Why Restaurants Are Building Their Own Ordering Apps Instead of Relying on Just Eat and Uber Eats

The Commission Problem

Third-party delivery marketplaces typically take 20–30% of every order's value in commission, on top of payment processing fees. For a restaurant running on typical food-service margins (often 5–10% net), that commission isn't a rounding error — it can be the difference between an order being profitable and an order actively losing the restaurant money once ingredient cost, packaging, and staff time are accounted for.

Worse, the marketplace owns the customer relationship, not the restaurant. The customer's order history, contact details, and loyalty all sit inside Uber Eats or Just Eat's systems. A restaurant that relies entirely on marketplace traffic has no way to email a customer about a new menu item, send a loyalty discount, or even know that "the person who ordered three times last month" exists as a distinct customer at all.

What Building Your Own Platform Actually Buys You

When we built the ordering platform for Italian Pizza UK, the brief wasn't "replace the marketplace apps" — most independent restaurants still list on those platforms for discovery. The brief was to give the restaurant a *direct* channel that didn't cost 25% of every order, and that the restaurant actually owned.

The Driver App Is the Hard Part

Anyone can build a menu and a checkout form — that's genuinely the easy 20% of this kind of project. The part that actually differentiates a serious restaurant ordering platform from a glorified online menu is the driver-facing mobile app and the live tracking experience it powers.

For Italian Pizza UK, this meant building a dedicated driver app (not just a web page) that handles:

Getting the location-streaming architecture right (batching updates sensibly instead of streaming every GPS tick, handling app backgrounding on both iOS and Android correctly, degrading gracefully when signal drops) is where most "we'll just build a quick driver app" projects go wrong. It's the single most technically demanding part of an otherwise fairly standard ordering platform.

Collection vs Delivery Changes the Product

A detail that's easy to underestimate: collection orders and delivery orders are genuinely different products wearing the same menu. Collection needs accurate "ready by" time estimation (nothing frustrates a customer more than standing at the counter waiting for an order that isn't ready), while delivery needs the driver tracking and live ETA experience described above. Building both into the same platform means the order flow has to branch early — fulfilment method should be one of the first choices a customer makes, not an afterthought at checkout, because it changes what information the platform needs to collect and display afterward.

Loyalty Without Handing It to a Third Party

Once a restaurant owns its customer data, loyalty mechanics that would be impossible on a marketplace app become straightforward — a tenth-order discount, a birthday-month offer, a "we haven't seen you in six weeks" win-back email. None of this works if every customer relationship lives inside Uber Eats' or Just Eat's database instead of the restaurant's own. This is arguably the single biggest long-term value of owning the ordering platform: marketplace commission savings are a one-time-feeling win, but an owned customer database compounds — every repeat customer becomes cheaper to keep than the last one was to acquire, and the restaurant is the one who gets to decide how to use that relationship instead of asking a marketplace's permission.

Realistic Timeline and Scope

For restaurant owners actually planning this, it helps to know roughly what's involved rather than treating it as an abstract "build us an app" request. A production-grade build typically breaks into three connected pieces: the customer-facing web ordering experience (menu, cart, checkout, order tracking), an admin dashboard for the restaurant to manage the menu, view and fulfil orders, and see basic reporting, and — if a driver app is in scope — the separate mobile application covering order acceptance and live location streaming described above. Each of these can be scoped and delivered in phases; a lot of our restaurant clients start with web ordering and collection only, prove it out, and add delivery and the driver app as a second phase once the core ordering platform is live and generating real repeat-customer data.

Payment and Checkout Trust for a First-Party Platform

A subtle risk of moving away from marketplace apps: customers trust Uber Eats and Just Eat's checkout implicitly because they've used it dozens of times across different restaurants. A restaurant's own checkout doesn't get that inherited trust for free — it has to earn it through visible payment security signals (recognisable card logos, a proper SSL certificate, clear order confirmation messaging) and, just as importantly, through actually being reliable. A checkout that errors out or double-charges even occasionally will push customers straight back to the marketplace apps they already trust, undoing the entire point of building a direct platform in the first place. This is worth budgeting real testing time for — payment failure edge cases, network interruption mid-checkout, and duplicate submission handling all need to be deliberately tested, not just the happy path.

Do You Still Need the Marketplace Apps?

For most independent restaurants, yes — at least for now. Just Eat, Uber Eats, and Deliveroo bring discovery traffic that a standalone restaurant website simply can't replicate; people browsing those apps for "what's nearby" would never have found the restaurant's own site. The realistic strategy isn't "replace the marketplaces," it's "stop being *entirely dependent* on them." A restaurant with its own ordering platform can:

What This Costs to Build Properly

A genuinely production-grade ordering platform with a driver tracking app is not a weekend project — it's web ordering (menu, cart, checkout, payment), an admin dashboard for the restaurant to manage orders and menu items, and a separate mobile app for drivers with live location streaming. Budget and timeline vary a lot by scope, but this is realistically a multi-month build done properly, not a template site with a "buy now" button bolted on.

If you're running a restaurant or food business and the marketplace commission math has started to bother you, [talk to us](/contact) about what a direct ordering platform — with or without a driver app — would actually look like for your business.