How to escape delivery-app commissions: a step-by-step plan for restaurants

Aggregator commission — typically 20–30% per order on platforms like Pyszne.pl, Glovo, Uber Eats or Wolt — is, for many venues, the biggest variable cost after food and staff. The good news: it's also the only big cost a restaurant can actually control. The bad news: "let's just leave the platform" is not a plan. This is the plan.

The principle: you're not escaping platforms, you're escaping commission

An aggregator does one thing well for you: it can be the first point of contact. The problem is you let it do a second thing too — serve the customers who would have come back anyway. Paying for discovery is fair. Paying for loyalty is not.

The goal: a new customer may arrive through a platform, but their second order happens with you.

Step 1 — stand up a commission-free channel (week 0–4)

Nothing else works without this. You need a site with direct ordering: cart, online payment, the order landing in the kitchen (printer or screen), plus reservations if you seat guests. Two routes:

SaaS on your own domain — e.g. Gastronaut: flat monthly fee, live in weeks, zero commission.
A custom site with an ordering engine — when the brand and non-standard needs justify it. Costs broken down here: how much a restaurant website costs.

One hard requirement: ordering on your site must be easier than on the platform, not just cheaper for you.

Step 2 — migrate your regulars (month 1–3)

Customers don't check daily whether you've launched a website. Tell them — at the exact moment they're holding your food:

A flyer in every platform delivery bag: "Order direct next time — [10% off your first one]". You've already paid commission on that order; treat it as acquisition cost.
QR codes on receipts, tables, the counter.
Google Business Profile — the "order online" link should point at your site, not the aggregator. Same for Instagram and Facebook bios.
The phone — everyone calling in an order hears: "you can do it faster on our site".

A first-direct-order discount isn't a cost — it's cheaper than the commission you'd pay for that same order on the platform.

Step 3 — build the base, give reasons to return (month 2+)

On a platform you don't know who your customers are. On your own channel you do — with their consent:

a simple email/SMS opt-in at checkout,
occasional, concrete messages: new menu, seasonal offer, free tables this Friday,
a loyalty scheme, even the simplest kind (every 10th order).

Step 4 — shift the ratio, don't burn bridges (month 3+)

Keep the platform as a storefront for new customers — with a trimmed menu, or platform prices that absorb the commission if your contract allows it (check the price-parity clauses in your agreement).
Measure one number monthly: direct share of all online orders. If the trend rises, the plan is working.
With a stable direct channel you can negotiate with the platform from a position of "I can leave" — or genuinely shrink its role.

What it's worth — in numbers

At 600 online orders a month with a 90 zł average basket, a platform at 25% takes 13,500 zł every month. Moving half of those orders to your own channel returns ~6,750 zł of margin monthly — over 80,000 zł a year. Full plug-in-your-numbers table: let's do the commission math.

How it looks in practice

This is exactly the plan I'm running with Sushi Zushi in Warsaw — a premium brand operating since 2008 whose online sales used to live entirely on platforms and the phone. The new site takes orders and reservations commission-free, 24/7, and for the first time the customer base belongs to the restaurant.

Want this plan for your venue? Tell me your monthly online order volume — I'll reply with where to start and when it breaks even for you.