What the delivery apps actually cost you, per order

Commission on the major platforms typically runs between 15% and 30% of order value, depending on whether they deliver or you do, and what you negotiated.

On a £25 order at 25%, that is £6.25. Against food and labour costs, the margin on many delivery orders is close to nothing before you account for packaging.

Plenty of operators know this and accept it as marketing spend. That is a defensible position. It just needs to be a decision rather than a default.

You do not get the customer. No name, no email, no phone number you can use. Someone who ordered from you eleven times is a stranger you cannot contact. You cannot tell them you have opened on Sundays. You cannot win them back when they drift.

You do not get the data. Which dishes get abandoned in the basket, what time your real peak is, which postcodes are worth expanding into. You get the platform's dashboard, which shows you what the platform wants you to see.

Your pricing gets pulled around. Many agreements include price parity — you may not undercut the app on your own channel. So the mechanism that would move customers to a cheaper-for-you route is contractually closed.

The relationship belongs to the app. When someone thinks "let's order Thai," they open the app, not your site. The app owns the moment of intent, and rents it back to you per order.

You are not paying 25% for delivery. You are paying 25% for delivery, and giving away the customer, the data, and your pricing freedom for nothing.

Do not guess. Work these four numbers out.

1. Monthly orders through apps, and average order value.
2. Total commission per month. Orders × value × rate.
3. What direct ordering would cost. Card processing is roughly 1.5–2.5%. A hosted ordering system is typically a modest monthly fee, or a one-off build. Add delivery if you are not doing it yourself — either your own driver or a courier on a per-drop rate, which is usually far below platform commission.
4. What proportion would actually switch. Be pessimistic. New customers who found you on the app will keep using the app. Regulars will switch if you ask them and make it easy.

If 30% of your app volume moved direct, what is the annual difference? For a lot of kitchens that number is five figures, and it pays for a proper ordering system several times over in the first year.

If it is not, stay on the apps and stop worrying about it. That is a real outcome and worth knowing.

Almost nobody should leave the platforms outright. The winning arrangement is usually both:

Let the apps do discovery. They are extremely good at putting you in front of people who have never heard of you. Treat that commission as a customer acquisition cost, which is what it is.

Own the repeat business. Once someone has ordered, the job is to move the second order direct.

• Put a card in every bag. Not a flyer — a card with a specific reason: a discount that only works on your site, or a dish that is only available direct.
• Make your own ordering genuinely easier than the app. If it is worse, nobody moves, and you will have proven the wrong thing.
• Collect the email address at checkout and actually use it, sparingly.
• Watch the split each month. The number that matters is the percentage of orders that are direct, and whether it is going up.

This is where most direct ordering fails. The apps have spent enormous sums making checkout frictionless, and a clunky alternative loses even to a 25% commission.

The bar: menu loads in under two seconds on a phone, no account required to order, address entry that autocompletes, card payment through Apple Pay or Google Pay in one tap, and an honest time estimate. Anything less and your regulars will go back to the app, and you will conclude direct ordering does not work — when what did not work was that particular checkout.