What the numbers say at the end of the day

Most venues look at one number at the end of the day: takings. It says whether the day was good and nothing about why.

Here are the ones that earn their place — and the ones that do not.

The four that are enough

Takings against the same day last week. Not against yesterday. Tuesday against Tuesday is a comparison; Tuesday against Saturday is noise.

Number of paid orders. With the takings this gives the average per order. If takings fall while orders hold, people are ordering less per table — a different problem from fewer people arriving.

Which dishes sold, by count and by takings. The two rankings never match, and the gap is the interesting part. A dish that is first by count and ninth by takings fills your kitchen and not your till.

The hours. When the takings come in. This shows whether lunch works, whether there is a dead hour, and whether staff are spread along the curve rather than by habit.

What a phone menu shows that a till cannot

This is the difference a digital menu brings and paper has no way to give.

What gets browsed. Which dishes are opened and read. If a dish is opened a hundred times and ordered five, it is attractive and something about it puts people off. Usually the price, sometimes the description, sometimes a missing photograph.

What is searched for and not found. A guest who types “vegan” into the menu search and gets nothing is telling you what to add. It is the only source of information about a dish that is not there.

Which language they read in. This shows which guests actually come through — and whether a fourth language is worth it.

Numbers that look important and are not

Total scans of the QR code. It shows traffic, not behaviour. It grows with the number of tables and with time, not with how well you are doing.

Average time in the menu. Long can mean reading carefully and can mean not finding what they want. On its own it means nothing.

A single record figure. The best day of the month is usually explained by something outside — a match, a holiday, rain. A two-week trend says more than a record.

How to read them

Five minutes at the end of the day, not half an hour once a month.

A monthly review shows what happened. A daily one shows what is happening while something can still be done. If a change to the menu does not show in the numbers within a week, it did no work — and you only learn that by looking often.

In short

Takings against the same day last week, order count, the two dish rankings, and the spread across the hours. Everything else is for when those four have stopped surprising you.