Why Restaurant Turnover Comes in Waves — and How to See the Next One Coming
September's departures didn't come out of nowhere. They were decided weeks ago, and the next wave is already forming. Here's how to see it coming.
Now that the September Exodus has come to an end, take a look at who left. Did any of those departures really take you by surprise? The line cook who carried you through July. The server who never turned down a double. It may have felt sudden. But it wasn't.
September is the biggest month for documented departures on the Push platform, every year without exception. In 2025, 8,807 people left in September, up 24% in two years.
By the time someone hands in their notice, the decision is weeks old. The resignation is the last step, not the first. The good news is that turnover runs on a calendar. The bad news is that the next wave is already in motion.
The Wave That Just Hit
September has topped the departure calendar in Push data three years running: 7,093 documented departures in 2023, 7,417 in 2024, and 8,807 in 2025.
One pattern tells most of the story. Time-off requests peak ~30 per location per month through June, July, and August. Then in September they drop off sharply, right as departures hit their annual high.
September isn't the month employees want time off. It's the month they leave.
September departures aren't the only figure increasing. Departures as a whole are rising. In 2021, 58% of hires on the Push platform were for growth. By 2025, that share was 18%. The other 82% were backfills. When almost every hire is filling someone that has left, a seat you keep is worth more than a seat you fill.
Why the Decision Happened in August
Look back at this month's departures and a lot of them were predictable. Seasonal contracts ended. Students went back to class. Patios closed and hours got trimmed. The people leaving knew for weeks. The question is whether anyone asked them.
The rest of the wave was quieter. People rarely quit on impulse. They drift first. In a January 2026 global survey of 460 chefs by Bournemouth University and The Burnt Chef Project, 69% said they often or always think about leaving their job, and two-thirds were actively looking for other work. Those aren't people who've resigned. They're people still showing up, still on the schedule, already halfway out.
That's the uncomfortable part. A kitchen that looked fully staffed on paper in August may have been filled with employees who were never planning on staying.
The schedule tells you who's working. It doesn't tell you who's leaving.
Peak season makes the drift easy to miss. It's when managers are stretched thinnest, and it's also when the decisions get made. By the time things slow down, the notices are already written.
The Next Wave Is Already Forming
September isn't the only wave. January is the second: 5,945 documented departures in 2023, 6,349 in 2024, 7,126 in 2025, and 7,932 in 2026. Up every year.
The pattern is the same, just earlier in the year. The people who leave in January will make up their minds over the holidays, while you're running some of your busiest weeks of the year. If September caught you off guard, November is your chance to be ready.
The Signals Worth Watching
None of these guarantees someone is leaving. But together, they're the earliest warning you'll get.
- Availability quietly narrowing. Someone who used to take anything now blocks off weekday mornings or suddenly can't do Sundays. That's often a new schedule forming somewhere else: classes, or another job.
- More shift drops and swap requests. Picking up shifts is engagement. Dropping them is distance. A steady rise from someone who used to be reliable is worth a conversation.
- Seasonal staff you haven't checked in with. If you don't know who's staying past the holidays, you may already be short-staffed for January. You just don't know it yet.
- The crew that carried the rush and heard nothing. The people who worked the doubles and covered the call-outs are the ones most likely to feel taken for granted once things slow down.
- Kitchen staff nearing the one-year mark. Line cooks have the shortest median tenure of any role in Push data, at 11.2 months. If someone started last winter, they're in the window now.
- Pulse scores slipping, or silence. A drop in engagement scores is a signal. So is a survey nobody bothers to answer. Low response is its own kind of answer.
The pattern matters more than any single signal. One dropped shift is life happening. Three signals from the same person in the same month is a decision in progress.
What to Do Before the January Wave
1. Run a Pulse Survey in November
Not in January. By then, you're running an exit interview with extra steps. A short check-in during the busy stretch tells you where people stand while you can still change something: a schedule, a role, a pay conversation you've been putting off. Next summer, run the same check-in during late July.
2. Ask Seasonal Staff Directly
"Are you staying after the holidays?" is a two-minute conversation. Build the January schedule around the real answer, not the hopeful one.
3. Recognize the People Carrying the Load
It's the cheapest lever you have. In a Toast survey of US restaurant workers, 44% of those planning to leave the industry cited a lack of recognition for their hard work, second only to low pay at 47%. You can't always fix pay mid-season. You can always say thank you, specifically and publicly, to the people who earned it. If the crew that stuck with you through September hasn't heard it yet, start there.
4. Put Retention Effort Where Churn Is Worst
Back-of-house median tenure runs 11 to 17 months in Push data. Front of house runs 18 to 26. Line cooks post a 33.2% documented separation rate, the highest of any role. If your retention effort is spread evenly, it's spread too thin. Start in the kitchen.
5. Start January Hiring in November
The Restaurant Workforce Benchmark Report's rule is simple: start hiring two months before each wave. This September's gap was set in motion by hiring decisions made in July and August. January's will be set by what you do in November, not by what you scramble to do after New Year's.
6. Make It a System, Not a Memory
For multi-location operators, the same calendar plays out in every unit at once, and every GM is buried at the same time. If the check-in lives in one manager's head, it happens at your best location and nowhere else.
That's where Push comes in. Push Engagement Surveys let you send short check-ins to staff at every location. You see which locations are drifting before the notices land, not after.
The Most Expensive Survey You'll Ever Run
An exit interview tells you exactly why someone left. It just tells you too late to do anything about it.
The operators who keep their people through each wave aren't luckier. They ask earlier. They notice the narrowing availability, thank the crew that carried the rush, and plan around real answers instead of hopeful ones.
Download the 2026 Restaurant Workforce Benchmark Report for the full churn calendar, role-by-role tenure data, and where your retention dollar goes furthest.
If you'd like to see how Push helps multi-unit operators spot churn risk early across every location, book a demo and we'll walk through it with your numbers.
