Which report answers which question (and which ones to ignore)

Most restaurant software ships with more reports than anyone opens. The problem is not the number — it is that nobody attached a decision to each one.

So here is the map, organised by the question rather than by the menu it sits under.

Operational reports: what happened, and when

The hourly one earns its place. “Tuesday is slow” is a feeling; “Tuesday is fine until 8, then it dies” is a schedule change. You can send someone home at eight. You cannot send them home from a monthly total.

Control reports: what should not have happened

Discounts, voids and refunds, by branch and by user. It is worth being clear about why these exist, because staff assume it is suspicion and it usually is not.

An outlier here is nearly always one of three things: a process nobody explained, a menu item that keeps getting rung wrong, or a till layout where two buttons sit too close together. Theft exists, but it is the least common explanation and the last one to reach for.

Read them weekly, look for patterns rather than incidents, and start the conversation with “what happens when…” rather than “why did you…”.

Cost reports: what it took to earn it

• Theoretical versus actual food cost. The single most valuable number in the system, and the one that requires recipes to exist. Anything over about three points of gap is money leaving the building.
• Waste by reason, item and branch. Not the total — the ranking. One item usually dominates.
• Prime cost by branch. Food plus labour against that branch’s own sales. The fastest health check there is.

Financial reports: what it means

These come out of the ledger rather than being assembled: a trial balance that proves the books are internally consistent, a profit and loss for a period and a branch, a balance sheet at a point in time, and tax reports built from the same entries the receipts produced.

The test of whether these are real is simple: can you click from a number in the P&L back to the order that caused it? If yes, your accountant can work from the system. If no, they will rebuild it, and you will pay for that twice.

People reports

Attendance and lateness, overtime by type, leave balances, and labour cost as a percentage of sales. The last one belongs beside the sales-by-hour chart, because that pairing is the whole of rostering: how many people, at which hours, against how much revenue.

What to actually look at

Weekly, four things:

1. Sales by hour, this week against last.
2. Prime cost, per branch.
3. Theoretical versus actual food cost, per branch.
4. Discounts, voids and refunds, scanning for outliers.

Monthly, one thing: the close. P&L against last month, the balance sheet, and a look at anything that moved more than you expected.

That is five reports. Everything else in the system is there for the week you have a specific question — which supplier is late, which dish stopped selling, which cashier needs help — and that is exactly what a long report list should be: an archive you visit deliberately, not a dashboard that competes for attention every morning.

The habit that makes reports work

Pick a fixed time each week, look at the same four, and write one sentence about what you will change. A number you did not act on is entertainment. The whole point of measuring is the sentence at the end.