Staff leave every few months. Build a restaurant that does not depend on who is on shift

Plan for the person who will leave

Turnover in restaurants is not a failure of hiring. It is the nature of the work: young staff, seasonal peaks, a better offer two streets away. Planning as if the team will stay is planning for a restaurant that does not exist.

The useful question is not how to keep everyone. It is how quickly somebody new can be useful, and how little is lost when somebody experienced goes.

A system a new person can use on their first shift

If taking an order means memorising item codes, a new waiter is a liability for a month. If it means tapping the dish on a phone, with its variants and notes laid out, they are taking correct orders on day one.

The same goes for the kitchen. Orders that arrive on a screen, grouped by station, need no explaining. Orders shouted across the pass need a person who has learned how this particular kitchen shouts.

Give each role only the buttons it needs

Shared logins and a single all-powerful password are how discounts and cancellations stop being anybody's responsibility. Separate logins cost nothing and change behaviour immediately.

• Waiters take orders and cannot apply discounts
• Cashiers bill and take payment, and cannot delete a bill
• Only a manager can cancel an item after it was sent, and the cancellation carries their name
• When somebody leaves, their login is switched off that day — not whenever someone remembers

Write the house rules down once

What an experienced person knows and a new person does not is mostly small and specific: how much gravy goes on this plate, which table is always a birthday, what to do when the printer jams. Written once, it stops walking out of the door.

• Opening and closing checklists, a page each
• Portion sizes for the fifteen best-selling dishes
• The five complaints you hear most, and what staff are allowed to offer for each