There’s a moment every 24/7 team lead knows. It’s 4pm, someone has called in sick for the overnight, and the roster has a hole in it. You have two options that both feel bad: tell someone they’re staying, or work it yourself.
Teams that handle this well aren’t luckier with sick days. They’ve built a system for offering overtime before the gap becomes an emergency, and they’ve made sure the offer doesn’t always land on the same person.
This is that system. It assumes you can’t hire your way out and that mandatory overtime is a cost you’d rather not pay.
The hour costs 1.5×. That’s the visible part, and it’s the smallest part.
It converts a coverage problem into a retention problem. Being told you’re staying — after your shift, with plans already made — is one of the most reliably resented experiences in shift work. It rarely triggers a resignation on its own. It reliably contributes to one three months later, and the replacement costs far more than the shift did.
It concentrates. Managers under time pressure ask whoever is most likely to say yes. That’s the same two or three people every time, and they know it. The reward for reliability becomes more nights, which is a lesson people only need to learn once.
It degrades the shift you’re protecting. An agent working their thirteenth consecutive hour is not delivering the coverage the roster implies. You have filled the slot on paper and left the quality gap open.
In some places it’s restricted. Several US states limit or prohibit mandatory overtime for nurses, and working-time rules elsewhere constrain consecutive hours and minimum rest. If you operate across jurisdictions, “just make someone stay” may not be lawfully available. Check what applies to you.
None of this means never use it. Emergencies exist. It means treating it as the break-glass option rather than the default gap-filling mechanism.
Work these in order. Each stage costs more than the last, so the goal is to resolve as early as possible.
Most overtime is bought at a premium because it’s bought late. A gap spotted on Monday for Thursday night can be filled by someone who wants the hours. The same gap spotted at 4pm Thursday can only be filled by whoever answers the phone.
The mechanism matters less than the cadence: check coverage for the next 14 days at a fixed time each week, and again each morning for the next 72 hours. Most teams that “have an overtime problem” have a visibility problem — they can’t see gaps until the gap is today. We costed that out in the real cost of a coverage gap.
When a gap is known, publish it to everyone eligible at once and let people claim it.
This single change does more than any other. Asking individuals means you decide who gets the opportunity, which means you carry the fairness problem and the resentment that comes with it. Broadcasting means the people who want the hours self-identify, and nobody has to decline a direct request from their manager — which is a much harder thing to do than ignore a channel post.
Include four facts in every offer, or you’ll spend the time you saved answering them:
Broadcasting fixes who gets asked. It doesn’t fix who gets picked when several people claim the same slot, and if you always award it to whoever replied first, you’ve rebuilt the concentration problem with extra steps — fastest responder, not fairest outcome.
Track OT hours per person over a rolling window — a quarter works well — and when a slot is contested, award it to whoever has the fewest. Make the tracking visible. Two things follow: people stop suspecting favouritism, and the person who’s had the most stops feeling obliged to keep volunteering.
Sometimes nobody claims it. At that point you’re choosing between mandatory overtime, running short, and degrading the service deliberately — and that’s a decision worth making explicitly rather than defaulting to the first one.
Running short on a low-volume overnight is often the right answer. Whichever you choose, log it. A quarter’s worth of “nobody volunteered for Sunday nights” is a hiring case, a shift-differential case, or a rotation-redesign case — but only if you wrote it down.
If nobody ever volunteers, the offer is wrong. In rough order of cost-effectiveness:
Predictability. Standing, known OT slots that people can plan around get claimed far more readily than ad-hoc requests. Some of your team want extra hours — they just want to schedule them, not be surprised by them.
Time back instead of pay. Where employment rules and your policy allow, some people would rather have a day than the money. Costs less, valued more by the right person. Check what’s lawful for your staff classifications first.
Split the shift. A 4-hour block gets claimed when an 8-hour block doesn’t. Two people covering half each is usually easier than one covering all of it.
Ask agents on a public holiday. On a distributed team, someone else’s public holiday is an ordinary day off. If your roster tracks holiday groups per person, the people already not working today are your most likely volunteers — and they’re often glad to be asked. This is one of the few genuine advantages of a multi-country roster; see handling country-specific holidays.
Differential, last. Paying more works, and it’s the option that permanently raises your cost base. Try the other four first.
Four numbers, reviewed monthly:
| Metric | What it tells you |
|---|---|
| OT hours per agent, rolling quarter | Whether it’s concentrating. A top-to-median ratio above ~3× means it is. |
| Voluntary vs mandatory split | Your actual reliance on break-glass. Target 90%+ voluntary. |
| Lead time on OT requests | Gaps found late cost more. Rising lead time is the leading indicator that things are improving. |
| Unfilled gap hours | Where you chose to run short — the input to your hiring case. |
The first is the one that predicts resignations. If the same person tops it three quarters running, that is a countdown, and no amount of thanking them changes the arithmetic.
Is mandatory overtime ever the right call? For genuine emergencies — a mass call-out, an incident, a safety-critical gap — yes. As a routine gap-filling mechanism it’s a loan against retention at a bad rate. The distinction is whether it’s the exception or the process.
How do I stop the same people volunteering for everything? Track OT per person and use it to award contested slots. Then talk to your top volunteer: often they’re covering because they believe nobody else will, and seeing the distribution changes that. If they need the money, that’s a compensation conversation, not a rostering one.
What if nobody volunteers, ever? The offer is wrong or trust is low. Check the rate against local norms, try splitting shifts, make slots predictable — and check that past volunteers were actually paid correctly and on time. Nothing kills volunteering faster than a disputed payslip.
How far ahead should I publish OT opportunities? As far as you can see the gap. A week’s notice reliably gets volunteers; same-day rarely does. The constraint is usually how early you can see the gap, which is a tooling problem before it’s a policy one.
Should overtime be first-come or fairest-first? Fairest-first for contested slots, first-come for slots nobody else wants. Pure first-come rewards whoever watches notifications most closely, which correlates with neither need nor fairness.
How does this interact with reducing overtime overall? They’re complementary. This post is about filling gaps you already have; reducing how many gaps you generate is a rostering problem covered in how to reduce overtime costs with smarter scheduling. Do both — the second reduces how often you need the first.
We’re a 6-person team. Is this overkill? The tracking is lighter but the principles hold, and small teams concentrate worse, not better — with six people there’s nowhere for the load to spread. Broadcasting and rotating matter more, not less.
Mandatory overtime is a loan against your retention, taken out at short notice, at a rate you don’t see until someone resigns.
The alternative isn’t heroic planning. It’s four unglamorous habits: see gaps early, broadcast instead of asking individuals, rotate the opportunity using tracked hours, and log it when you break glass. Most teams that run comfortably on voluntary overtime aren’t better staffed — they just found out about the gap on Monday instead of Thursday afternoon.
If you want the tooling side of that: Manage Roster surfaces coverage gaps across the upcoming window, runs OT events where agents claim the slots they can cover, and flags agents who are on a public holiday — often your most willing volunteers. Free during beta for up to 5 agents with code BETA2026, $29/month on Pro for unlimited agents.
Related: building a fair night shift rotation and reducing support team burnout with smarter scheduling.
Overtime rules, mandatory-overtime restrictions and time-off-in-lieu eligibility vary by jurisdiction and staff classification. Verify what applies to your team before changing policy — this is not employment law advice.