Takeaway.com and Uber Eats, managed for Luxembourg restaurants

Because a delivery radius holds a finite number of people who are hungry at 19:40 on a Tuesday. Past a point, extra budget buys repeat impressions to people who have already seen you: cost per order climbs while order count barely moves. What still works is when the money is spent, not how much of it there is.

So we pace budget across hours and weekdays — concentrated on your peaks, paused in windows where your kitchen was never going to win the order anyway. Across the Dutch storefronts we manage that has cut roughly 30% of wasted ad spend, and the best of them have reached up to 14x return on ad spend by spending in fewer, better hours.

By reading it rather than assuming it. A storefront here can serve people whose weekday lunch and whose weekend evening happen on different sides of a border, which makes one restaurant behave like two businesses sharing a menu. So we read your own hour-by-hour pattern before recommending a promotion calendar or an opening-hours change.

Language behaves differently here too. French, German, Luxembourgish and English can sit inside one delivery radius rather than being split across separate regions, so the answer is rarely a menu per language. It is item names that stay legible to a mixed audience, with the description doing the persuading.

It makes what happens after the first order worth more than the first order. In a small catchment the same households come back, so the levers that compound are the ones attached to every visit: menu conversion, combos and sides, and a storefront that still sells when no promotion is running. The 38% basket-size gain across the Dutch storefronts we manage came from exactly that kind of work.

That is why we start with the menu rather than the ad account. A storefront converting under 20% of its visitors is where we begin, because in a small market you run out of new customers to waste on a page that loses them.

A plan per storefront with the expected result written next to it: which offer, on which items, in which hours, and what it should do to the payout you keep from that peak window rather than to the order count on a dashboard. You approve it, amend it, or reject it. Nothing goes live until you have.

We work remotely from Amsterdam, inside your platform accounts and with your permission, and our 60+ partner restaurants are in the Netherlands rather than here. In a small catchment that gap matters more than it would elsewhere: there is no room to learn on your customers, because the households you would spend twice to reach are the same ones either way. So the first weeks go on the menu and the baseline, where the work is safe and the effect is readable, before any budget moves.

Pricing is performance-based against that baseline, so a quiet month costs us too. Reporting runs against it as well, rather than against the platform dashboards, which are generous about gross orders and quiet about what reached your bank account.