What Staff Turnover Really Costs a Restaurant

Nobody invoices you for a resignation, so the cost gets spread across four accounts and never adds up anywhere. Here is how to total it, and what to do once you have.

The cost with no invoice

Ask an operator what their food cost is and you get a number. Ask what their labour cost is and you get a number. Ask what turnover cost them last year and you get a pause, then a guess, then usually a shrug. It is not carelessness. There is genuinely no report that shows it, because the cost is dismantled across the accounts before it ever reaches a page.

The recruiting fee is an administrative expense. The training hours are indistinguishable from any other payroll. The overtime that covered the gap looks like a busy week. The productivity lost while a new hire found their feet is not recorded anywhere at all — it shows up only as slower service, more comps and a manager who spent the shift on the pass instead of in the office.

Four fragments, none individually alarming. Assembled, they are often the third largest controllable cost in the business, behind food and behind wages themselves. The turnover cost calculator assembles them, and the first time most operators see the annual figure they assume the arithmetic is wrong.

Pricing one departure

Take a line cook on $18 an hour, loaded. They resign on a Tuesday with two weeks' notice. Here is what leaves your bank account.

You spend around $400 finding a replacement — job board listings, the manager hours spent reading applications and interviewing, the two candidates who did not turn up for their trial, and the onboarding admin. You then pay roughly 40 hours of training once you count both the new hire's paid time and the senior cook standing next to them instead of running their own section: $720.

Between the resignation and the new starter working a section unsupervised, roughly five weeks pass. Those shifts do not vanish. They are absorbed by the cooks who stayed, at time-and-a-half, at around 30 hours across the gap: $810. And for six weeks after that, the new cook is producing perhaps 70% of what the person who left produced, which at a 30% shortfall across a 40-hour week is another $1,296.

Total: $3,226 for one line cook. If that kitchen employs twelve people and loses eight in a year, it is spending $25,808 annually to end up exactly where it started. That is not an efficiency problem. It is a number roughly the size of most independents' entire marketing spend.

What the rate actually tells you

Hospitality turnover runs high everywhere, and a headline rate on its own is close to meaningless. A city site staffed largely by students will always show a higher number than a neighbourhood restaurant with a settled team, and neither figure says anything about how well either is managed.

The useful split is between departures you could not have prevented and departures you could. Someone finishing a degree and moving cities is the first kind. Someone leaving after nine weeks because the rota came out on a Saturday night and they could never plan anything is the second. Both cost you the same $3,226; only one is a management problem.

Sorting a year of departures into those two piles takes about twenty minutes and is more informative than any benchmark. If the avoidable pile is small, your rate is structural and you should focus on making replacement cheap and fast — a standing pipeline, a shorter induction, a written training plan. If the avoidable pile is large, the rate is a symptom, and recruiting harder will not touch it.

One pattern shows up almost universally when operators do this exercise: early departures cluster hard. A disproportionate share of leavers go within the first ninety days, which means you paid the full recruiting and training cost and recovered almost none of it. That is the cheapest turnover to fix and the most expensive to ignore.

How it distorts your labour line

Turnover does not sit quietly in its own corner. It expresses itself as a labour cost that runs a point or two above what the schedule predicted, week after week, with no obvious cause. The usual explanation offered is that trade was softer than forecast.

Sometimes it was. More often the schedule was written for a fully staffed team, and the hours actually worked included training shifts, overtime cover for an open role, and a supervisor spending service watching a new starter rather than running the floor. None of those hours appear as turnover. They appear as labour, and they push your labour cost percentage up while the schedule on the wall insists everything is fine.

The same distortion reaches food cost, though less obviously. Inexperienced cooks over-portion because they have not yet built the muscle memory for a spec, mis-trim because they are working slower under pressure, and generate more remakes. A kitchen that is permanently one-third new runs a measurably higher food cost than the same kitchen fully staffed, on the same menu with the same recipes and the same suppliers. That is why turnover belongs in any serious conversation about prime cost rather than being filed under human resources.

What moves the number

In rough order of return on effort, and with the cheap ones first:

Notice how few of these are about money. Pay matters, and underpaying relative to the sites around you will cost you people no matter what else you do. But among restaurants paying broadly market rates, the ones that keep staff usually differ on scheduling, induction and whether anyone noticed the new starter having a bad first month.

The trap to avoid

The natural response to a turnover problem is to recruit harder. Post more listings, use an agency, shorten the induction so people reach the floor sooner, keep a permanent pipeline of candidates warm.

All of that reduces the time a role sits open, which is genuinely worth something. None of it reduces the number of roles that open. If the reason people leave is a rota published at the last minute or a section that is chronically one person short, faster recruiting simply cycles more people through the same conditions at $3,226 a head. The visible number — days to fill — improves while the expensive one holds steady.

It is the same mistake as buying more stock to solve a waste problem. You address the symptom you can see and leave the cause exactly where it was. Total the cost first, split your leavers into avoidable and unavoidable, and let that split decide whether you are solving a recruitment problem or a retention one. They look identical on the surface and have almost nothing in common underneath.

Frequently asked questions

What counts as a departure for this calculation?

Anyone who left and had to be replaced, whether they resigned, were dismissed, or simply stopped turning up. If you spent money recruiting and training a replacement, it counts. Internal transfers and promotions do not, unless they opened a role you then had to fill externally.

How do I estimate recruiting cost if I do not use an agency?

Add your job board spend for the year and divide by the number of hires it produced, then add manager time at their hourly equivalent for screening, interviewing and onboarding. Two to four hours per hire is typical. Most independents land between $200 and $600 for hourly roles.

Is high turnover always a management problem?

No. Some of it is structural — student workforces, seasonal trade, transient neighbourhoods. The way to tell is to sort a year of leavers into avoidable and unavoidable. If the avoidable pile is small, focus on making replacement cheap and fast rather than on retention.

Does turnover affect food cost as well as labour?

Yes, though it is rarely attributed. New cooks over-portion, mis-trim and produce more remakes than experienced ones. A permanently understaffed or perpetually new kitchen will run a higher food cost on an unchanged menu, which is why food cost variance sometimes has nothing to do with suppliers or pricing.

Should I count management turnover the same way?

Use the same four buckets, but expect much larger figures. Agency fees run as a percentage of salary, searches take months, the ramp-up period is a quarter or more, and a departing manager frequently takes staff with them. One management exit can cost more than a year of hourly turnover.

How quickly can turnover cost come down?

Scheduling and induction changes show up in early attrition within a quarter, because that is where most preventable departures sit. The full annual figure takes a year to re-measure. Track ninety-day retention monthly in the meantime — it moves early and predicts the annual number well.

Run the numbers

Use the free Employee Turnover Cost Calculator to apply everything above to your own figures.

Employee Turnover Cost Calculator All guides All calculators