On call scheduling: the rules, the pay and a rota that holds up

The boiler in a 120-room hotel fails at 2:40 on a Sunday morning. Someone has to pick up the phone, and someone has to drive in. If your on-call rota lives in a group chat, you find out at that moment who is actually available, and on Monday you find out what it costs.

This guide is about on call scheduling for frontline teams: care homes, hotels, facility and cleaning crews, funeral services, event technicians. Not software incident rotations, which follow a different logic. We cover when on-call time becomes paid working time, what a late change costs where predictive scheduling laws apply, and a four-week rota you can copy.

What is on call scheduling, exactly?

On call scheduling means planning who must be reachable, and ready to work, outside their regular shifts. The person is not working yet. But they are not entirely free either.

That grey zone is the whole difficulty. Three set-ups cover almost every case we see, and they are treated very differently by the law:

The third line is the one most managers assume they are running. The second line is the one they are often actually running.

When does on-call time count as paid working time?

In the United States, the federal rule sits in two short paragraphs of the Fair Labor Standards Act regulations. Under 29 CFR 785.17, an employee who must stay on call on the employer’s premises, or so close that they cannot use the time effectively for their own purposes, is working. An employee who only has to leave word where they can be reached is not. And 29 CFR 785.15 adds the idea that sums it up: someone who waits as part of the job is engaged to wait, and that time belongs to the employer.

Europe arrived at a similar line, but by a stricter road. The Working Time Directive 2003/88/EC knows only two categories, working time and rest, with nothing in between. So the question becomes: how much does the standby restrict your life?

The case that set the tone is Belgian. In Ville de Nivelles v Matzak (C-518/15, 21 February 2018), a volunteer firefighter from Nivelles had to stay close enough to reach the station within 8 minutes during his standby at home. The Court of Justice ruled that this time was working time. Eight minutes leaves no room for anything else.

Three years later, the Court refined the test. In the Offenbach case (C-580/19, 9 March 2021), a fire service commander had 20 minutes to reach the town boundary in uniform with his service vehicle. Between 2013 and 2015 he did 126 standby periods and was called out 20 times. The Court said standby counts entirely as working time only if, looking at all the circumstances, the constraints very significantly limit the person’s ability to manage their free time. Response time and how often people are actually called out are both part of that assessment.

Concretely, the closer your callout window gets to Matzak’s eight minutes, the harder it is to argue the time is rest. A Liège care home that asks its night standby nurse to be on the ward within 10 minutes, and calls her in most nights, is in a very different place from a Ghent events company that phones a technician twice a month with an hour to get there.

Mind one thing the directive does not do. It regulates time, not pay. Whether standby earns an allowance, and how much, comes from national law, the sector agreement or the employment contract.

What does a late change to an on-call shift cost?

In most places, nothing beyond the hours worked. In cities and states with predictive scheduling laws, quite a lot.

Oregon is the clearest example. Its rules apply to retail, hospitality and food service employers with 500 or more employees worldwide, and the written schedule posted 14 days ahead must include on-call shifts, not just work shifts. Inside that 14-day window, adding an on-call shift costs one extra hour at the regular rate. And if you schedule someone on call and then do not ask them to work, you owe one-half of their regular rate for each scheduled hour they did not work.

Take a maintenance technician at a Portland hotel that belongs to a chain big enough to be covered. Say the rate is 24 dollars an hour and the Saturday on-call shift runs 8 hours. If nobody calls them, the hotel owes 0.5 x 24 x 8, so 96 dollars for that Saturday. Do that twice a month and it reaches 2,304 dollars a year for one person, for hours nobody worked.

Seattle’s Secure Scheduling Ordinance follows the same idea: an employee scheduled on call and not called in is owed half the hours not worked. Oregon also blocks scheduling anyone during the 10 hours after the end of a work or on-call shift, unless the employee asks for it or agrees. Hours worked inside that window are paid at one and a half times the regular rate.

Our guide on predictive scheduling laws covers the rest of the Oregon and Seattle rules.

How do you build a fair on call rota?

Start from the cover you need, not from the people you have. A 120-room hotel in Antwerp with four maintenance technicians needs one person reachable every night from 18:00 to 07:00, plus the weekend. That is 13 hours a weeknight and a block from Friday evening to Monday morning.

Here is the four-week rota we would suggest for that team:

Each technician holds one weeknight block and one weekend in four weeks. Nobody is on call two weeks in a row. And nobody does a weeknight block and a weekend in the same week, which is the combination that wears people out fastest.

The rota itself is the easy part, though. What makes it hold up is a handful of rules decided before anyone is assigned:

• Set the response time in writing. It decides how the time is treated legally, so it should be a choice, not a habit.
• Cap the frequency. One week in four is a sensible ceiling for a small team; one in three is already heavy.
• Protect the rest after a callout. Someone called in at 2:40 should not open the next morning at 7:00. In the EU that is the 11 consecutive hours of daily rest under article 3 of the directive.
• Put on-call slots on the same schedule as regular shifts. A separate spreadsheet hides every clash.
• Record each callout with its real start and end times. Payroll needs the hours worked, not the hours planned.

Swaps will happen. A rota with four names and a written rule for who covers whom survives a sick child on a Thursday. One without it collapses into the manager covering the night personally.

Why on-call rotas break in practice

At Shyfter, the pattern we see most often is not bad intent. It is the on-call rota living somewhere other than the main schedule. The standby list sits in a spreadsheet or a WhatsApp group, the regular shifts sit in the planning tool, and nobody sees that the person holding the Sunday night standby is also opening at 7:00 on Monday.

The second classic: callouts that never reach payroll. A technician drives in at 3:00, fixes the boiler, goes home. Nobody writes it down. Three weeks later the payslip is wrong, and that conversation costs more trust than the hours ever cost money.

And the third, quieter one, is fairness. Without a count, the same two reliable people end up carrying the standby, and they are the ones who resign first.

In Shyfter’s staff scheduling, on-call slots sit on the same schedule as regular shifts, so a clash between a Sunday night standby and a Monday opening is visible while you build the week, and Shyfter alerts you when a schedule breaks the legal rules on rest and maximum working time. Employees declare their unavailabilities in the employee app, so you are not assigning a standby to someone who already said no. Clocked hours from a callout then flow into payroll preparation as actual hours, not as a note on a whiteboard.

Healthcare teams have their own layer on top: qualifications, minimum staffing and night premiums. Our page on scheduling software for the healthcare sector goes through it.

Want to see your own on-call rota on a real schedule, with the rest checks switched on? Book a free demo and we will build it from your team, not from a prepared example.

Sources

• Code of Federal Regulations, 29 CFR 785.15 (on duty) and 785.17 (on-call time), Fair Labor Standards Act.
• Court of Justice of the EU, Ville de Nivelles v Matzak, C-518/15, 21 February 2018; RJ v Stadt Offenbach am Main, C-580/19, 9 March 2021. Directive 2003/88/EC, articles 2 and 3. Oregon Bureau of Labor and Industries, Predictive Scheduling (ORS 653.412 to 653.490, OAR 839-026). Seattle Office of Labor Standards, Secure Scheduling Ordinance, SMC 14.22.

Frequently asked questions

Do you have to pay employees for being on call?

It depends on how restricted they are. In the US, someone who must stay on the employer’s premises or so close they cannot use the time for themselves is working and must be paid, while someone who only has to stay reachable is not. In the EU, standby with constraints that very significantly limit free time, such as a response time of a few minutes, counts as working time. Many employers also pay a standby allowance set by contract or sector agreement.

How many on-call shifts per month is reasonable?

No general law sets a number, so it comes down to cover and team size. One week in four is a sensible ceiling for a small team, and one week in three is already heavy over several months. Count weekends separately, because they weigh more on people than weeknights.

Does an on-call night count toward the 11 hours of daily rest?

In the EU, if the standby counts as working time, it is not rest, so the 11 consecutive hours start again after it ends. Even when loose standby counts as rest, the hours of an actual callout are working time. In practice, plan a later start or a day off after a night with a callout.

Can employees refuse an on-call shift?

That depends on the contract and on local law. Where on-call duty is written into the job, refusing a properly scheduled shift can be a disciplinary matter. Seattle gives employees the right to decline hours that were not on the schedule as first posted, and Oregon requires its voluntary standby list to be joined by request or written agreement.