Most 24/7 roster templates you’ll find are a grid with names dropped into it. They show you a schedule without telling you the thing you actually need to know: how many people the pattern requires, and what it does to the people working it.
These five patterns cover the realistic options. For each one: the shape, the headcount it needs, what it’s genuinely good at, and where it breaks.
| Pattern | Shift length | Min. headcount | Best for | Main cost |
|---|---|---|---|---|
| 3 × 8 fixed | 8h | 5–6 | Predictability, simple payroll | Three handovers a day |
| 2 × 12 continental | 12h | 4–5 | Fewest handovers, long breaks | Fatigue on the back half |
| 4-on-4-off | 12h | 4 teams | Stable, well-liked rhythm | Rota drifts across weekends |
| DuPont | 12h | 4 teams | Long recovery block | Seven-night stretch is brutal |
| Follow-the-sun | 8–9h | 3 regions | No night shifts at all | Needs staff in 3 time zones |
Every pattern below depends on one calculation, and getting it wrong is why rosters that look fine on paper collapse in week three.
A 24/7 operation is 168 hours a week. One person on a standard full-time week covers roughly 37.5–40 of those. So bare minimum:
168 ÷ 40 = 4.2 people ← to cover one seat, 24/7, with zero slackThat 4.2 assumes nobody takes leave, nobody is sick, nobody trains, and nobody quits. Add the industry-standard allowance for time your staff are paid but not available — shrinkage of 30–35% covering leave, sickness, training, meetings and admin:
4.2 ÷ 0.70 = 6 people ← to cover one seat, 24/7, realisticallySix people per seat. If you want two people on at all times, it’s twelve. This number surprises teams constantly, and staffing at 4–5 rather than 6 is the single most common reason a 24/7 roster runs permanently on overtime.
Shape: 06:00–14:00, 14:00–22:00, 22:00–06:00. Each person works one band consistently.
Headcount: 5–6 per seat.
Good for: Predictability. People know their hours indefinitely, which makes childcare, study and second jobs possible. Payroll is simple. An 8-hour shift is short enough that alertness holds to the end.
Where it breaks: Three handovers every day, and every handover is a chance to drop context — three times the exposure of a 12-hour pattern. Permanent nights is also the hardest assignment to staff and retain; if you run fixed bands, the night band needs a genuine differential or a rotation onto it.
Use when your work is context-heavy per shift and you’d rather trade handover frequency for shorter, sharper shifts.
Shape: 07:00–19:00 and 19:00–07:00, typically rotating days/nights on a fixed cycle.
Headcount: 4–5 per seat.
Good for: Half the handovers. Staff get more whole days off — usually a 3–4 day block every fortnight — which people value more than shorter shifts. Fewer bodies needed than 3×8.
Where it breaks: Hours 10–12 are measurably worse than hours 1–3, and if your work is cognitively demanding or safety-critical that decline is a real cost. Twelve-hour nights are harder to recover from than eight. It also assumes staff can reliably work a 12-hour day, which excludes some people entirely.
Use when handover cost is high and the work has natural quiet periods.
Shape: Four 12-hour shifts, then four days off. Usually two days then two nights, or four of one then a switch.
Headcount: 4 teams.
Good for: The rhythm people like most, in our experience. Four consecutive days off every eight is genuinely restorative, and the cycle is easy to remember and plan around. Coverage is even.
Where it breaks: The cycle is 8 days against a 7-day week, so your days off drift through the calendar — you work some weekends and miss some. That’s fine for some people and a dealbreaker for others, and you’ll find out which after they’ve joined. Team-based patterns also make individual leave harder: one person’s holiday leaves a hole their whole team feels.
Use when you have enough people to form four stable teams and your staff value long breaks over fixed weekends.
Shape: A 4-week cycle per team — 4 nights, 3 off, 3 days, 1 off, 3 nights, 3 off, 4 days, then 7 consecutive days off.
Headcount: 4 teams.
Good for: That seven-day block. Nothing else on this list gives staff a full week off every month without using leave, and it’s a powerful retention argument.
Where it breaks: Getting there costs a punishing stretch — four nights, then later three more, with short recovery between. The cycle is also complex enough that people genuinely lose track of where they are without a published roster in front of them.
Use when the seven-day block is a strong enough draw for your team to accept the front half, and you have four solid teams.
Shape: Three regions roughly eight hours apart, each working their own daytime. Nobody works nights.
Headcount: 3 regional teams, ~2 people each minimum.
Good for: It removes the night shift entirely, which removes the health and retention costs that come with it. Everyone works local business hours.
Where it breaks: It needs staff in three time zones, which is a hiring and legal problem before it’s a rostering one. Handovers happen at the seams and must be genuinely good, because context is crossing between people who never overlap in a working day. Public holidays land differently per region and will surprise you until you track them per person.
Use when you already have or can build a genuinely distributed team. We covered the mechanics in follow-the-sun scheduling explained.
| If… | Use |
|---|---|
| Work is context-heavy, handovers are risky | 3 × 8 fixed |
| You’re short-staffed and handovers cost you | 2 × 12 continental |
| Staff value long breaks, you have 4 teams | 4-on-4-off |
| Retention is the binding problem | DuPont |
| You have or can hire across 3 regions | Follow-the-sun |
Two rules that override the table:
Don’t pick a pattern you can’t staff. A DuPont roster with three teams instead of four isn’t DuPont, it’s permanent overtime with extra admin.
Ask your team before you commit. Fixed weekends versus long blocks is a genuine preference split, and the pattern people chose is tolerated far better than the pattern that was imposed — even when it’s objectively harder.
How many people do I need for 24/7 coverage? Six per concurrent seat, realistically — 4.2 for the raw hours, divided by 0.70 for shrinkage. Four or five works only if you accept routine overtime as part of the design.
Are 12-hour shifts bad? They’re a trade. Fewer handovers and more whole days off, against measurably reduced alertness in the final hours and harder recovery from nights. Fine for monitoring-heavy work, worse for sustained cognitive load.
What’s the fairest way to allocate nights? Rotate them, track the count per person over a rolling quarter, and make the tracking visible. Fixed nights are acceptable only when genuinely volunteered for and properly compensated. More in building a fair night shift rotation.
Can I run 24/7 with four people? Only with routine overtime and no resilience — one resignation collapses it. Four covers 160 of 168 hours at full-time before any absence at all. Treat it as a temporary state with a hiring plan attached.
Do I need software, or will a spreadsheet do? A spreadsheet holds a fixed pattern for a small team. It fails at the interaction between leave, swaps and coverage, and it can’t show you an overnight shift clearly — a rest day and a night shift look identical in a cell. Why Google Sheets fails for 24/7 scheduling covers the specific failure modes.
How often should I change the pattern? Rarely. Changing a rotation is disruptive and people plan their lives around it. Fix staffing levels, handovers and overtime policy first — most “the pattern is wrong” complaints are actually one of those three.
Pick the pattern you can staff at six people per seat, that your team had a say in, and whose failure mode you can live with. The pattern matters less than the headcount behind it — a well-chosen rotation staffed at four collapses, and a mediocre one staffed at six holds.
If you want the roster to show you all 24 hours at once, flag coverage gaps before they become tonight’s problem, and track holidays per person across regions: Manage Roster does that, free during beta for up to 5 agents with code BETA2026, $29/month on Pro for unlimited agents.
For the full build, start with how to create a 24/7 customer support schedule.
Working-time limits, maximum consecutive hours and minimum rest periods vary by jurisdiction and are not covered here. Check what applies to your staff before adopting any pattern — several of these exceed limits that apply in some countries.