Whether that promotion worked

Every restaurant has run a promotion that "worked": the room was full, the kitchen was slammed, the sales figure at the end of the night was the best in weeks. And a surprising number of those nights made less money than an ordinary Tuesday. Selling more and earning more are different things, and the gap between them is made of arithmetic nobody does at the time. This topic is that arithmetic, and the harder question underneath it — whether those people were coming anyway.

➖ What a discount actually costs

A discount does not come out of your price; it comes out of your contribution, and that is a much smaller number. Take the pasta from topic 3: it sells at eighteen pesos, costs five, and leaves thirteen. Take twenty percent off and it sells at $14.40 — but the food still costs five, so it now leaves $9.40.

The price fell by a fifth. The money you keep fell by more than a quarter. So to make the same total contribution as before, you do not need twenty percent more customers — you need thirteen divided by $9.40, which is almost forty percent more plates, produced by the same kitchen, in the same hours, with the same staff.

That is the calculation to do before agreeing to any discount, and it gets worse as the discount deepens, because the contribution shrinks faster than the price does. It is also why "two for one" is a much more expensive offer than it sounds: you are giving away an entire plate's contribution and paying its food cost. None of this means never discount — it means know the number you have to hit, and say it out loud before the promotion rather than after.

🤔 Would they have come anyway?

The second question is harder than the arithmetic and it is where most promotion analysis quietly fails. If you offer a discount to everyone who walks in on a Friday, most of them were walking in on Friday regardless — so you did not buy extra sales, you bought a rebate on sales you already had. The promotion looks successful because the night was busy, and the busy night would have happened without it.

You cannot measure this perfectly, but a few habits get you close enough to decide.

• Compare against a fair baseline. Not against a bad week: against the same weekday for the four weeks before, or the same period last year.
• Promote where there is room. An offer on your dead Tuesday is nearly all incremental. The same offer on a full Saturday is nearly all rebate.
• Make it need an action. A code, a specific dish, a booking — something that lets you count who used it, instead of applying it to everyone automatically.
• Watch the whole check, not the offer. A cheap starter that brings in a table which then buys drinks and dessert can be excellent business; the offer item's own margin says nothing on its own.
• Look at the week after. A promotion that pulls next week's customers into this week has moved sales, not created them, and the following week's dip is part of the result.
• Ask whether they came back. A discount that acquires a customer who returns at full price is an investment. One that acquires a discount-hunter is a cost — and which of the two happened is a customer data question.

That last point is the difference between a promotion and a marketing strategy. The loyalty and email courses are about giving people a reason to return; this course only tells you whether a particular offer left money behind. Both matter, and confusing them is how restaurants end up running permanent discounts and calling it marketing.

📋 Judging one properly, in half an hour

Judging a promotion is not sophisticated work, and doing it once badly is worse than not doing it, because a wrong verdict gets repeated. What follows is enough for any restaurant.

Before it runs, write down three things: what you expect it to do, the fair baseline you will compare against, and the number of extra units or covers that makes it worth doing — which you now know how to calculate. Writing them down beforehand is the whole trick, because a promotion evaluated afterwards is always judged against whatever number flatters it.

Afterwards, compare contribution, not sales. Take the units sold at the promotional price, work out what they left, add anything else those customers bought, and compare the total against the baseline period's contribution. Then subtract what the promotion itself cost you: printing, the platform's fee for featuring it, the extra staff you rostered for a night you expected to be busy. The number that comes out is the answer, and it is often smaller than the sales figure suggests and occasionally negative.

Then write the verdict down next to the plan, in a sheet you keep. After a year of doing this you will have something genuinely valuable and rare: a record of which of your offers actually pay in your restaurant, on which days, and which ones only ever made you busy.

💡 Things that lift sales without a discount

Because discounting is expensive, it is worth knowing that most of the reliable ways to raise a check do not involve giving anything away. They are less exciting than a promotion and they compound quietly.

The biggest is what the floor recommends. A server who suggests the star and the puzzle from topic 3 moves the mix towards contribution without changing a single price — and unlike a discount, it costs nothing and works every service. Then there are the small structural moves: a drink or a side offered naturally at the right moment, a menu where the profitable dishes are where the eye falls first, and a dessert list that arrives while people are still comfortable rather than after the check.

Two more, in the same spirit. Raising a price by a small amount on a workhorse is, arithmetically, the mirror image of the discount calculation — a one-peso rise on five hundred plates is five hundred pesos, and unlike a discount, you keep all of it. And filling capacity you already have, at full price, through the marketing channels from the earlier courses beats filling it at a discount every single time. Discounts are a tool for a specific job: a genuinely empty period, a new dish nobody has tried, an occasion. Used as a habit, they train your customers to wait for the next one — which is the most expensive outcome of all, and the one no single promotion's numbers will ever show you.