A point of sale produces more reports than anyone will read, which is functionally the same as producing none. Most owners look at one number — today's total — and nothing else.
So this is the short list. Six reports, what each one is for, and roughly how often to look. If you only ever read these, you will know more about your business than most people running one.
Daily, and it takes two minutes
The close-out variance. What the drawer holds against what the system expected, plus whether the card batch balanced.
That is it. One number, every day, and the reason to look daily rather than weekly is that a variance you find the next morning is traceable and a variance you find in a month is not.
A consistent small shortfall is usually process — change given wrong, a tender rung in as the wrong type. A sporadic large one is worth taking seriously. Either way you cannot tell the difference without a run of daily numbers.
Weekly, fifteen minutes
Sales by item. Sorted by contribution in dollars, not by units and not by percentage.
This is the report that tells you what your business actually sells, which is reliably different from what you think it sells. It is also the input to any pricing or menu decision. For restaurants the full method is menu engineering; for salons the same logic applies to services and it is usually more lopsided than owners expect.
Look at dollars because a high-margin item nobody orders is not a good item, and a low-margin item that sells constantly may be paying your rent.
Labour against sales, by day part. Two columns, same time buckets.
Not a total. The total hides the problem, which is almost always specific hours — a Tuesday afternoon with three people on for four covers, a Saturday morning understaffed. Fix the hours, not the headcount.
Weekly if you have staff on a variable rate
Voids and comps, by reason and by person.
Both are money leaving without a sale and both are legitimate in normal amounts. What you are looking for is pattern: one person with three times everyone else's rate, one item that gets sent back constantly, one reason code that is being used as a catch-all.
If your system does not require a reason, that is worth changing before reading this report, because a comp with no reason attached tells you nothing.
Monthly, half an hour
Repeat and rebooking rates. What proportion of customers came back, and within what window.
For a salon this is the closest thing to a single health metric. A salon with a good rebooking rate can survive a bad marketing month; one without it needs new clients forever. For a restaurant the equivalent is harder to measure and still worth the attempt if you have any customer identification at all.
Sales by hour, across the month. Which tells you your real shape.
Almost every business has a shape different from the one it staffs and opens for. A quiet first hour, a peak that is narrower than assumed, a day that is carrying less than it appears. This report is the input to opening hours, staffing, and whether that early shift earns its cost.
The one people ask for and should ignore
Average transaction value, on its own.
It moves for reasons that have nothing to do with anything you did — one large party, one retail sale, a shift in mix. Watched monthly it is noise, and it invites decisions based on noise. If you want it, look at it as a trend over a quarter, next to volume, or not at all.
Three that are worth building once
Not standard reports, but derivable and unusually useful.
Requests you could not fill. By day and hour. If your phone or booking page records attempts, the turned-down ones are the highest-value data in the business, because they are demand you already had.
Time from booking to appointment. If it is lengthening, you are turning people away without noticing. If it is shortening, demand is softening before the revenue shows it.
Retail attachment rate, for salons. What share of service tickets included a product. It is usually much lower than the owner believes and it responds quickly to attention.
How to actually do this
The failure is never the reports. It is that reading them is nobody's scheduled job.
Pick a fixed fifteen minutes each week, the same slot, and put it in the calendar. Daily variance goes to whoever closes. The monthly half hour goes wherever you do your other monthly admin.
And write down what you changed as a result. A report you read and never act on is entertainment. The test of whether this is working is that something in the business is different because of a number you looked at.