Eight numbers a shop should know, and the three that lie

A shop needs eight numbers, because each ends in a decision: buy, reprice, chase, stop or wait. Everything else on a report is context for one of them. The three figures owners reach for first — revenue on its own, margin on the selling price, and the bank balance read as profit — mislead, and each misleads towards good news.

Which eight numbers, and what does each one decide?

A number earns its place by the decision on its other side; if nothing changes when it moves, it is a curiosity.

Number	The decision it drives
Revenue for the period	whether the month is on plan, and which category gets attention
Gross margin	whether the price list still works after a supplier increase
Stock turns	whether the shelf earns the space it takes
Days of cover	when to place the next order, and how much
Average basket	whether to push a second item at the till
Footfall, if counted honestly	whether the problem is traffic or conversion
Returns	whether a product, size or supplier has to go
Cash position	whether the next buy can be paid without borrowing

Two need defining before they can be read. Footfall counts only where people have one way past a counter that counts them; a stall, a shared door, or a sensor the staff walk past measures the counter. Days of cover is stock divided by the rate it leaves the shelf over the past few weeks, not the year, or a season hides inside the average.

Why does revenue without the month's purchases mislead?

Because revenue is a claim on cash, not cash, and a month can look strong while the shelf empties. A shop that sells out of stock bought three months ago records revenue and no purchase: the page says trading is good, the stockroom says the buffer is gone, and the next order is larger than the month's margin can pay for. Read beside purchases and stock movement, the figure says whether you sold what you bought or sold what you had — and whether the rise came from a price increase, a bigger basket, or a delivery charge.

Why is margin on the selling price wrong?

Because it is a different number from the one the shop manages to, and the two get mixed within a month. Margin on the selling price says what share of the till is yours; mark-up on cost says what the sale earned on the money in the goods. Switching between them makes a stable business look volatile and a deteriorating one look acceptable.

Two things make it worse. The invoice price is the easy part of the cost: freight, customs, the courier fee on a returned parcel, the free sample and the damaged carton are what it took to get the item onto the shelf. And a discount looks smaller against the selling price than against the profit it removes; a shop that discounts on that figure finds at month end that the promotion moved stock and produced no money.

Why is the bank balance not profit?

Because a statement records when money moved, not what was earned. Three ordinary events raise the balance in a month that earned nothing: a supplier paid late, a loan, a customer's deposit for undelivered goods. Three lower it in a month that earned well: a large stock purchase, a quarter's rent falling due, last month's invoices settling. A shop can be profitable and unable to pay, or comfortable for months on a business that is losing money. The balance answers only whether you can pay what falls due.

Why does a number need an owner and a rhythm?

Ownership is not a job title; it is the name of the person who would be asked why the number moved, and who wants it right. Beside that name goes the definition in one line: what is counted, from which system, excluding what. Two people computing revenue differently is an argument that never reaches the business.

Each of the eight has a natural interval, and producing it faster or slower wastes it: a till count a week old cannot be acted on, and a margin refreshed every morning invites a decision on noise. Daily for cash, weekly for basket, footfall, returns and cover on the fast movers, monthly for margin, stock value and the purchases behind them.

A figure without its definition and its date is not evidence, which is why the provenance and versioning the W3C asks for alongside a dataset separates comparing this month with last from comparing two measures that share a name.

What does the month-end ritual look like?

The ritual has a fixed order, because each step corrects the next.

1. Count the stock: fast movers and disputed lines first, in the base unit. Record every difference with a reason, and never type the right number over the wrong one.
2. Reconcile the bank. Match each line to an invoice, a supplier payment or a transfer, and put what does not match into a queue with a name against it.
3. Post the month's purchases and credit notes, including those dated inside the month and entered after it closed.
4. Compute the eight from their source systems, not from a sheet assembled the night before.
5. Write one paragraph: what moved, why, and what changes next month.
6. Sign and date it, and keep last month's version beside it.

The order is not ceremonial. A count taken after the report is written becomes a correction to a published number, and people stop trusting the report rather than the count.

Where this goes wrong

• The impressive number. Total revenue on a wall chart, only ever going up, before a team that cannot change it. A mood, not a measure.
• A definition changed without a date. Margin is improved in June and history is not restated, so every comparison across the year is wrong invisibly.
• Footfall from memory. Counting customers by eye on a busy Saturday measures the person counting; the conversion rate built on it is fiction.
• The report written by the person judged by it. Where one person prepares and is blamed, rounding drifts towards the audience.

What this does not cover

This is not accounting advice and names no benchmark. What a good margin is depends on the category, the city, the way you buy and what the customer will pay; a produce shop and a hardware shop cannot share a target, and a figure borrowed from another business is worse than none. Which costing method you use — how freight and shrinkage enter the cost, when a write-down is recorded — is your accountant's decision. Multi-branch shops also need a rule on which branch owns a transfer.

We build the eight figures with a source, an owner and a date on each, so the month ends with a decision instead of a discussion.