How Many Employees Do You Need for 24/7 Coverage?

What is the actual arithmetic?

Seven days at twenty four hours is 168 hours a week. Every other number in this article comes out of that one.

If one position has to be staffed at all times, and you treat a full-time week as 40 hours, you need 168 divided by 40, which is 4.2 people. Not four. That 0.2 is the part managers tend to wave away, and it is the part that costs money.

Four people working 40 hours each cover 160 of the 168 hours. The remaining eight hours still have to be worked by somebody. In practice they get absorbed as overtime, every week, permanently, because the gap is structural rather than occasional.

Two things about the 40 are worth stating plainly, because they are the assumptions the whole calculation rests on.

The first is that 40 is your choice, not the law's. The Fair Labor Standards Act does not define full-time employment at all. As the Department of Labor puts it in its guidance on full-time employment, this is a matter generally to be determined by the employer. If your full-time week is 36 hours, the denominator changes and so does the answer.

The second is that 40 is where overtime starts. Federal overtime rules attach to hours worked beyond 40 in a workweek, so any plan that runs people past 40 to close the gap is a plan to pay time and a half indefinitely. That can be a reasonable decision. It is not a reasonable accident.

Does shift length change the headcount?

No, and this is the point most people get wrong.

A 24 hour day split into three eight hour shifts is 21 shifts a week. Split into two twelve hour shifts it is 14 shifts a week. Both come to 168 hours. The arithmetic is identical, so the headcount is identical.

What shift length actually changes is everything around the number:

• Handovers. Twenty one handovers a week against fourteen. Each one is a chance for something to be dropped.
• Fatigue profile. Twelve hour shifts concentrate the same hours into fewer, longer stretches.
• How the rota lands. Some patterns need a specific crew count to close cleanly. A four crew rotation on twelve hour shifts is the classic example, and it averages 42 hours a week rather than 40.
• Recruitment. Some people will only take twelve hour shifts, and some will only take eight.

Here is the same position at different assumptions about a full-time week.

The row worth sitting with is the 42 hour one. Four crews on twelve hour shifts is a genuinely popular pattern, used by the DuPont schedule and by 2-2-3 rotations like the Pitman schedule, and it gets you to four people per post. It does that by building roughly two hours of overtime per person per week into the design. The pattern is not cheating the maths. It is paying for the gap in a predictable way rather than an improvised one.

The trade only holds if you can see it. Averaging 42 hours is a reasonable decision when the overtime is visible while you are still building the schedule, and an unwelcome discovery when it first shows up in a payroll run.

What does that look like at eight hours and at twelve?

Take one post, staffed continuously, and build it both ways.

Three eight hour shifts a day. That is 21 shifts a week to fill. Put five people on a 40 hour week and you have 200 person-hours available against the 168 you need. The 32 hours of slack is what absorbs a holiday, a sick day or a training course without anybody working extra. You are paying for that slack, and what it buys is a rota that survives contact with an ordinary week.

Two twelve hour shifts a day. That is 14 shifts a week. Four crews at an average of 42 hours gives you 168 person-hours against 168 needed. There is no slack in it at all. Every absence, without exception, is covered by overtime or by a gap in cover.

That contrast is the real difference between the two models, and it does not show up in the headcount. Five people on eight hour shifts and four crews on twelve hour shifts both cover the week on paper. One of them has a margin and the other one does not, and the one without a margin is the one where the manager spends Sunday night making phone calls.

Why is 4.2 never the real answer?

Because 4.2 assumes every person on the roster is available for every hour they are scheduled, every week of the year. Nobody is.

The things that take someone off the roster are ordinary and unavoidable:

• Paid time off
• Sick days and family illness
• Training, certification and onboarding
• Jury duty, bereavement, parental leave
• The gap between somebody resigning and their replacement being useful

None of these are failures. They are the normal operating condition of employing people. A staffing model that ignores them is not a lean model, it is a wrong one, and the way it announces itself is that the schedule works on paper and falls over roughly once a fortnight.

How do you work out your own buffer?

Use your own numbers rather than an industry average, because the average is made of businesses that are not yours. The method takes about ten minutes.

Step one. Count the shifts one person is scheduled in a year. Somebody on five shifts a week is scheduled about 260 times.

Step two. Count what they actually missed. Go through last year: holiday taken, sick days, training days, anything else. Say it comes to 18 days across those three categories.

Step three. Turn that into an availability rate. 260 minus 18 is 242. Divide 242 by 260 and you get 0.93, so each person is genuinely available about 93 per cent of the time they are on the roster.

Step four. Divide the raw number by the rate. 4.2 divided by 0.93 is 4.5.

Step five. Round up. You cannot hire half a person, and rounding down means the shortfall reappears as overtime or as a gap in cover. 4.5 becomes 5.

Replace 18 with your own figure and the answer moves. At 10 days missed a year you land at 4.4, which still rounds to 5. At 30 days you land at 4.8, which also rounds to 5. That stability is the useful finding here: across a wide range of ordinary absence rates, one continuously staffed post needs five people. Four is the answer only when you have decided to run permanent overtime.

If you do not have a year of absence history, do not invent one. Track it for a quarter and multiply, and in the meantime plan on five.

What does it cost to run with four instead of five?

Cost it properly before deciding, because the decision is usually made by default rather than on purpose.

Take a single post, four people, $20 an hour. The uncovered eight hours a week get worked as overtime at time and a half, so $30 an hour, which is $240 a week and a little under $12,500 a year. For one post. Before anybody takes a holiday.

That is the visible half. The invisible half is that a four person rota has no slack in it, so every absence lands on the people who are left.

I worked as a nursing assistant in a long-term care facility, on units staffed around the clock by three shifts: days, afternoons and nights. On an ordinary week that was three people to a unit and it held. The moment somebody called out it stopped holding, because there was no bench to call. What happened instead is that one of us picked up a double, and two people covered the full twenty four hours. Nobody designed that. It is just what is left when the rota has no slack in it, and the shape is the same whether you are short on a four person rota or a three shift one.

A run of those turns into the resignation that makes the maths worse. Working out overtime correctly is worth doing before you commit to a model that generates it every week by design.

There is also a safety argument, and it is not decorative. OSHA is direct that worker fatigue raises the risk of incidents, and NIOSH's review of the evidence on long work hours links extended shifts to reduced performance and higher injury rates. A rota that only works when nobody is ill is a rota that runs tired people.

What if you genuinely cannot hire the fifth person?

Sometimes the budget is the budget. In that case, pick which compromise you are making rather than discovering it later.

Build the overtime in and pay it predictably. A four crew twelve hour pattern does exactly this, and people can plan around it because it is the same every cycle.

Use part-time cover for the gap. Eight hours a week is one short shift. A single reliable part-timer closes the structural gap without a fifth full-time salary.

Reduce the hours that need covering. The most honest question is whether the post really needs to be staffed 168 hours a week. Overnight cover is expensive, and it is worth being sure it is earning its cost rather than inherited from a decision nobody remembers making.

Cross-train so cover is possible at all. A fifth person who cannot legally or practically do the job is not cover.

What should you do with the number on Monday?

Work out your own figure, one post at a time. Take the hours the position must be staffed, divide by your full-time week, divide again by your availability rate, and round up.

Then compare it to the people you actually have. If you are short, you already knew, because you have been feeling it as overtime and last-minute calls rather than seeing it as a number. Naming the number is what turns a recurring irritation into a hiring decision or an explicit choice to keep paying for it. That is the reasoning behind putting labour cost reporting in Shift Amp from the first version rather than adding it later: the spend is most useful to a manager while the schedule can still be changed, not once it has been worked.

Once you know how many people the post needs, the next question is how to arrange them. That is a separate decision with its own trade-offs, covered in which shift pattern you should actually use and in the guide to rotating shifts.