Ask a shop why a job lost money and the answer is usually "the timber went up" or "it took longer than I thought". Both happen. Neither is where most of the money goes. The recurring losses are structural, they repeat on every job until the structure changes, and four of the five are arithmetic rather than craft.
1. Charging markup and calling it margin
This one costs more than the other four combined, because it applies to every job you have ever quoted. Markup is a percentage added to cost. Margin is a percentage of the selling price. A shop that adds 30% to a £1,000 cost sells at £1,300 and keeps 23.1% — not 30%.
The formula
price = cost ÷ (1 − margin). Not cost × (1 + margin).
On a £40,000 year of work the gap between those two formulas is several thousand pounds that was never charged. Nothing about the workmanship changed; one division did.
2. Hours nobody pays for
Quoting, sourcing, collecting timber, machine setup, sharpening, cleaning, delivery, installation travel, snagging visits, invoicing and chasing payment are all hours the job consumed. Most shops quote a rate that only covers bench time, then work a full week and bill perhaps 60% of it.
If you work 1,800 hours and bill 1,080, a rate calculated by dividing what you need to earn by 1,800 recovers only 60% of it. The fix is not to work faster; it is to divide by the hours you can actually bill.
3. Overhead that lives nowhere
Rent, power, insurance, tooling, blades, abrasives, software, accountancy, the van. None of it belongs to a particular job, so it tends to be left out of every particular job — and then paid for out of whatever margin survives.
Total your annual overhead, divide by your annual direct costs, and you have a percentage you can apply to every quote instead of hoping it is covered. For a small shop it is commonly 15–30%, which is not a rounding error.
4. Scope that moved and a price that did not
The finish changed. They wanted soft-close after all. The site was not ready and you went twice. Each is small, each feels awkward to charge for, and together they are frequently the difference between the margin you quoted and the margin you got.
The practical defence is a quote that lists what is included in enough detail that "that is a change" is a statement of fact rather than an argument. A one-line price invites scope creep because there is nothing to compare the new request against.
5. Never checking afterwards
This is the one that makes the other four permanent. An estimate is a prediction, and a prediction nobody scores does not improve. Shops that never record what a job actually cost repeat the same optimism for years, because nothing ever tells them it was optimism.
Recording actuals is unglamorous and it is the highest-leverage habit in the list. After a handful of closed jobs the pattern is usually obvious and specific — not "we should charge more" but "our last five jobs took 60% more hours than quoted", which is a 1.6× correction you can apply to the next one.
Where to start
In order of money recovered per hour spent: fix the margin formula first, because it is one calculation and it applies retroactively to how you quote everything. Then count your billable hours honestly. Then put a real overhead percentage in. Then start recording actuals, so that in three months you are working from evidence rather than from this article.