Direct online ordering vs third-party apps.

Commission commonly runs 25 to 30% of the order value, depending on the platform, the country and whether you're using their drivers or your own. On a €30 order that's €7.50 to €9 before you've bought a single ingredient.

The commission isn't the whole bill. Add card and service charges, any promotion you've agreed to inside the platform, and the discounts you fund when you join a campaign to stay visible in the app. Then there's the pricing question: if you raise your in-app prices to cover the commission, the customer sees a more expensive restaurant, and some platforms have rules about that in the contract you signed.

Do the sum with your own statements. Pull 90 days, add commission, in-app marketing and card fees, divide by orders, and you'll have your real cost per order. Then compare that with your gross profit per order. Plenty of restaurants discover the busiest platform is the one they make least on.

More than the average sales pitch admits, and it's worth being straight about it.

They're where a lot of people start looking for food, particularly in a new area or at eleven on a wet Tuesday. They have a driver network you'd struggle to replicate at small volume. They handle the payment, the fraud risk and most of the refund arguments. They spend heavily on their own advertising and you're inside it. And for a new restaurant with no reputation and no list, they can fill a kitchen in week one, which nothing else will.

You keep the money, and you keep the customer.

The money part is arithmetic. The customer part matters more over a few years: names, contact details and order history, which means you can bring people back rather than paying to reach them again. You also control the things the platform decides for you today. Your prices, your delivery radius, your minimum order, your opening hours, what's promoted at the top of the menu, whether you close orders at nine because the kitchen is drowning.

There's a resilience argument too. When a marketplace changes how it ranks restaurants, the ones that only exist inside it find out on a Friday. A direct channel with a list behind it doesn't move because somebody else changed a setting.

Do the sum on your own numbers rather than anybody's marketing.

Say you're doing 200 online orders a month at an average of €28. That's €5,600 in sales. At commission of 25 to 30% you're handing over roughly €1,400 to €1,680 a month. Against that, a direct channel costs you a platform fee, card processing on the order, and delivery if you're doing it yourself. On Eclyde that fee is €59 a month plus 5% per order, or €199 a month flat with no commission at all, €169 if you pay for the year.

Be honest about the two costs marketplaces are absorbing for you: getting the customer, and getting the food there. Direct ordering doesn't remove those, it moves them onto your side of the ledger, where they're cheaper but not free. Budget something for the bag inserts, the signage and the messages that bring people across, and cost your delivery properly using the next section.

One of three ways, and the right answer depends on how tightly packed your orders are.

Your own driver is the cheapest per drop once you have density, and the most expensive when you don't. Cost it properly: total shift cost including wages, fuel and insurance, divided by the drops actually made in that shift. If you can't consistently get three or more drops an hour, either your radius is too wide or your delivery fee has to cover the gap. Tighten the radius before you drop the idea, because a shorter radius makes the maths work and the food arrives hotter.

Second option, a courier company on a per-drop fee rather than a percentage. That fee is fixed, so it hurts less as order values climb. Third, use the marketplace fleet for delivery while taking the order yourself, which some platforms offer at a lower rate than full marketplace commission. Collection is the quiet fourth option: promote it properly with a real incentive and a lot of customers within a kilometre will happily come to you.

Usually not, and certainly not on day one of running your own channel.

The test is what proportion of your app orders are genuinely new customers. If a large share are people who already know you and are ordering through the app out of habit, the app isn't buying you discovery, it's charging you 25 to 30% to serve your own regulars. Move those. Keep the platform for the people who genuinely find you there, and watch what happens to that share over six months.

There are real reasons to stay. A new site with no reputation. A quiet area where the app's demand carries your Tuesdays. A cuisine people browse for rather than search for. And there's a real risk in going: you'll lose some orders, at least at first, from customers who won't leave the app for anybody. Going hybrid first means you learn how much that number is before you bet the business on it.

Physically, and with a reason worth changing a habit for.

Every marketplace delivery is a bag going into somebody's house. Put a printed card in it with a QR code to your ordering page and one clear offer for the first direct order. Print it on card that survives grease. A sticker on the box or a branded sleeve does the same job and doesn't get thrown away with the receipt. At the counter, ask, and mean it: tell collection customers the site is cheaper and takes a minute.

Then look after the ones who cross. Their first direct order should be smooth, quick and slightly rewarding, because that's the order that decides whether there's a second. Capture the email or number, start the loyalty count, and send a thank you that doesn't read like a mailshot. There's more detail in the guide on getting more restaurant orders.

In principle yes, since the commission isn't there to fund. In practice, read your platform agreement first.

Some marketplaces have rules about pricing lower elsewhere, and they change from time to time and by country. Check what you signed before you build a campaign around undercutting them. Where you can price differently, keep the gap modest and visible: the same menu at a slightly better price, or the same price with a free side and free collection. A large gap makes your app menu look like a rip-off to the same customers you're still serving there.

An alternative that sidesteps the whole question: keep prices identical and put the value in things the app can't do. Loyalty that only counts on your own site. A wider menu. Bigger portions on family deals. Order-ahead for collection at a set time. You're not in a price war, you're giving people a reason to deal with you directly.

The technology takes days. The habit takes months.

Getting an ordering page live, the menu loaded, payments connected and printing to the kitchen is a short piece of work with a decent provider. What takes time is the drip: bag inserts on every order, the QR code on the box, the counter conversation, the follow-up message after a first direct order, the loyalty count ticking along. Expect a slow first month and a noticeably different picture by month three or four.

Judge it on one number: the share of your online orders that came through your own channel this month. Watch that climb rather than watching total orders, which will move around with the weather regardless of what you do.