Is Uber Eats worth it for restaurants? The honest maths

Short answer: Uber Eats is a paid customer-acquisition channel wearing a sales-channel costume, and it is worth exactly what acquisition is worth to you. After the March 2026 fee changes (entry tier now 20%, effective rates near 30% on much mid-tier volume, 30% premium), an average order contributes a thin-but-real margin IF it is genuinely incremental: a customer you could not reach, on a night with idle capacity. The same order is a straight loss of margin when it replaces a direct order from a regular. So the honest answer is a strategy, not a yes or no: curated menu, marked-up prices, deliberate tier choice, and a standing programme that graduates repeat customers to your own channel. Here is each piece with numbers.

What Uber Eats is genuinely good at

• Discovery: app-native customers scrolling at 8pm who have never heard of you, this audience is real and hard to reach any other way.
• Idle capacity: a staffed, rent-paid kitchen on a quiet Tuesday sells marginal orders at marginal cost, thin margins on top of covered fixed costs are still contribution.
• Speed to delivery: fleet, tracking, support and payments exist on day one, standing up your own delivery is its own project (see delivery strategy).
• New-market entry: opening in a new suburb, the app is a legitimate launch amplifier while your own channels build.

The per-order maths, honestly

Both columns are true at once, which is why blanket judgements about delivery apps are useless. The strategic question is the MIX: what share of your marketplace volume is genuinely incremental? For a venue with strong direct channels, it is often under half, and every cannibalised order pays roughly ten dollars for the privilege of anonymising your own customer. The commissions guide has the full rate tables and a calculator for your numbers.

Run it deliberately: menu, prices, tier

Three controls most venues never touch. Menu: curate for the box, high-margin travellers only, no fragile dishes that arrive as refund requests, bundles that lift order value. Prices: mark up the marketplace menu 10–20% (within current platform rules), most venues do, customers broadly expect it, and it narrows the commission wound while making your own channel visibly cheaper. Tier: choose by measured order volume, not fear; the premium tier’s visibility is an advertising spend and should be judged like one (cost per incremental order), and stepping DOWN a tier is a legitimate experiment the platform will never suggest. Watch your weekly numbers split by channel so each tier decision has data behind it.

The dependence trap

The structural risk is not any single order’s margin, it is drift: the app is convenient, volume grows, marketplace share creeps from 15% to 40%, and one day the platform is effectively your landlord, setting rent (commission), controlling the storefront (ranking) and owning the tenants (your customers, whose names you never learned). The 2026 fee rises land precisely on venues that reached that position, because dependence is what a take-rate increase prices. The antidote is a number: set a maximum marketplace share (many operators choose 15–20%), review it monthly, and treat every point above the line as a project for the migration playbook below.

The migration playbook: graduate the regulars

• Make direct visibly cheaper: your channel at menu price against a marked-up app menu is a standing advertisement for switching.
• Make direct earn something: points, a first-direct-order offer, member specials that never appear on the app.
• Put the link where app customers already look: Google profile, Instagram bio, receipts, packaging, window decals (follow platform rules on what goes inside the order bag).
• Capture and remarket: every direct order joins the list; one good weekly message beats any ranking algorithm.
• Measure graduation: track marketplace vs direct share monthly; celebrate every point that moves.

Mistakes that make the app the owner

• Full menu, in-store prices, default tier, and no direct alternative: the passive setup that maximises platform take.
• Judging the channel by revenue instead of contribution per order.
• Running platform promotions permanently “for visibility” without pricing them as advertising.
• Letting fragile dishes travel, and paying for it in refunds and ratings.
• Having no marketplace-share ceiling, drift only ever goes one direction.

Frequently asked questions

Is Uber Eats worth it for a small restaurant?
It depends entirely on what job you give it. As a discovery channel that brings customers you could not otherwise reach, and as a way to sell idle kitchen capacity, it can be worth a 20–30% commission. As the default home for your regulars’ weekly order, it is a margin transfer from your business to the platform. The venues that win run it deliberately: marketplace for acquisition, their own channel for repeats.

How much does Uber Eats charge restaurants in 2026?
After the March 2026 fee changes: the entry tier is 20% (up from 15%), the mid tier lists at 25% but carries surcharges on member orders that push effective rates to about 30% on much of the volume, and the premium tier is 30%. Pickup orders are cheaper. Sponsored placement and self-funded promotions come on top, which is why measured all-in costs commonly land at 30–40% per order.

Do Uber Eats orders actually make restaurants money?
Run the contribution maths per order: order value minus commission, minus food cost, minus packaging, minus the incremental labour. At a 30% all-in rate and a 31% food cost, a $40 order leaves roughly $15 to cover packaging, labour and everything else, thin but positive for genuinely incremental volume on an already-staffed kitchen. It turns negative the moment the order replaces one you would have received directly.

Should I put my whole menu on Uber Eats?
No. Curate a delivery menu: dishes that travel well, carry strong margins, and survive 20 minutes in a box. Cut the loss-makers and the fragile items (they generate refunds and bad ratings), and consider delivery-only bundles that lift order value. Most venues also price the marketplace menu 10–20% above in-store to claw back commission, within the platform’s current rules.

How do I get my Uber Eats customers to order direct?
Give them a visible reason plus a friction-free path: your own ordering link on Google, socials, receipts and packaging; better prices on your channel (you can beat your own marked-up app menu comfortably); loyalty points and first-order offers that only exist direct; and faster pickup for direct orders. Follow the platform’s rules on inserts and messaging, and let regulars discover that the same order costs them less and earns them something.

What is a healthy marketplace share of my sales?
Many operators aim to keep marketplaces under roughly 15–20% of revenue: enough to keep the discovery engine running, small enough that the platform is not your landlord. The number that matters more than the share is the trend: if marketplace share is rising quarter over quarter while direct is flat, the platform is winning the relationship war, and the migration playbook needs to start now, not at 40%.