What a job really costs you
Where cost comes from, why the quote and the outcome disagree, and how to find the work you should be turning down.
Most small manufacturers know their overall margin and almost none know it per job. That average hides the thing that matters: some of your work is subsidising the rest. This is about finding out which, because the answer usually changes what you quote.
Part 01
Where a job cost comes from
2 minutes
Four things accumulate against a works order while it is open. Nothing is estimated after the fact — each is recorded as it happens, which is why the recording habits matter so much.
| Component | Where it comes from | What makes it wrong |
|---|---|---|
| Material | The BOM, valued at what the stock actually cost when it was consumed. | A BOM that does not match how the thing is really built. |
| Labour | Time booked against the job, at the work centre rate. | Time booked at the end of the week from memory, or not at all. |
| Machine and overhead | The work centre running cost for the time the job occupied it. | Work centre rates never revisited after setup. |
| Scrap and rework | Material and time consumed making a part twice. | Scrap quietly replaced without being logged. |
Material is hard to hide — it leaves the shelf. Labour is easy to lose, and it is often the larger number. A job with no time booked against it looks like your most profitable work, which is exactly backwards.
Part 02
Which cost you are looking at
2 minutes
The word "cost" means three different things, and mixing them up is the most common reason two people in the same business quote different figures for the same part.
| Which cost | What it answers | Use it for |
|---|---|---|
| Standard cost | What we expect this to cost. | Quoting, and spotting when reality drifts away from the plan. |
| Actual cost | What this particular job did cost. | Judging whether that job made money. |
| Average cost | What the stock on the shelf is worth, blended across receipts. | Valuing inventory. Not for quoting. |
If you bought steel at £40 last year and £52 last week, the bar you consume today is valued from what your stock actually cost, not from the newest invoice. Quote off the latest price and you will over-quote settled work; quote off average and you will under-quote when prices are climbing. Both matter, which is why the two figures are kept apart.
Part 03
Why the quote and the outcome disagree
2 minutes
A quote is a prediction. The job is the result. The gap between them is the single most useful number in the business, and it is worth understanding the four ordinary reasons it opens up.
You quoted a run of 20 brackets at £18 each — £360. The job closes with £171 material, £140 labour, £48 machine time and one scrapped part at £9: £368 actual. You lost £8 on a job that felt fine, because the quote assumed four blanks per bracket and the shop has always used five. Run twelve times a year, that is the BOM costing you £96 nobody noticed.
- The BOM was optimistic. The drawing says four brackets; the shop has always used five because one always splits. The quote used four.
- Setup was priced as if it were free. A one-off of five units carries the same setup as a run of five hundred. Averaged across the year this looks fine; per job the small runs are losing money.
- Rework was absorbed. The second attempt consumed material and time that the quote never contemplated.
- Material moved. You quoted in March and bought in June.
Judge a quote against the actual cost of the same job, not against your average margin. An average tells you the business is fine; the comparison tells you which customer is being subsidised by the others.
Part 04
Where margin quietly leaks
2 minutes
These are the patterns that show up once you have two or three months of job costs to look at. None is visible in a monthly profit figure.
- Small runs of complex parts. Almost always the worst work in the book, and almost always quoted as a proportion of a larger run.
- The obliging customer. Frequent small changes, each absorbed as goodwill. Individually trivial, collectively a discount nobody agreed to.
- One machine everything waits for. The cost is not the machine — it is the jobs sitting around it and the overtime spent catching up.
- Scrap treated as normal. A predictable scrap rate that nobody prices is a discount you give away on every unit.
Part 05
What to actually do about it
1 minute
You do not need a costing project. You need three habits and one monthly hour.
- Book time to the job, every day. Nothing else in this article works without it.
- Log scrap when it happens. Unlogged scrap turns a losing job into an apparently profitable one.
- Correct the BOM the moment a job disproves it. The recipe is the basis of every future quote for that part.
- Once a month, read quoted against actual and sort by the gap. The bottom of that list is your answer.
Most shops doing this for the first time find between one and three regular jobs that lose money on every run. Repricing or declining those is typically worth more than any efficiency the software gives you.