Efterkalkyl: Do You Know Which Jobs Actually Make Money?

An efterkalkyl is a post-job costing review: once a job is finished you put what you quoted or estimated (the förkalkyl) next to what the job actually cost — real hours, real material, real travel — and see whether it made the money you thought it would. The number at the end is not the point. The point is the variance between estimate and actual, because that gap is the feedback that fixes your next quote. Run one on every job and a pattern shows up within weeks: a handful of job types, customers or estimators quietly lose money while the busy weeks hide them. Most small trades firms never do this, so they keep quoting the losing jobs the same way and wonder why a full calendar never turns into cash.

What efterkalkyl means: the estimate versus the outcome

The word is Swedish, but the idea is plain job-cost accounting that every trade needs. It comes as a pair.

• Förkalkyl — the pre-calculation. The hours, material and cost you assumed when you priced the job. It is baked into every quote whether you wrote it down or not.
• Efterkalkyl — the after-calculation. What the job truly consumed once the van is unpacked and the invoice is out.

Bigger contractors call it estimate-versus-actual, or budget-versus-outcome, done per job instead of per year. The value is not in either number alone. It is in laying them side by side. A quote is a prediction. An efterkalkyl is the only thing that ever tells you whether the prediction was any good, and a prediction you never check does not improve.

Why your year-end accounts hide the losing jobs

Your annual profit and loss nets everything together. A month with four strong jobs and three bad ones reads as a fair month. The three bad ones vanish into the average, so nothing tells you to stop selling them, and they keep coming back.

Job-level review breaks the averaging. It is worth insisting on in 2026 specifically because the inputs have moved. Cable, timber, insulation, switchgear and copper have all repriced since 2023, and wages moved with them. Any estimating habit you set before 2024 — "materials plus 20%", "two days will cover it", "round it to a nice number" — was calibrated against a cost base that no longer exists. The habit survives in your head long after the maths under it stopped being true. An efterkalkyl is how you catch that, one job at a time, instead of finding out at year end that a busy year barely paid.

What to put side by side

Keep both columns strictly ex VAT, or the comparison lies. Compare like for like: the same cost lines in the estimate and in the actual.

Loaded labour matters here: not the hourly wage, but wage plus employer contributions, holiday, sick, pension, tools and training. For most trades firms that is 1.35–1.5× the gross wage. If your förkalkyl used the bare wage and your efterkalkyl uses loaded cost, the variance is noise. Use loaded cost in both.

A worked example: the fixed-price job that felt fine

A two-person firm quotes a fixed price for a small commercial fit-out. Here is the förkalkyl they priced on.

Estimate (ex VAT)

• Labour: 24 h × £34 loaded = £816
• Material at cost = £1,800
• Travel and van = £120
• Estimated direct cost: £2,736
• Quoted fixed price: £4,900
• Expected gross margin: £2,164 (44%)

The job went out the door feeling like a good win. Then the efterkalkyl.

Actual (ex VAT)

• Labour: 32.5 h × £34 = £1,105
• Material at cost: £2,050 (a price rise since the quote, plus offcuts)
• Two extra wholesaler runs: 1.2 h × £34 = £40.80
• Travel and van, actual: £165
• One uncredited wrong-part return: £70
• Actual direct cost: £3,430.80
• Invoiced: £4,900 (fixed price, no variation captured)
• Actual gross margin: £1,469.20 (30%)

Variance: £695 of margin gone, and 14 points off the expected percentage. The job still made money, so nothing screamed. But the estimate was wrong in four specific, findable ways, and without the efterkalkyl every one of them gets repeated on the next fit-out quote.

Reading the variance: what each gap is telling you

A variance is only useful if you name its cause. Line by line, this job says:

• Labour ran 8.5 hours over. That is a third more than priced. Either the estimate was optimistic or the scope grew on site. Both have fixes, but only if you know which one it was.
• Material came in £250 high. Part price rise, part waste. If you quoted off a stale price list, the miss will repeat on every quote until you re-import the supplier file.
• The extras were never billed. On a fixed price, unrecorded scope is a straight gift. The customer added work, it was done, and no variation was captured, so the firm ate it.
• £70 walked out as an uncredited return. Pure margin. Book returns the same week or they become someone else's stock.

Now the next quote can change. You add a realistic hour buffer for that job type, quote off live material cost, and agree upfront how variations get signed off. That is the whole reason to do an efterkalkyl: not to feel bad about a finished job, but to price the next one right.

Turning it from a spreadsheet chore into a by-product

None of this survives if it needs a spreadsheet evening. The comparison is easy arithmetic. The hard part is capturing the actuals honestly, because memory rounds everything toward the estimate. It only works if the real numbers land on the job while the work happens:

• Hours logged against the job as they are worked, with travel kept separate from site time
• Material at true cost, pulled from imported supplier price files, with supplier invoices attached to the job (photograph them from the job card before they get lost)
• Variations recorded the moment the customer says yes, on site, not reconstructed weeks later
• The invoice built from the same records, so the estimate and the actual sit on one object and the variance calculates itself

This is exactly what a field-service system is for. In OdinTask the quote holds the förkalkyl, the geofenced time clock puts real hours on the job, supplier price-file import and job-card invoice capture put true material cost on the job, and variations and the invoice come off the same record — so the efterkalkyl is a by-product of running the work, not a monthly ordeal. Whatever you use, the rule holds: a cost you do not capture against the job is a cost that never enters your variance, and it will mislead the next quote.

Start this week: your last ten jobs

1. Take your last ten completed jobs. For each, write the estimate you priced on next to the actual cost. Do not average them — list them.
2. Look at the three with the worst variance. They will share something: a job type, a customer, a distance, or one person's estimates.
3. Re-price that pattern, or stop selling it, and change one estimating habit before the next quote goes out.

The uncomfortable finding is almost always the same: busy and profitable are different things, and the estimate you trust is the one costing you money. The good news is that an efterkalkyl turns that into a specific fix rather than a vague worry. Try OdinTask free for 14 days and see what your last ten jobs actually left behind.