The hours budget

Most restaurants find out what labor cost when payroll runs, which is between four and fourteen days after they could have done anything about it. The fix is not more discipline while scheduling. It is deciding the number of hours before the schedule exists.

🔮 Forecast the week first

You cannot budget hours without a sales estimate, and you already have everything you need to make a decent one:

1. Take the same week last year, if you have it, or the last four weeks if you do not.
2. Adjust for what you know. A holiday, a local event, a road closure, a menu change, the school calendar.
3. Adjust for the trend. If the last eight weeks are running above last year, carry that through instead of pretending it is not happening.
4. Break it down by day, and for the busy days by day part. A weekly total cannot be scheduled against.

Your forecast will be wrong. That is fine and it is not the point — a forecast you compare against reality every week gets better fast, and even a mediocre one beats scheduling from memory. Write it down before the week starts, because a forecast you adjust afterwards teaches you nothing.

🧮 From forecast to hours

Two ways to do it, and the second is better once you have the habit:

Whichever you use, split the result before you schedule anything:

• Fixed hours. Opening, closing, prep, receiving, cleaning, management. These barely move with sales, and cutting them is how you end up opening late with no mise en place.
• Variable hours. The extra server, the second person on the line, the busser on a Saturday. This is your real lever, and it is the only place a mid-week correction can come from.

Then hand the schedule a number, not a feeling: "we have this many variable hours this week". The person building the schedule can now make trade-offs instead of guessing whether they are being generous.

⚖️ When the budget and the schedule disagree

They will, most weeks. The order to resolve it in:

1. Check the forecast first. If Saturday genuinely looks bigger, the budget was wrong, not the schedule. Change it deliberately and note why.
2. Trim the edges before the middle. Half an hour off the start or end of several shifts is invisible to guests. Removing a whole person from the peak is not.
3. Stagger harder. The cheapest hour you will ever save is one where somebody was standing around waiting for the rush.
4. Move hours between days, not out of the week. Often the problem is a Tuesday, not a total.
5. Only then, cut. And say why to the person affected, before they read it in the schedule.

One rule worth holding: if you are over budget in a way that would break service, publish the schedule anyway and log it. A ruined Saturday costs more than the hours you saved, and the log is what turns a bad week into a better forecast.

📅 Mid-week, while it can still change

The budget is only useful if somebody looks at it before the week ends:

• Wednesday check. Actual sales and actual hours so far, against forecast. Ten minutes.
• If sales are behind, cut variable hours from the back half of the week: a late start here, an early cut there, a shift that no longer needs two people.
• If sales are ahead, spend some of it. Being understaffed on a busy Saturday to protect a percentage is a false economy that costs you tables and goodwill.
• Cut with notice and fairly. Rotate who gets sent home early, ask first, and remember that people came to work expecting to earn a shift's pay. This is where a labor decision becomes a retention problem — topic 5.