- costs
- productivity
- what to automate
What an hour of your team's time actually costs you
· 6 min read
Salary divided by contract hours isn't your real number. The four-line calculation, a worked example, and the honest limits of what it tells you.
Twelve euros an hour. That's what you get if you divide a gross salary by the hours in the contract, and it's the number most owners use when they decide whether a task is worth taking off someone's plate. Trouble is, nobody pays that hour twelve euros — your real cost sits a lot higher. This is for the owner working out what to hand off and wanting to decide on the right number. It takes ten minutes to work out, and it moves more decisions than you'd expect.
The easy number leaves three things out
Salary divided by contract hours has one thing going for it: you already have both figures. What it ignores is three costs you pay every single month.
- What you pay on top of the salary. Employer social security contributions aren't a rounding error — together they come to roughly 30% of gross pay before you even add the workplace-accident premium, which varies by sector. The 2026 rates are set out in Orden PJC/297/2026 if you want the decimals.
- Hours you pay for that aren't worked. Holiday (Spain's Estatuto de los Trabajadores sets a floor of thirty calendar days), fourteen public holidays a year, sick leave, personal leave, the morning at the doctor's. Same salary, fewer hours — and those missing hours never make it into the denominator.
- The cost of the seat. Laptop, software licences, phone, the payroll firm's share, training, the desk itself. If that person weren't there, you wouldn't be paying any of it.
None of the three is arguable. They're just missing from the number you're deciding with.
The calculation, in four lines
You don't need a nice spreadsheet. You need four figures per role:
- Annual employer cost. Gross salary plus contributions. If you don't want to be precise, multiply gross pay by 1.32 — a reasonable approximation for office work. Your payroll firm can give you the exact figure in a minute, because they already have it.
- Annual cost of the seat. Everything you spend on that person that isn't salary. In a small business it usually lands between €1,500 and €3,000, unless expensive software is involved.
- Hours actually worked. Start from the hours in your collective agreement (full-time typically runs 1,750–1,800 a year) and subtract the absences you genuinely had last year. Don't invent a percentage — go and look.
- Divide. Total cost by hours worked. Done.
Do this by role rather than by person: admin, workshop, sales. Three numbers will cover almost every decision you make this year.
A worked example
Full-time admin, €22,000 gross a year.
- Employer cost: 22,000 × 1.32 = €29,040
- Cost of the seat (laptop, licences, phone, payroll, training): €2,000
- Total: €31,040
- Contract hours: 1,792 (52 weeks × 40 h, minus 22 working days of holiday and 14 public holidays)
- Minus last year's actual absences, about 90 hours: 1,700 h
- 31,040 ÷ 1,700 = €18.26 an hour
The quick version gave you 22,000 ÷ 1,792 = €12.28. That's a gap of nearly 50%, and it isn't an accounting nicety — it's the difference between a task feeling tolerable and feeling intolerable.
Now attach the number to something real. Preparing and sending quotes takes five hours a week. Five hours across roughly 46 working weeks is 230 hours a year. At €18.26, that task costs you €4,200 a year. On the easy number it looked like €2,825. That's exactly where the decision flips: there are fixes that aren't worth €2,800 and are worth €4,200.
Your own hour works differently
Yours doesn't come off a payslip, because you aren't paid by the hour. Your hour is worth whatever you're not doing while you spend it.
A rough proxy that holds up: take last year's gross margin and divide it by the hours you personally put in. If the business threw off €120,000 in margin and you worked 2,200 hours, your hour is somewhere around €54. It's a blunt figure, but it does the one job it needs to do — it lets you look at the three admin tasks you still do on Sundays and see them at €54 an hour instead of "it only takes me a minute".
With an honest caveat: not all your hours are worth that. The hour where you close the big client is worth far more; Thursday afternoon filling in a form is worth considerably less. The point of the number isn't to admire it, it's to sort the order in which you start letting go.
What to do with the result, and what not to
What it's good for:
- Ranking automation candidates by annual cost rather than by how annoying they feel. The two lists rarely match.
- Setting a ceiling on what you're willing to pay to fix each one.
- Comparing hiring, outsourcing and automating on the same yardstick.
What it isn't good for — and this one doesn't flatter us: saved hours aren't euros until you do something with them. If you take five hours a week off your admin's plate and those hours dissolve back into the same workload at a gentler pace, you've made someone's job better, which counts for plenty, but you haven't saved €4,200. They turn into money when you stop paying overtime, when you postpone a hire you were about to make, or when those hours go to something that bills. If none of the three happens, log it as an improvement rather than a saving, and keep it out of the P&L.
A saving that doesn't move a single line in your accounts isn't a saving. It's a less punishing working day — worth having, but it goes by a different name.
And there's a category of task where this calculation decides nothing at all. A quote that sits unanswered for two days costs zero hours and can cost you the whole client. An appointment nobody confirms costs five minutes on the phone and leaves a hole in the diary. There the cost isn't in hours, it's in what slips away. Those get automated because of the risk, not the arithmetic — and if you wait for the hours to justify it, they never will.
Where to start
Work the number out this week, three roles, unpolished. Then take the five tasks that repeat most in your business, put annual hours against each and multiply. You'll find at least one that costs far more than you thought and one that costs far less. Both surprises earn their keep: the first tells you where to start, the second saves you paying to fix something that wasn't hurting much.
Once the list is ranked, the decision has mostly made itself. If the one at the top is repetitive and runs on clear rules, see whether there's an agent that already does it. And if there isn't, you at least know precisely what you're paying to keep doing it by hand.
