Most takeaway owners check one number: what came through the till. It’s the obvious one, and it’s not useless, but on its own it’s the least informative figure you have. A good week and a bad week can post similar totals for completely different reasons, and revenue alone won’t tell you which you had. It tells you the score, not the game.

The point of looking at data isn’t to admire it. It’s to notice something changing while you can still do something about it — a dish quietly losing money, regulars drifting away, refunds creeping up. That doesn’t take a dashboard habit or a spreadsheet you’ll abandon by August. It takes five numbers and fifteen minutes on a Monday.

1. Average order value

What it is: total revenue divided by number of orders — what a typical customer spends per order.

AOV is the number you have the most direct control over, which is what makes it worth watching. You can move it this week with changes you already know how to make: modifier placement, a sensible minimum order, a bundle that nudges people toward a drink and a side. When AOV drifts down, it’s usually not that customers got poorer — it’s that the menu stopped asking for the add-on. When it drifts up, whatever you changed is working, and it’s worth knowing what.

2. Repeat rate

What it is: the share of customers who order more than once — the proportion coming back versus ordering once and vanishing.

This is the number that separates a business that’s growing from one that’s just churning through customers. A takeaway that fills the top of the funnel but leaks out the bottom can look busy and still be going nowhere. Repeat rate is where owning your customers pays off: on an aggregator you can’t see it and can’t act on it, because the customer was never yours. On your own storefront, a falling repeat rate is an early warning, and a rising one is the single healthiest signal a food business can have.

3. Gross profit percentage

What it is: the share of a sale left after the cost of the ingredients — roughly, revenue minus food cost, as a percentage.

Revenue flatters; GP tells the truth. Your bestselling dish and your most profitable dish are often not the same item, and until you look at GP you’re flying blind on which is which. A busy week selling low-margin food can bank less than a quieter week with the mix right. Most kitchens aim somewhere in the region of 68–72% GP on food, and knowing where each dish sits tells you what to feature, what to reprice, and what’s quietly costing you money every time it sells well.

4. Refund rate

What it is: the share of orders that end in a refund, and — more useful — why.

Refunds are the cheapest market research you’ll ever get, because every one is a customer telling you exactly what went wrong. A rising refund rate is never random. It’s a dish that doesn’t travel, a prep time you keep missing, an out-of-stock item still on sale, a kitchen slammed past its capacity. The number matters, but the reasons matter more. Read the why behind your refunds and you’ll usually find one or two fixable causes behind most of them.

5. Prep time accuracy

What it is: how long orders actually take versus how long you told the customer they’d take.

The prep time you quote is a promise, and it’s the one you’re most likely to break on exactly the nights you can least afford to. When your real times run consistently over the quoted ones, everything downstream suffers — cold food, chased-up orders, refunds, a customer who won’t come back. This is an operational metric hiding as a customer-experience one. If the gap is widening, it’s usually a sign the kitchen’s capacity limits and busy-mode settings need to reflect reality rather than optimism.

The numbers to ignore

Just as useful is knowing what not to look at. Some numbers feel like progress and tell you nothing actionable.

• Total page views or site visits on their own. Traffic without conversion is just weather. What matters is how many of those visits became orders.
• Social media followers. A big following that never orders is a vanity metric with a food-shaped logo. Orders attributed to a channel are the real number.
• Lifetime totals. “We’ve done 40,000 orders” feels great and changes nothing about this week. Trends beat totals every time.

The test for any metric is simple: if the number moved, would you do something different? If not, it’s decoration.

Make it a fifteen-minute habit

The value isn’t in any single reading — it’s in the trend. One Monday’s numbers are a snapshot; eight Mondays in a row is a story, and the story is what tells you whether last month’s change worked.

• Set a baseline. Write down where the five numbers sit today. You can’t spot a change without a starting point.
• Same time, same day. A fixed fifteen minutes on a Monday beats an occasional deep-dive you’ll skip when it’s busy. Consistency is what turns numbers into instinct.
• Watch the direction, not the decimal. You’re not auditing. You’re asking one question of each number: better or worse than last week, and do I know why?

None of this is about becoming a data business. It’s about the difference between running a takeaway on gut feel and running it on gut feel informed by five numbers. The gut was never the problem — it just works a lot better when it knows which way things are moving.