Meet Tsuno at Sirha Lyon 2027 · Stand S.57Learn more
EN FR
Try the demo
Menu

Team and management

Restaurant staff turnover: which French figures to use (and which to throw out)

Hiring plans, hiring declarations, rotation rates, headcounts: each French source measures something else. Which to quote, which to drop, how to measure yours.

Published on

On a desk, printed statistical reports marked up with a highlighter, next to a closed laptop.

There is no single, reliable "turnover rate for French restaurants". The serious public sources each measure something different: hiring plans (France Travail, the national employment agency), hiring declarations (AKTO, the industry's skills body, using social security data), contract rotation rates (DARES, the labour ministry's statistics office) or headcounts on a given date (AKTO, INSEE). Each is useful if you quote its year, its scope and its limits. Figures that circulate without a source belong in the bin, and the only rate that tells you anything is your own, always calculated the same way.

Tuesday, 10:15 a.m., in the office behind the kitchen

Victor runs three restaurants. He is preparing a meeting with his bank and his two partners, and he wants a figure to show that his hiring problem is not his alone. In twenty minutes he finds four: a percentage of people leaving within their first three months, a turnover rate presented as "the industry average", a replacement cost in euros, and a number of hiring plans. None of them cites the same document. Two cite nothing at all.

Selma, his HR manager, asks one simple question: "Do these figures count people or contracts?" Victor does not know. He puts his notes away and decides to go back to the sources themselves.

(Victor, Selma and their restaurants are fictional.)

Why turnover figures cannot be compared

They do not count the same unit

A hiring plan is an intention declared by an employer. A hiring declaration is a signed contract. A rotation rate is a ratio between movements and a headcount. A headcount is a snapshot on a given date. Putting these side by side is like comparing bookings, covers served and seats in the dining room.

They do not cover the same scope

In France, the hotels, cafés and restaurants branch (HCR) is defined by a collective agreement: the businesses that apply it. INSEE's "accommodation and food service" sector is a statistical scope, wider, which also includes tourist accommodation and contract catering, for example. Some surveys only cover businesses above a certain size. A figure without a scope cannot be compared with another.

Short contracts inflate the flows

In restaurants, the same person can sign several contracts in a year, especially as casual or seasonal staff. Each contract counts as a hire and then as an exit. A rate calculated on contracts can be very high while the permanent team barely moves.

They do not share a date

A survey run at the end of 2025 on hiring intentions for 2026, declarations from 2023 and a headcount on 31 December 2023 do not describe the same year. The labour market has moved in between.

The reliable French sources, and what each one actually says

SourceWhat it measuresFigureLimit
France Travail, BMO 2026 surveyHiring plans declared by employers for 2026319,000 plans in hotels and restaurants, 44% of them expected to be difficult, out of about 2.3 million across all sectorsThese are intentions, not actual hires. Says nothing about departures
AKTO, HCR 2024 report (national profile)Pre-hiring declarations (DPAE) filed in 2023, general social security scheme, temp work excluded63% of declarations for fixed-term contracts under one monthCounts contracts, not people. Every rehire counts again
AKTO, HCR statistical overview, 2024 editionEmployees of the HCR branch on 31 December 2023811,700 employees; 82% of the branch's companies employ fewer than 10 peopleA snapshot on one date. Branch scope, not the whole sector
DARES, "CDD, CDI: how have hires and terminations changed over 25 years?" (2018)Rotation rate: average of entry and exit rates29% in 1993, 96% in 2017, driven mainly by very short fixed-term contractsAll sectors, study scope, old data. Not a restaurant figure
INSEE, employment estimatesSalaried jobs in accommodation and food serviceOver 1.2 million at the end of 2024, up over the yearStatistical scope wider than the branch. Does not measure movements

We checked these figures against the publications of France Travail, AKTO, DARES and INSEE; we did not find a recent, official rotation rate specific to restaurants, so we do not print one. The DARES figure, for its part, is instructive precisely because it covers all sectors: a rotation rate close to 100% does not mean everyone leaves within the year, it means many very short contracts follow one another.

AKTO's 63% calls for the same care. It does not say that 63% of restaurant employees are on short contracts. It says that nearly two in three contracts signed in 2023 lasted less than a month. The share of people on permanent contracts on any given day is a different question, and this figure does not answer it.

The figures to throw out

Some figures keep coming back, in the trade press, on software vendors' blogs or in sales decks. Before you quote one, ask four questions: which organisation produced it, in which document, for which year, and over which scope. If one answer is missing, the figure does not leave your office.

Three families almost always fail this test:

  • the percentage of people leaving within their first weeks or months, attributed to a trade body without anyone being able to find the study;
  • the "average industry turnover rate", often dramatic, given without saying whether it counts people or contracts;
  • the cost of replacing an employee in euros: we know of no French data published with a method that can be checked.

Throwing a figure out does not mean denying the problem. Hiring difficulties are documented by France Travail, and the weight of short contracts by AKTO and DARES. The fragile figure is the one that claims to sum it all up on its own.

How to measure turnover in your own business

National figures set the context. To make decisions, you need your own. Here is a simple method, to keep unchanged from one year to the next.

  1. Pick the period: a rolling twelve months, recalculated every month. A single month means nothing in restaurants, because of the seasons.
  2. Count people, not contracts. A server who signed four fixed-term contracts in the year is one joiner, and one leaver only if she is gone at the end.
  3. Work out the average headcount: the average of the people on staff on the first day of each month.
  4. Apply the formula: rotation rate = ((joiners + leavers) ÷ 2) ÷ average headcount × 100. This is the definition DARES uses.
  5. Calculate the exit rate on its own as well: leavers ÷ average headcount × 100. It is often the one that speaks most to a manager.
  6. Separate three groups: the permanent team, seasonal staff whose end date was planned, and casual staff. A planned end of season is not an unwanted departure.
  7. Record the reason and length of service at departure: resignation, end of trial period, end of contract, termination. Someone leaving after three weeks and someone leaving after three years call for different answers.

A fictional example. Pascal's restaurant has an average of 20 people over the year. In twelve months, 12 people joined the permanent team and 10 left. His rotation rate is ((12 + 10) ÷ 2) ÷ 20 × 100 = 55%, and his exit rate 50%. Had he counted his 40 casual contracts as joiners and leavers, he would have shown over 250%, an impressive number that says nothing about his team.

Common mistakes

  • Comparing your rate, calculated on people, with a national figure calculated on contracts.
  • Changing method from one year to the next, then concluding things have improved.
  • Mixing planned end-of-season exits and resignations into a single figure.
  • Quoting a hiring plan as a hire, or a hire as a person.
  • Repeating a figure because it is everywhere, without having found the original document.

What to measure

  • The exit rate of the permanent team, over a rolling twelve months.
  • The share of departures before the end of the first month, calculated on your own joiners.
  • The time it takes a new joiner to run their station alone, recorded by their trainer.

The third indicator is often the most useful. You do not control the labour market, but you do partly control how long it takes a new joiner to become useful.

Where Tsuno fits in

Tsuno will not give you an industry turnover rate, and it does not promise to bring yours down. It works on what a departure takes away and on what an arrival costs. Employee records hold the contract, the role and the location; the onboarding path sets the steps from day one to working alone; skills are signed off and dated by a manager. You therefore have the start and end dates to calculate your own rate, and a record of what each new joiner has learned. With Tsuno San, a new joiner can ask for the procedure for their station instead of interrupting a colleague, and every action still waits for their confirmation. The details are on the Tsuno Competences page and the Tsuno San page.

Further reading

What a departure costs beyond recruitment is covered in restaurant turnover: the hidden cost of lost knowledge. On short contracts, see seasonal staff and extras ready from the first shift and per-employee restaurant software pricing: the seasonal staff trap. To measure a change without drowning in indicators, read measuring a pilot in 30 days: three numbers, not thirty and how to reduce restaurant onboarding time without cutting corners. These figures make more sense alongside the future of work in hands-on jobs.

Key takeaways

A turnover figure is only as good as its source, year, scope and unit: without all four, throw it out. Calculate your own on people, over a rolling twelve months, and track seasonal and casual staff separately.

Frequently asked questions

What is the average staff turnover rate in French restaurants?

There is no single, recent figure from an official French source that measures how many people leave restaurant jobs. Public sources measure hiring plans, hiring declarations or contract rotation rates, which are different things. The most useful figure is the one for your own business, calculated the same way every year.

Why do restaurant turnover figures vary so much from one article to another?

Because they count different things. One survey counts hiring intentions, another counts contracts signed, another counts people on the payroll on a given date. The scope changes too: the hotels, cafés and restaurants industry agreement, the wider statistical accommodation and food service sector, or only businesses above a certain size.

How do I calculate staff turnover for my restaurant?

Over a rolling twelve months, count the people who joined and the people who left, take the average of the two, divide by your average headcount and multiply by one hundred. Count people, not contracts, and track casual staff and planned end-of-season departures separately.

What does 63% of contracts under one month mean for French hospitality?

According to AKTO's 2024 national report on the French hotels, cafés and restaurants sector, 63% of the hiring declarations filed in 2023 were for fixed-term contracts of less than a month. That is a share of contracts signed, not a share of employees: someone rehired several times is counted several times.

Is there a reliable figure for the cost of replacing a restaurant employee in France?

We know of none published for France with a method that can be checked. It is better to estimate your own, role by role: the trainer's hours, the learning period, the recruitment itself, rather than reuse an amount that circulates without a source.