Every restaurant has a shape. Not a vague feeling about which nights are busy, but an actual distribution of money across the hours and days of the week, and it is usually far more lopsided than the people working there believe. The kitchen remembers the nights it got hammered; the numbers remember the afternoons nobody was there. This topic is about getting the real shape out of your POS, because almost every operational decision you make — hours, staffing, ordering, when to promote — is a bet on it.
🕐 Where the money actually is
Pull sales by hour for a month and one thing becomes obvious immediately: activity and money are not the same curve. A lunch service can produce a great many covers and very little contribution; a quiet Thursday dinner can quietly out-earn a frantic Saturday lunch, because the checks are bigger and the labour needed is lower.
Look at each daypart in three columns rather than one: sales, covers, and average check. That is the same split the previous course used, and it does the same work here — the three together tell you what kind of service each period actually is. A period with many covers and a small average check is a volume service, which lives or dies on speed and on labour cost. A period with few covers and a large average check is an occasion service, where the room, the wine and the service matter far more than throughput. They need different menus, different staffing and different marketing, and running them identically is how restaurants end up mediocre at both.
Two hours deserve special attention because they are where money hides. The shoulders — the last hour before a rush and the first hour after it — often carry a surprising share of sales while being the first thing cut when hours are tight. And the opening and closing hours, which frequently produce almost nothing while still costing a full crew, are worth measuring precisely before you defend them out of habit.
📆 The week has a spine
Do the same by day and the week gets a spine: two or three days that carry the business, a middle that pays its way, and usually one day that does not. Knowing which is which in money rather than in feeling changes what you do with each of them.
| Kind of day | What it needs |
| Your two or three big days | Protection: full staffing, no experiments, nothing out of stock |
| The solid middle | Efficiency: this is where a small improvement repeats fifty times a year |
| The day that does not pay | A decision, not a habit — see below |
| The day you have never tested | Curiosity: a brunch, a set menu, an early service |
The big days are worth defending harder than most owners do. A weekend service that runs out of the popular dish, or is understaffed by one person, costs more than any saving elsewhere in the week — those are the days where every seat is spoken for and a mistake cannot be recovered on the next table.
🚪 The day that does not pay
Almost every restaurant has one, and almost nobody has run the numbers on it. The previous course gives you exactly the tool: you know your daily break-even figure, so compare it against what that day actually does, consistently, over a couple of months. If it sits below the line, that day is being subsidised by the rest of the week, and it is worth deciding that on purpose rather than by inertia.
Deciding does not automatically mean closing. There are usually four options, and they are worth working through in order.
- Make it cheaper to open. A smaller crew, a shorter menu, one section of the room. Often a day that loses money at full staffing breaks even at two thirds of it.
- Give it a reason to exist. A different offer that suits that day rather than a discount on the usual one — the marketing courses can fill a specific day if it has something specific to say.
- Move the hours. Sometimes the day is fine and only the service is wrong: closing the lunch and keeping the dinner, or the reverse.
- Close it. The option owners resist most, and sometimes the right one — a closed day costs you its contribution but saves its labour, and gives your team a genuine two days together, which the training course will tell you is worth real money in retention.
Whatever you choose, check one thing before you act: whether that day's customers are the same people who come on the good days. If your Monday regulars are your Saturday regulars, closing Monday does not just lose Monday's sales, and this is exactly the question the customer data course is built to answer. The sales report alone will tell you the day is weak; only the customer data will tell you whether it is load-bearing.
🌦️ Seasons, and the calendar's tricks
Zoom out to the year and a second shape appears, and it matters most for cash rather than for pride. You already know your quiet months roughly; measuring them precisely tells you how much cushion the good months need to build, which is the forward planning the previous course's cash sheet exists for.
Be careful about what you are actually comparing, though, because the calendar quietly moves your numbers. A month with five weekends beats one with four by more than most menu changes. Holidays land on different weekdays in different years. Weather moves a terrace restaurant more than any promotion will. And a single event — a concert, a match, a street closure — can make a week look like a trend when it was a one-off.
The habit that protects you is comparing like with like: the same month last year rather than last month, and, for the week, the same weekday rather than the day before. And write down what happened next to the numbers, as the previous course suggested — "roadworks", "long weekend", "the chef left" — because in eleven months you will be looking at this data trying to forecast, and you will have no memory of why that week was strange.