In this article
Every other stage of the staff lifecycle already has an article on this site: finding staff, checking references, onboarding, letting someone go. Nothing covers the moment that comes before all of those — the unpaid "come in and show us what you've got" shift, the most common and least thought-through hiring step there is.
You've got a vacancy, a candidate calls, and you say: "come in Friday and help out for a couple of hours, we'll see how it goes." No contract, no payslip, no registration with anyone. On paper that's an extended interview that happens to take place behind the stove. In practice, within fifteen minutes there's a plate in front of a paying guest that this "candidate" cooked — and from that moment on, the question isn't whether this is a trial shift, it's whether it already became work.
Almost every hospitality business does this. Almost nobody has ever worked out exactly where the line sits, because there's no form for it and no colleague who ever explained it — it's something you "just do," the way your predecessor did it. That's exactly why it goes wrong so often: the one moment you'd want to know the answer is when a labour inspector, an employment tribunal, or a disappointed candidate asks the question instead of you.
The tension sits in one sentence that shows up across the guidance that does exist: a trial shift can test a skill, but it can't deliver value to the business. The moment a candidate plates a dish that gets sold, pours a drink that gets charged for, or is left alone on the till, that line has been crossed — no matter how short the shift was or what you called it.
This article lays out the seven numbers that draw that line: how many hours still count as a genuine test, what makes Belgium's rule stricter than most of its neighbours, what a mishandled trial shift can actually cost you, and why simply paying for it — for a shift that only lasts a couple of hours — is the cheapest insurance you'll buy this month.
Why it matters
A trial shift isn't a smaller version of onboarding — it's a separate transaction with its own risk on three fronts at once: wages you still owe once the shift gets reclassified as work, inspection risk because unpaid real labour counts as undeclared work, and an insurance gap — an undeclared "candidate" who cuts themselves on a mandoline may fall outside the venue's own accident cover.
And doing it properly costs almost nothing. A short, paid, documented trial shift is cheap insurance against something far more expensive: a bad hire that — per widely cited estimates — can cost up to 30% of a first year's wage once training, uniforms and lost productivity are counted in.
This isn't about scrapping the trial shift. It's about knowing where the line sits, so you choose to cross it on purpose rather than by accident — and so the candidate you hire got the same fair test as the four you didn't.
Ultimate guide Everything about restaurant staff From hiring to leaving — every article in this category, on one page. See the guideThe 7 numbers
Seven numbers, in the order they actually drive the decision: first the test that decides whether something is still a trial shift, then what it costs you once it isn't.
1. Where the legal line actually sits
No European law spells out the words "trial shift." What does show up consistently — in UK ACAS-style guidance, in EU labour-law commentary, in how social inspection services actually treat these cases — is the same one test: did the trial deliver real value to the business, or did it only assess a skill?
A cook who spends three minutes showing how they dice an onion is being assessed. A cook who spends that same evening cooking twenty mains that get served and charged for is doing labour — whatever anyone chose to call it. The distinction isn't the label you put on it; it's what actually happened that night.
2. How many hours before it stops being a test
There's no single number that applies everywhere — but the direction most guidance points in is strikingly consistent. UK employment-law advice (ACAS-style guidance and the law firms that build on it) cites 1 to 2 hours as a reasonable window to assess a skill, and explicitly warns that a full shift — or multiple shifts — should be paid at least the national minimum wage.
Those same sources put the tipping point at roughly 3 hours: past that, it becomes steadily harder to argue you were only assessing, and steadily more likely you simply had an extra pair of hands on the floor.
Each from a separate, publicly checkable source — together the reason to hold your own trial-shift habit up to the light.
Sources: ACAS-style guidance and the UK employment-law firms that build on it (number 1); Belgium's unified-statute law, 26 December 2013 (number 2); U.S. Department of Labor, widely echoed in HR literature (number 3); hospitality staff-turnover benchmarks (number 4).
3. Belgium doesn't have a grace period anymore
Anyone assuming Belgium still has a statutory "trial period" during which real work can stay unpaid is thinking of a rule that stopped existing on 1 January 2014. The Law introducing a unified employment statute (26 December 2013) abolished the standard trial clause for ordinary employment contracts — there is no statutory grace window left once someone is genuinely working, only the narrower "this was recruitment, not employment" exception from numbers 1 and 2 above.
That makes Belgium stricter than neighbours that still allow an informal settling-in period: the moment a Belgian "trial shift" crosses the line from step one, there's no middle ground left — it's wages, from the first hour, with no exception for "the first time."
4. What a reclassified trial shift actually costs
If an unpaid trial shift is later reclassified as work — by an inspection, an employment tribunal, or simply a disappointed candidate who files a complaint — you owe the wages for those hours after the fact, plus the employer's social charges on top. That's rarely the big number; the real cost sits in what follows the finding: an undeclared-work file rarely stays confined to one shift, and the conversation that follows almost always costs more time and stress than the few dozen euros you could have paid up front.
The calculator further down this page sets that figure next to what paying it properly would have cost — the gap is usually smaller than you'd expect.
5. The insurance gap during an unpaid trial
Everyone formally on your payroll is covered by the venue's own workplace accident insurance. An undeclared "candidate" who cuts themselves on a mandoline, bumps into a hot pan, or falls off a stepladder may sit outside that cover — precisely at the moment the cost could be highest.
That's why "it's only a couple of hours" is the wrong question. The right one is: if something goes wrong right now, who's actually covering it? For a properly paid, briefly documented trial shift, the answer is clear. For an informal, unpaid arrangement, it often isn't.
The same arithmetic as the calculator below, run for a longer trial each time — the amount stays small, even once the shift no longer does.
Compare that with the €8,714 a bad hire could cost you instead — number 3 above, run on the wage entered in the tool.
6. Why just paying for it is the cheaper bet
Put the cost of a properly paid two-hour trial shift next to what a bad hire can cost, and the comparison isn't close: a widely cited estimate (originating with the U.S. Department of Labor, since echoed across international HR literature) puts the cost of a bad hire at up to 30% of a first year's wage once training, uniforms, lost sales and the cost of hiring all over again are counted in.
And that's exactly where a good trial shift earns its keep: roughly 40% of all hospitality staff turnover happens in an employee's first year — precisely the group a short, honest test up front is meant to filter out. A few euros now, against a multiple of that later.
7. How to run a trial shift that holds up
The version that holds up under both an inspection and an honest conversation with the candidate is simple: cap it at 2 hours or less, say clearly up front that it's a test, never leave the candidate alone on the till or handling an order that counts, and pay for it anyway — €37 is, in most cases, less than what a dispute afterwards costs.
That last part is the easiest thing to forget and the cheapest thing to fix: a short, paid, documented moment is immediately legible for everyone involved — the candidate, an inspector, and you.
Work it out for yourself
The figures above are illustrative — the exact rules differ by country and by sector. What stays the same everywhere is the arithmetic: hourly wage, hours, and how many trial shifts like this you run in a typical month.
Enter your own numbers. The tool sets the cost of doing it properly next to what a bad hire could cost you instead.
What does your trial shift actually cost?
Enter your own numbers — the rest works itself out.
—
Assumption: 27% employer social charges on top of gross wage — a common EU-ballpark rule of thumb, not an exact figure for your own sector.
This is a planning tool, not legal advice — trial-shift rules differ by country and by sector. Check with your local labour authority or payroll provider if in doubt.
The gap between the first and third tile above is this article's whole argument in one glance: a few euros now, against hundreds or thousands later.
Keep that number somewhere visible — on the kitchen whiteboard, in the folder of applications — so the next time someone says "come in and show us what you've got," it automatically comes with: and we pay you for it.
How to run a trial shift that holds up
Three moments, three choices — plan them before the candidate walks in, not after.
Before the trial shift
- Say clearly, out loud, that this is a test, not a hire — a text or email confirming it is even better.
- Agree a length of 2 hours or less, and stick to it.
- Decide in advance who's supervising — a candidate left alone stops being a candidate and becomes staff.
During the trial shift
- Never leave the candidate alone on the till or handling an order that counts — that's exactly the moment the line from number 1 gets crossed.
- Test skill (knife technique, pace, how someone handles pressure), not output — a trial shift that turns as many covers as a normal one IS a normal one.
After the trial shift
- Pay for the hours, even if you don't hire — €37 is the price of a fair test, not a favour.
- Note down what you saw: date, length, who supervised. Those few lines are exactly what makes the difference if the question is ever asked.
Where this leaves you
A trial shift isn't a free test drive — it's a short, targeted assessment with a clear line: the moment real value flows to the business, it's work, and in Belgium there's no statutory trial period left to soften that.
The fix costs almost nothing: keep it short, say so up front, and pay for it. €37 for a fair test is a fraction of what a bad hire can cost you — and it's the difference between a trial shift everyone can feel good about and a dispute nobody wanted to have.
The same discipline that matters here — short, clear, documented — is exactly what the rest of the hiring cycle on this site recommends too: from the first reference check to the day someone eventually leaves.