How Much Notice Does a Duty Roster Really Need?

The short answer for my team and my guests

In Germany, a duty roster does not automatically need a specific statutory notice period. The often-mentioned four-day rule applies to on-call work under Section 12 TzBfG, not across the board to every normal restaurant roster. In other cases, the key points are the employment contract, any collective agreement, any works agreement, and the employer’s right to issue instructions within fair discretion under Section 106 GewO. Relevant jurisdiction and status: Germany, sources verified as of 30 September 2026. So if someone asks me whether “tomorrow for the day after” is enough, I do not give a blanket yes before I know which contract model applies.

For my business, that means I do not plan on the edge. Even where the law does not force every roster to be published four days in advance, my team still needs reliable notice so childcare, second jobs, medical appointments, and ordinary life can work. My operating standard is therefore a published roster at least seven days ahead, ideally fourteen. Only then do I consider the schedule dependable rather than just a patch for open gaps. A hypothetical example: if I staff Saturday evening only on Tuesday, I may not necessarily have broken a law, but I have very likely already created a fairness and leadership problem in the team.

When the four-day rule applies and when it does not

The four-day notice rule matters, but only in the right context. It applies to on-call work under Section 12 TzBfG. That is a different model from a regular shift roster with an agreed weekly working time. So if someone is explicitly employed on an on-call basis, I must give at least four days’ notice for the assignment. For a regular service employee with an agreed volume of hours, that does not create a general four-day duty for every shift plan. In practice, this distinction is often blurred, especially when short-notice changes are sent by chat or app and start to feel like a permanent expectation of availability.

I organise that difference clearly so no one has to remain constantly reachable out of uncertainty. For regular shifts, I publish a binding roster version; when someone is absent, I actively ask for extra cover instead of silently presuming it. A hypothetical situation: Lisa works a fixed 30 hours per week, Mehmet is on call for events. I do not manage them as if they were the same. Lisa’s shifts belong in a reliable weekly structure. Mehmet may only be deployed within the on-call rules. That distinction does not belong only in the contract; it also has to be enforced through everyday management discipline in the restaurant.

How I tell whether a notice period is fair and sustainable

For me, a duty roster is only fair when it is not merely distributed, but realistically livable. I check four points: first, enough lead time for private planning; second, predictability of recurring shifts; third, genuine voluntariness for extra services; and fourth, the strain placed on key people. A restaurant can work operationally and still be unfair if the same bartender is always expected to step in at short notice. The employer’s power to instruct does not stop at simply entering a shift; it is limited by fair discretion under Section 106 GewO. That is exactly why I document why I assign shifts the way I do and when I really need a change.

A concrete example from a typical, purely hypothetical week: Friday brings a city festival, Saturday a private event, Sunday a brunch service. If I have known all of that for weeks, a late roster publication is poor organisation, not an “unavoidable” surprise. It looks different when someone calls in sick the same day. Then the issue is no longer the notice period of the original roster, but fair replacement planning. I first ask according to the open task: who can do the closing shift, who knows the terrace section, who is authorised for cash closing? That way I do not simply burden the next available person, but the right one for the actual need.

The legal limits I check before I publish any roster

Notice alone does not make a roster lawful or sensible. Before publishing, I also check the working time limits under the German Working Time Act. Daily working time is generally eight hours and can be extended to up to ten with balancing, see Section 3 ArbZG. Breaks are mandatory after more than six hours, and longer breaks after more than nine, under Section 4 ArbZG. Between workdays there is generally a rest period, with a narrow hospitality-specific exception under Section 5 ArbZG. A late-published roster tends to increase exactly the kind of mistakes these rules are meant to prevent.

For my team, that means I do not publish a roster that merely “looks manageable.” A hypothetical example: someone finishes on Thursday at 00:30 and is then supposed to set up breakfast service on Friday morning. Even if that person agrees, I first check the rest period before I confirm the assignment. The same applies to split or double coverage around a banquet and evening service. Guests usually notice those planning errors only indirectly, through tired staff, delayed handovers, or rushed breaks. Good notice therefore helps me not only on the human side, but on the professional side too, because it gives me time to correct the roster before it becomes binding.

Especially tricky: second jobs, works councils, and silent roster changes

A common hospitality case is staff with more than one job. In that situation, working time is added together under Section 2 ArbZG. Two individually acceptable shifts can become impermissible once combined. So if someone agrees to a Sunday lunch shift with me, I may need to ask whether they are already booked for an evening shift elsewhere. If a works council exists, the timing and distribution of working time can also be subject to co-determination, and technical systems used for monitoring can be as well; the basis is Section 87 BetrVG. In that setting, notice periods are not just a leadership style issue, but also a participation issue.

Silent changes are, in my view, the biggest trap. A roster is published, then individual shifts get moved “just quickly” in a chat, and by the end nobody knows which version actually applies. That is exactly what I avoid through a clear publication status and defined responsibilities. Give your team a clear start to every working shift shows the organisational framework behind that: planning, publishing, and separating rights. My team should be able to see whether something is a draft, a request, or a confirmed change. Without that distinction, any notice period becomes practically worthless because the roster loses its binding quality.

How I set a sensible notice period in a restaurant

My rule of thumb is simple, but not arbitrary: regular roster fourteen days in advance, binding no later than seven days before, and after that only changes for a real reason. That is not a claimed legal norm, but my own organisational decision for the business. It fits well with reservation patterns, purchasing lead times, and holiday coordination. If I plan on Wednesday for the week after next, I can distribute known groups, expected terrace demand, prep workload, and closing routines properly. Later planning may feel convenient in the short term, but it creates more questions, more shift swapping, and more hidden extra work for management.

I do not set the notice period in the abstract, but according to volatility and responsibility. For a steady lunch business, I can plan very early. For weather-dependent outdoor service, I plan the core staffing early and keep additional demand as an explicitly open option. The wording matters: open means open, not secretly reserved. A hypothetical example: for Saturday I fix six service staff and mark two more as possible reinforcement in good weather. Those two people are not automatically scheduled. That helps me avoid the feeling that everyone is permanently on call, even though there is no on-call contract in place.

How I organise the process fairly, clearly, and in a way guests can feel

In day-to-day operations, I work in three steps. First, I build a draft from opening hours, reservations, known events, and recurring tasks. Second, I check strain, qualifications, and legal limits. Third, I publish one clear roster version from which any later changes must be actively confirmed. For that organisational basis, a clean reservation view and a shared daily context help a lot, as in A well-prepared service starts with a clear table plan. That way I am not planning bodies on paper, but real tasks: reception, terrace section, runner, closing duty, brunch setup, or sole responsibility at the bar.

When an employee asks how much notice a duty roster needs, the real question is often very practical: can I plan my life, or do I have to keep waiting? My fair answer is therefore this: the law does not always impose one rigid notice period, but in practice a roster needs enough lead time to be lived as a binding plan. Guests are affected as well, because stable shifts usually mean calmer service. If my business regularly plans only at the last minute, that is not a sign of flexibility; it is a sign of poor preparation. Good management shows itself in keeping short-notice exceptions rare and making the normal case early, clear, and understandable.