The five numbers you watch

There is a version of this subject that buries you: forty indicators on a screen, updated hourly, none of which changes what you do on Monday. A restaurant does not need that and cannot sustain it. It needs about five numbers, looked at every month, understood well enough that a movement in one of them makes you ask a specific question. This topic is those five, why each earns its place, and what to do when one of them moves.

🖐️ The five

Here they are, with the same month from topic 2 as the example. Everything here comes from the P&L and the POS, so none of it requires new systems or new work beyond writing it down.

The middle two are there for a reason that is easy to miss: sales on their own are a dead end diagnostically, because sales are just covers multiplied by average check. Two thousand covers at twenty-five pesos is fifty thousand in sales — and when sales fall, the only useful first question is which of the two moved. Fewer people is a marketing and reputation problem, and this library has five courses about it. The same people spending less is a menu, pricing and service problem, and it is an entirely different fix. A sales figure alone cannot tell you which conversation to have.

🔁 Read them as a set, not a list

The five are worth much more together than separately, because a movement in one is only interpretable next to the others. That is what turns a monthly review from staring at figures into a diagnosis.

• Sales up, prime cost % up, result % down. You are growing on volume that does not pay. Look at what is selling — the sales analytics course — before you celebrate the top line.
• Covers down, average check up. Often a price rise doing its job, but check it is not the quiet loss of your everyday regulars, replaced by fewer, bigger occasions.
• Covers up, average check down. Usually discounting or promotions. Fine if it was deliberate and priced; expensive if it just happened.
• Everything flat but result % down. Look at the fixed lines: a rent increase, a new subscription, a repair, insurance.
• Prime cost fine, result % thin anyway. Your fixed costs are too heavy for your sales volume, which is a much bigger conversation and usually a rent or opening-hours one.

Notice how each pattern points at a different course in this library. That is deliberate: the numbers are the diagnosis and the other books are the treatment. The value of the five is not that they solve anything — it is that they tell you which problem you actually have, which is where most owners lose months.

📏 Compare against yourself, and against the right month

A number needs something to be compared with or it says nothing, and the choice of comparison is where most restaurant reporting goes wrong. The instinct is to compare against last month, but restaurants are seasonal and the calendar is uneven, so last month is often the worst available benchmark.

Three comparisons that do work, in order of usefulness. The same month last year, which handles seasonality and is the one to trust most. The last three months as a rolling trend, which smooths out the noise of a single strange month. And, only for costs, published industry ranges — used gently, because a bar, a bakery and a full-service dining room have genuinely different structures and an average that mixes them describes nobody.

Watch out for the calendar too, since it quietly moves your numbers without anything changing in the restaurant. A month with five Saturdays beats one with four, and a public holiday landing midweek can move a month by more than a menu change. If you are comparing months, it is worth knowing how many trading weekends each of them had — and it is another argument for the covers-and-average-check pair, because they show you immediately whether a month was really different or simply longer.

📝 Making it a ten-minute habit

The five numbers are worthless if they are calculated once in a burst of enthusiasm and never again, and that is what usually happens. What makes the difference is that they get written in the same place every month, by hand if necessary, in a row under the last one.

A single sheet with a row per month and five columns beats any dashboard, because it is the comparison that carries the information and a sheet like that is nothing but comparison. After a year you have something genuinely valuable: the shape of your own restaurant across the seasons, which no benchmark can give you and which makes next year's version of this month readable at a glance. Add a sixth column for notes — "raised prices", "chef left", "roadworks" — and you will be able to explain your own history a year later, which you otherwise will not.

Keep the discipline of the set as well. It is tempting, in a bad month, to write down only the numbers that came out well; that is the same failure the customer data course called using data as decoration. The five are chosen precisely so that the uncomfortable one has nowhere to hide, and the month you least want to write them down is the month they are worth the most.