materials, suppliers. Sometimes the result is good, sometimes a lot worse than expected. Either way it arrives weeks after the moment when something could still have been done about it.

This is not an accounting problem. The numbers, when they arrive, are correct. It is a timing problem: a margin you learn about once the work is finished is a photograph, not a management tool. And because the photograph always arrives late, the company learns very little even from the jobs that went badly.

Why the quote and the final cost never talk to each other

The quote is born between sales and the technical office: estimated hours per stage, a bill of materials, a price. The final cost is reconstructed afterwards from supplier invoices, picking notes and hours collected loosely. They are two documents with two separate lives, often in two different systems, and they meet exactly once, at the end. Neither is wrong: they simply never talk to each other while the work is in progress.

The three places where the margin disappears without a trace

• Unplanned hours. Rework, changes asked for verbally, setups that took longer than expected. They go unrecorded precisely because they were not planned, so they vanish from the count.
• Changes accepted without reopening the quote. The customer asks for a modification, production carries it out so as not to stop the line, and the price stays what it was.
• Materials consumed beyond the bill of materials. Scrap, replacements, urgent purchases paid at full price because they were needed yesterday.

The shop floor knows about all of this. The point is that it stays personal knowledge belonging to whoever was there, and never becomes a number someone else can read while it still matters.

What to track during, not after

You do not need a full cost-control system, and it is not the right first step for an SME. You need three pieces of information kept up to date while the job is still open.

• Hours recorded per stage, not against the whole job. If you only record the total you learn that you overran but not where, and the information is useless for the next quote.
• Materials committed against the bill of materials, updated when they are picked rather than when the invoice arrives, which can be months later.
• Changes recorded as events: who asked, when, and whether they were priced. An unpriced change is a cost nobody decided to take on.

The threshold that turns data into a warning

Tracking is not enough: data only helps if someone looks at it at the right moment, and nobody opens a report every morning. The simplest rule that works is a threshold. When a stage passes a set percentage of its quoted hours without being finished, say seventy per cent, the job flags itself to whoever is running it.

From there the decision goes back into the company's hands, and it is a decision that no longer existed once the work was done: absorb the overrun because the customer is strategic, renegotiate the change that caused it, or stop and revise the schedule. Three different roads, all of them better than finding out at the end.

What does not need to change

Your management system or ERP stays where it is for master data, inventory and invoicing. You need no new shop-floor terminals and no software to replace what already works. What gets added is the link between the quote and actual progress, in a place where both are visible to the same person at the same time.

The sign that it is working

It is not the end-of-job report, which existed before and arrived late anyway. It is the first time someone steps in on a job that is still open because they saw a number climbing, and brings it back to break-even instead of a loss. From that point on the next quote is built on real data rather than memory, which is the real long-term gain.