# What Staff Turnover Actually Costs Your Restaurant: A Practical Calculator (2026)

> There is no defensible flat replacement-cost figure for every restaurant role. Use this six-part method and calculator to cost your last real departure without double-counting wages or inventing turnover.

- Author: Alex Riesenkampff (Super44)
- Published: 2026-07-22
- Canonical: https://super44.ai/blog/cost-of-staff-turnover

## Key takeaways

- No current public study gives a reliable universal replacement cost for a server, cook, or restaurant manager. The familiar Cornell figures are from US hotels in 2006, not restaurants in 2026.
- Cost one real departure from your own records across six blocks: exit administration, recruitment, vacancy cover, training, ramp-up productivity, and evidenced service failures.
- Do not count every new-hire wage as a loss. Count only incremental cover, supernumerary training, displaced trainer time, and the measured productivity gap.
- Schedule stability is associated with lower subsequent turnover in service-sector research, but pay, workload, management, progression, and onboarding remain material retention factors.
- The most useful benchmark is your own cost per departure by role, reviewed after every exit and compared over time.

Ask what staff turnover costs and search results will offer a precise number within seconds. That precision is usually false. A busy cafe replacing a weekend server through referrals does not incur the same cost as a rural restaurant recruiting a head chef through an agency. The vacancy may be covered by ordinary hours, overtime, the owner, temporary labour, fewer covers, or some mixture of all four. Each route creates a different bill.

The useful question is not, “What does turnover cost restaurants?” It is, “What did this departure cost this restaurant?” **The defensible answer comes from six cost blocks in your own records, not from applying an old industry average to every person who leaves.** This guide gives you the method, a calculator, and a 30-day routine for producing a number you can act on.

## Why the famous flat figures do not fit your restaurant

**The often-repeated Cornell numbers are real hotel research from 2006, but they are not current restaurant replacement costs by role.** Tracey and Hinkin’s Cornell report cites an earlier convenience sample of 12 hotels, whose geography is not stated in the passage, with a $5,864 average. It then presents its own analysis of 33 US properties. That analysis grouped jobs by complexity and found materially different costs for lower- and higher-complexity work. It did not establish a 2026 cost for a server, cook, or restaurant manager.

That does not make the study useless. Its durable lesson is that cost changes with the job and with the components counted. Its age, geography, hotel setting, and small samples make the dollar amounts poor inputs for a current independent restaurant budget.

Monthly government labour data needs similar care. The Bureau of Labor Statistics defines quits as separations “initiated by the employee.” Its JOLTS series reports a monthly quits rate for accommodation and food services, a sector that includes more than independent restaurants. A monthly event rate is not the percentage of your team that left during a year. Adding twelve published monthly rates is not a valid shortcut to your annual staff turnover either.

**What the public benchmarks can and cannot tell you**

*Source: Cornell hotel research and BLS JOLTS*

- **They support:** Replacement cost varies substantially with job complexity, while official monthly data shows the direction of labour-market churn across accommodation and food services.
- **They do not establish:** A current dollar cost for replacing a server, cook, or manager in your venue, or your restaurant's annual turnover percentage.

If a benchmark cannot tell you exactly what population, period, cost components, and denominator it uses, do not put it in a budget. Use it as context at most.

## The six cost blocks to collect after a departure

**A clean turnover calculation separates incremental costs from ordinary payroll and records each item once.** Pull the following from invoices, time records, payroll, and POS notes. If an amount is not recorded, label it as an estimate rather than quietly treating it as fact.

**Build one replacement-cost record**

1. **Exit and administration:** Owner or manager time spent on the conversation, records, payroll changes, access, and handover, valued at a loaded hourly rate.
2. **Recruitment:** Job ads, agency or referral fees, screening tools, and interview time. Keep cash spend and internal hours visible.
3. **Vacancy cover:** Only incremental cash: overtime uplift or the premium paid for temporary labour, not the ordinary wage that would have been paid anyway.
4. **Training:** Trainer time displaced from normal work plus shadow shifts that were genuinely supernumerary, not every hour on the new hire’s payslip.
5. **Ramp-up gap:** Productive hours multiplied by the new hire’s loaded rate and an explicit estimate of the remaining productivity gap.
6. **Service failures:** Recorded cash refunds, remakes, waste, or comps that can reasonably be tied to the vacancy or ramp-up period.

*Keep notes for unusual items. A transparent estimate that can be corrected later is more useful than a precise-looking total with hidden assumptions.*

Two mistakes distort the result most. First, counting all new-hire wages as a loss even though the person is producing useful work. Second, counting an overtime shift at its full wage rather than the premium above the ordinary wage that would have been paid anyway. For owner cover, decide on an hourly opportunity value before looking at the result. That keeps the number from changing according to the point you want to make.

Do not add vague “lost morale” or a multiple of annual salary unless you can show how it was measured. Staff strain matters, but a financial model is stronger when non-financial risks sit beside the total instead of being disguised as cash. If the owner covered the vacancy or service capacity was reduced, show that opportunity cost separately from direct cash cost and use a non-overlapping period. Otherwise the same lost output can appear under cover, ramp-up, and lost contribution three times.

## Calculate your last real replacement

**Start with the most recent completed replacement, because an actual event exposes missing records faster than a theoretical annual model.** Use loaded hourly rates where possible, including the employer costs you normally include in labour decisions. For the ramp-up line, estimate the average productivity gap across the period, not the employee’s weakest first shift.

For a quick first estimate, use the role’s normal weekly employment cost — pay plus the employer costs you include in staffing decisions — then choose the nearest description of what happened. The choices convert recruitment, cover, and ramp-up into transparent fractions of that weekly cost. They are deliberately round working assumptions, not industry benchmarks. For a completed case, replace the estimate with the six-block record above.

**Roughly what did one staff replacement cost?** *(guided estimator in the web version)*

- **Normal weekly employment cost for this role:** £750
- **How did you find the replacement?:** Your own job ad — Assume three-quarters of a week for ads, screening and interviews.
  - Other choices: Referral or network / Agency or hard search
- **What happened while the role was empty?:** Extra shifts — Assume one week of employment cost.
  - Other choices: Absorbed by the team / Temps or less trading
- **How involved was onboarding?:** Normal onboarding — Assume one and a half weeks.
  - Other choices: Quick start / Specialist or leader

| Result | Example value |
| --- | --- |
| Working replacement-cost estimate | £2,813 |
| Exit admin and recruitment | £938 |
| Extra vacancy cover | £750 |
| Training and ramp-up gap | £1,125 |

*The half-week admin allowance and the choice multipliers are transparent working assumptions, not external benchmarks. Changing currency changes the symbol and number format; it does not convert your input. Exclude ordinary productive wages and do not add the same loss again elsewhere.*

Once you have records for several exits, compare like with like: FOH with FOH, kitchen with kitchen, managers with managers. Your median may be more useful than your average if one agency hire or prolonged vacancy is unusually expensive. Annual cost is simply the sum of actual departure records. If you must forecast, multiply a role-specific typical cost by the planned number of replacements, and show that replacement count as an assumption.

## What the evidence says about preventable turnover

**More predictable scheduling is a credible retention lever, but the research supports association and prediction more strongly than a guaranteed causal saving.** A peer-reviewed panel study followed 1,827 hourly retail and food-service workers. Schedule instability predicted subsequent turnover, with job dissatisfaction and work-family conflict explaining much of the relationship. A separate randomized stable-scheduling intervention at 28 Gap stores did not reduce turnover overall. Both findings matter.

The larger Shift Project survey reported six-month turnover of 24 percent among workers who received at least two weeks’ notice and 39 percent among those receiving less than 72 hours. The sample came from roughly 30,000 workers at 120 of the largest US retail and food-service firms. It was not an annual quit rate, not an independent-restaurant experiment, and not proof that changing notice alone will reproduce a 15-point difference in your venue.

**How to use the scheduling evidence**

*Source: Peer-reviewed panel research and The Shift Project*

- **A reasonable conclusion:** Unstable, short-notice schedules are associated with work-family conflict, dissatisfaction, and higher subsequent turnover among hourly service workers.
- **Do not claim:** That two weeks' notice guarantees a particular reduction, that scheduling matters more than pay or management, or that large-chain US findings are your venue's forecast.

Vendor surveys add useful colour but not causal certainty. A 2025 Sona and KAM survey of 300 UK hospitality workers found widespread frustration with processes and late rotas. Its £2.7 billion estimate covers lost productivity from manual paperwork, poor scheduling, chasing information, and repeated tasks together. It is not a turnover-cost estimate and should not be presented as one.

Look beyond the rota too. Exit notes may point to pay, workload, manager behaviour, progression, unsafe conditions, weak onboarding, commuting, or life changes. A good retention review asks which causes the venue can change and which departures were not failures at all.

## A 30-day measurement and retention routine

**The goal for the next month is to create one trustworthy baseline and remove one repeated source of avoidable friction.** You do not need a new HR platform to begin.

**Your next 30 days**

- [ ] **Reconstruct the last exit**: Complete all six cost blocks and flag every estimated input.
- [ ] **Record the reason carefully**: Use the employee’s words where available, and separate stated reason from your interpretation.
- [ ] **Audit four recent rotas**: Check notice, post-publication changes, close-open turns, uncovered shifts, and uneven unpopular shifts.
- [ ] **Check pay and workload**: Compare actual hours, breaks, overtime, and role expectations, not just hourly rates.
- [ ] **Choose one fix**: Set a fixed rota publication day, improve first-week training, or resolve the most repeated exit theme.
- [ ] **Review after the next exit**: Compare cost and cause by role; do not wait for an annual retrospective.

*Keep the operational cause and the financial cost linked in the same record. That is what turns an exit into a better decision next time.*

For the scheduling part, our [fair rota guide for small restaurants](/blog/fair-staff-rota-guide) gives a practical workflow for availability, demand, rest, publication, swaps, and emergency callouts. For the wider budget, compare the result with your [restaurant labour-cost percentage](/blog/restaurant-labour-cost-percentage), but keep replacement cost separate from ordinary weekly payroll.

## Where Super44 fits, without pretending software solves retention

**Software can reduce coordination friction, but the owner still decides what is fair, what caused an exit, and which cost is worth acting on.** Super44 can draft a weekly rota from roles, availability, approved time off, operating patterns, and connected demand data; the owner reviews and publishes it. It can also help with sick calls, swaps, and uncovered shifts, while keeping the decision with the operator. That supports one part of the routine above. It does not predict resignations or replace fair pay, sound management, and a useful conversation with the person leaving.

Super44 connects to your POS and shows you where your money leaks — then tells you what to do about it: https://super44.ai

## FAQ

### How much does it cost to replace a restaurant employee?

There is no reliable universal amount. The quick estimator starts with the role’s real weekly employment cost and uses clearly labelled assumptions for recruitment, vacancy cover, and ramp-up. For a completed case, replace that estimate with the six actual cost blocks in the guide.

### Is the widely quoted Cornell turnover-cost figure restaurant data?

No. The 2006 Cornell report concerns US hotels. It cites an earlier 12-hotel convenience sample averaging $5,864 and reports its own 33-hotel analysis by job complexity. Neither is a current restaurant-by-role benchmark.

### Should I count all wages paid during onboarding as turnover cost?

No. A new hire usually produces some useful work. Count supernumerary shadow hours, displaced trainer time, and only the estimated productivity gap during ramp-up. Counting the entire wage and a productivity gap would double-count the same shortfall.

### How do I calculate annual turnover cost?

Add the measured cost of each actual departure during the year. If records are incomplete, calculate representative FOH, kitchen, and manager departures separately, then multiply by the actual number of exits in each group. Do not multiply headcount by a guessed turnover percentage.

### Does a more predictable rota reduce turnover?

Research in hourly retail and food service finds schedule instability predicts subsequent turnover and that short notice is associated with higher six-month turnover. Those studies are largely observational, so treat predictability as one plausible retention lever, not a guaranteed cure or a substitute for fair pay and good management.

## Sources

1. [Cornell Center for Hospitality Research: The Costs of Employee Turnover, When the Devil Is in the Details](https://ecommons.cornell.edu/bitstreams/b556c92c-4579-4495-b5a3-5c2e1e9e13c7/download) — Tracey and Hinkin, 2006. Hotel data, including an earlier 12-hotel convenience sample of unstated geography and the report authors’ 33-property US analysis by job complexity; not restaurant data.
2. [US Bureau of Labor Statistics: Job Openings and Labor Turnover Survey definitions](https://www.bls.gov/jlt/jltdef.htm) — Official definitions for quits, layoffs and discharges, and other separations. Monthly rates measure events during a month and are not annual employee-turnover percentages.
3. [US Bureau of Labor Statistics: JOLTS Table 4, quits levels and rates by industry](https://www.bls.gov/news.release/jolts.t04.htm) — Current monthly accommodation and food services quits context; seasonally adjusted sector data, not restaurant-level annual turnover.
4. [Choper, Schneider and Harknett: Uncertain Time, Precarious Schedules and Job Turnover in the US Service Sector](https://pmc.ncbi.nlm.nih.gov/articles/PMC10691786/) — Peer-reviewed panel study of 1,827 hourly retail and food-service workers. Schedule instability predicted subsequent turnover; observational evidence does not by itself prove an intervention effect.
5. [The Shift Project: It’s About Time](https://shift.hks.harvard.edu/its-about-time-how-work-schedule-instability-matters-for-workers-families-and-racial-inequality/) — Survey of roughly 30,000 workers at 120 large US retail and food-service firms. Reports six-month turnover incidence by schedule notice, not an annual independent-restaurant quit rate.
6. [Stable Scheduling Study report](https://www.worklifelaw.org/publications/Stable-Scheduling-Study-Report.pdf) — Randomised bundled scheduling intervention at 28 Gap stores. It did not reduce overall turnover; a preliminary subgroup result concerned more-experienced associates. Retail, not restaurant, evidence.
7. [Sona and KAM: Poor systems and processes cost the hospitality sector £2.7 billion per year](https://www.sona.ai/blog/sona-and-kam-release-new-insights-report-poor-systems-and-processes-cost-the-hospitality-sector-2.7-billion-per-year) — Vendor-commissioned UK hospitality survey of 300 workers. Useful preference and process evidence, but the £2.7bn estimate combines several sources of lost productivity and is not a cost of turnover.