Understanding Labour Costs in a Restaurant

What I count as labour cost in a restaurant

Labour cost in a restaurant is not just the agreed hourly wage. I always calculate it on at least two levels: first the employee’s gross pay, and second the full employer cost per worked hour. Depending on the type of employment, that total can include the employer share of social insurance and other payroll-related costs. For regular social-insurance employment in Germany, the 2026 base figures are 18.6 percent for pension insurance, 2.6 percent for unemployment insurance, 14.6 percent as the general statutory health-insurance rate plus the individual additional contribution charged by the health fund, and 3.6 percent for long-term care insurance; the split is not identical in every detail, so I check the basic rule against the official values from DRV, BMG, and the DRV note on contribution sharing at DRV contribution rate guidance.

When someone on my team asks why two colleagues with a similar rota are not economically identical, the fair answer is simple: employment status, health fund, long-term care details, surcharges, and agreed hour models can change the employer cost. A practical example makes that visible. A server on 15.50 euros gross per hour does not automatically cost me 15.50 euros for each hour that person is present. I also have to allow for employer-side payroll costs, paid time that does not directly create revenue, and the organisational effort around the shift. That is exactly why an evening that looked strong on sales can still turn out weak in economic terms once the full staffing cost is honestly attached to it.

The number that actually helps me: cost per productive hour

The biggest thinking error is to look at labour cost only by month. To run the business, I need cost per productive hour. Productive does not mean every minute instantly generates sales. It means the hour belongs transparently to live operations: mise en place, service, pass work, direct cleaning linked to the shift, or cashing up at the end. If I count only the minutes spent face to face with guests, I systematically understate what an evening really costs. That is why I separate planned attendance, actually recorded working time, and the question of which of those hours belong to a shift, an area, or a revenue window. Otherwise I compare fantasy efficiency to real payroll.

A hypothetical Friday shows the difference clearly. We open from 5 p.m. to 11 p.m., but two service staff start at 4 p.m. because the terrace, coffee station, and side work need to be ready. After 11 p.m., another 45 minutes goes into closing tasks. If I look only at six opening hours, the shift appears lean. In reality I am paying for preparation and finishing time as well. With two people, eight paid hours each quickly become 16 paid hours of service coverage. Once I divide the full employer cost of that shift by only the visible opening period, I finally see what the available service capacity truly costs per operating hour instead of flattering myself with incomplete numbers.

How I connect labour cost to operating reality, not just sales

Many restaurants look only at labour cost as a percentage of revenue. That ratio is useful, but on its own it is too blunt. I compare several angles at once: labour cost per opening hour, per occupied seat, per check, per service shift, and per daypart. A lunch trade with many fast checks can have a lower ratio than an evening with more explanation, wine service, and longer table occupation. That does not automatically mean dinner is badly organised. It may simply require a different staffing pattern and should be judged against the right average check or contribution, not against the same expectation I use for a quick takeaway lunch.

In a fictional 42-seat bistro, the contrast becomes obvious. At lunch I might process 58 checks in two hours, while dinner produces only 34 checks across three and a half hours. Anyone looking only at count would cut the evening team first. But if dinner guests order more courses, ask more questions, and split bills at the table, the service time per party is much higher. So I do not judge only how much was sold; I judge how demanding the sale was to produce. Only then can I decide fairly whether the right answer is to change staffing, simplify the workflow, or recalculate prices rather than punish the wrong shift for doing harder work.

The costs people forget and then wonder why the plan misses

The most underestimated labour costs are the quiet ones: training time, handovers, rota coordination, sickness replanning, short-notice cover, idle patches caused by weather, and those awkward fragments of time when nobody is fully busy but somebody still has to be there. Paid minutes around tip clarification, shift handover, or checking a correction also belong here. None of that is waste by definition; it is part of a properly run restaurant. If I leave those hours out of the calculation, every shift looks cheaper than it is and every pricing decision becomes too optimistic. That is one of the fastest ways to create menu prices that look fine on paper and never carry the operation in real service.

Employment type matters too. With mini-jobs, I do not work from an old fixed hour limit as if it still applied regardless of pay. In Germany, the average monthly earnings limit for a mini-job is 603 euros in 2026. At exactly 13.90 euros per hour, that works out arithmetically to 43.38 hours, but there is no universal fixed hour cap that stays the same when the hourly wage changes. I check that directly with the official Minijob Centre guidance. In practice, that means I manage mini-job deployment through expected earnings and real scheduled hours, not through outdated rules of thumb or the vague feeling that someone can “probably still do one more shift.”

How I turn time data into usable labour-cost insight

If labour cost is supposed to be more than an estimate, I need clean time data. I therefore organise shifts so planned duties, actual working times, and exceptions remain traceable. Bonzumo helps here in a practical way through shift planning, time tracking, working-time accounts, and roles and permissions in the team area. The underlying process is the individual shift: I schedule the assignment, employees clock their actual start and end, and the shift lead clarifies exceptions where they occurred. Later I do not just see a total number of hours; I see the connection between planned staffing, the real course of the shift, and any meaningful deviation that needs a sensible explanation rather than a guess.

The benefit for me is not some promised automatic staffing optimisation. The benefit is a cleaner basis for decisions. If, after four weeks, I see that Saturday lunch regularly consumes one and a half more front-of-house hours than planned, I can ask the right question: is the extra time coming from setup, from slow closing work, or from sending people home too late? That kind of discussion is fairer when the time data is ordered and visible. For the wider daily picture, I also look at how transactions and closing documents line up in Understand the day’s activity before you close up. And when labour cost needs to be read against kitchen pressure rather than against sales alone, the shared order flow in Give service, kitchen and bar a shared view of orders helps me judge where time is really being absorbed.

A simple calculation logic I can actually use in my restaurant

I calculate labour cost in a sequence my team can also understand. First, I determine the gross hourly pay per person. Second, if I am doing an early forecast and do not yet have the exact payroll result, I add an employer-side uplift that is clearly marked as a planning assumption, not as a legal constant. Third, I record the real hours per shift, including setup and closing work. Fourth, I assign those hours to areas or dayparts. Fifth, I compare the resulting total not only with revenue but also with check count, seats, length of stay, and task load. That way I can see whether the problem comes mainly from prices that are too low, too many paid hours, or a workflow that is simply poorly structured.

A hypothetical example: three employees work a combined 22 paid hours on a Saturday. I estimate their average full employer cost at 22 euros per hour, explicitly as an internal planning assumption rather than an official legal figure. That makes the shift cost 484 euros. If the period generates 1,760 euros in revenue, the labour-cost ratio for that shift is about 27.5 percent. Useful, yes, but still not enough on its own. If those hours included unusual setup for a private event, I judge that Saturday differently from a normal one. I always need the number and the reason behind it. Otherwise I start “fixing” exceptional days with permanent measures and create a new problem in the weeks that follow.

What decisions should follow from labour cost — and what should not

High labour cost does not automatically mean I should cut staff. Often the better answer is a different task split, a shorter menu, clearer station logic, or a different start time for specific roles. If the kitchen waits 40 minutes every evening for the first meaningful wave of orders, that is not necessarily a headcount problem; it may be a reservation or production-timing problem. In the same way, a service team that looks expensive can in reality hold guests better, protect add-on sales, and prevent complaints. Labour cost therefore has to be read together with quality, speed, and guest flow. Otherwise I save money at exactly the point that makes the revenue possible in the first place.

It becomes fair when I use these figures as shared operating logic, not as a threat. I explain to my team which blocks of time are genuinely hard to carry economically and which deployments are worthwhile even if they look expensive at first glance. That creates practical rules everybody can understand: who sets up and when, how long the final closer should really stay, when one bartender is enough, and when understaffing ends up costing more than one extra paid hour. Understanding labour cost, for me, is not about reducing people to percentages. It is about bringing work, workload, and economic reality together in a way that keeps the business stable and makes good people willing to come back for the next shift.