The divisor is billable hours, and that is the whole trap
Overhead recovery sounds like accountancy and it comes down to one division: annual overheads divided by annual billable hours.
The word doing the work is billable. Not hours worked. Not hours paid. Hours you can put on an invoice.
Quoting is not billable. Invoicing is not billable. Chasing payment, buying materials, collecting them, driving to site, and fixing your own van are all real work that somebody has to do, and none of it goes in the divisor.
Which means the rate rises when you are quiet
£40,000 of overhead over 1,600 billable hours is £25 an hour.
The same £40,000 over 1,000 billable hours is £40 an hour.
Same overheads. Same business. A 60% higher cost on every job, arriving in the year you had least work, which is precisely the year most businesses are sharpening their pencils on price.
A recovery rate worked out once in a busy year and never revisited under-recovers on every single job in a slow one. Nothing announces it. The jobs still look profitable on paper, and the bank balance disagrees.
Utilisation, and the number that flatters everything
Billable hours as a share of paid hours is your utilisation. Between 55% and 80% is realistic for most trades and small consultancies.
Anything sustained above about 85% is usually a sign that unbillable work is being counted as billable. That is worth catching, because the error runs in the dangerous direction: too big a divisor gives too small a rate, which means every job you cost comes out looking better than it actually was.
Contingency keeps its own line
The free job costing calculators are Labour plus Material plus Overhead plus Desired Profit. Jobber's is exactly that, with no contingency field at all, and it is a funnel into a paid subscription.
Contingency belongs on its own line, and folding it into the margin is worse than leaving it out. Here is why: at the end of the job you want to know which of two things happened.
- The job was priced properly and it went smoothly. Keep doing that.
- The job was underpriced and the margin absorbed it. You got away with it.
Those need opposite responses and they produce an identical profit line. A visible contingency is what separates them, because you can see whether it was consumed.
When a job loses money, look at where it turns
If the direct costs alone exceed the price, this job was mispriced or it overran. That is a problem with this job, and the fix is estimating or site control.
If it only goes negative once overheads are added, the job was priced as though the business had no fixed costs. That is not a problem with this job at all. It is the same error sitting on every job you have quoted, which sounds worse and is actually much easier to fix: you change the rate once.
The second is far more common, and it is the shape of a business full of jobs that all look fine while the year does not.
Do it again when the job is finished
Costing a quote tells you what you hoped for. It is useful and it is a guess.
Costing the same job afterwards, with the hours that were actually worked and the materials that were actually bought, tells you whether your estimating is any good. That is the number that improves the next quote, and it is the one almost nobody produces.
A business that does this on ten jobs a year knows its own estimating error and can price around it. A business that only ever costs quotes is repeating the same optimism indefinitely and has no way of finding out.
Common questions
How do I work out an overhead recovery rate?
Annual overheads divided by annual billable hours. The important word is billable. Not hours worked, not hours paid: hours you can actually put on an invoice. Quoting, invoicing, chasing payment, buying materials, driving and maintaining your own van are all real work and none of them are billable, so they do not go in the divisor.
Why does my overhead rate go up when I am quiet?
Because the overheads do not fall with the work. £40,000 of overhead spread over 1,600 billable hours is £25 an hour. The same £40,000 over 1,000 hours is £40. The rate rises exactly when you are quiet, which is the opposite of what most businesses do with their prices in a slow month. A rate worked out once in a good year silently under-recovers on every job in a bad one, and nothing tells you it has happened.
What utilisation should I assume?
Somewhere between 55% and 80% of paid hours is realistic for most trades and small consultancies. Sustained figures above about 85% usually mean unbillable work is being counted as billable, and that matters more than it sounds: if the divisor is too big, the overhead rate is too low, and every job you cost looks better than it was.
Should contingency go in the margin?
No, and this page keeps it on its own line for a specific reason. If contingency is folded into the margin, then afterwards you cannot tell a well-priced job that ran smoothly from an underpriced job that survived on its margin. Those two look identical on a profit line and they need completely different responses: one says keep doing that, the other says you got away with it.
This job lost money. How do I tell what went wrong?
Look at where it turns negative. If the direct costs alone exceed the price, it was mispriced or it overran: that is a problem with this job. If it only goes negative once the overhead line is added, it was priced as though the business had no fixed costs, and that is the same error on every job rather than a bad estimate on one. The second is far more common and, oddly, far easier to fix.
What is the difference between margin and markup here?
Margin is profit as a share of the price; markup is profit as a share of the cost. They are different numbers and the page shows both, because working on "40%" while meaning markup and believing it is margin is one of the most reliable ways to earn less than you think. The markup and margin tool covers that in detail.
Should I do this before or after the job?
Both, and the second one is where the value is. Costing a quote tells you what you hoped for. Costing the same job when it is finished, with the hours that were actually worked, tells you whether your estimating is any good, and that is the number that improves the next quote. Very few businesses do the second one, which is why so many of them keep making the same estimating error for years.