Own Website vs Uber Eats & DoorDash: The Real Cost for Australian Restaurants in 2026

A restaurant doing $40,000 a month in delivery on Uber Eats Marketplace hands over $12,000 of it in commission. Every month. That's $144,000 a year. The same volume through the restaurant's own website, with payment processing and software included, costs closer to $2,000 a month.

That gap, $10,000 a month, is the real Uber Eats commission story for Australian restaurants. Not the headline percentage. The dollars.

Australia's delivery market changed sharply on 26 November 2025 when Menulog ceased operations after 20 years. Two marketplaces remain: Uber Eats and DoorDash. Less competition between them, more pricing power for the pair that survived. In March 2026, four months after Menulog's exit, Uber Eats raised its Lite-tier delivery commission from 15% to 20% and pickup from 6% to 7%. More hikes will follow. The economics are not running in the operator's favour.

What follows is the actual rate table for each marketplace as of May 2026, the hidden costs that don't appear on the dashboard, the honest trade-offs of running your own webstore, a worked AUD example, and a decision framework you can apply to your own numbers.

What Uber Eats actually charges Australian restaurants in 2026

Uber Eats publishes its Australian merchant rates on merchants.ubereats.com/au/en/pricing. As of May 2026, the structure looks like this.

Marketplace pricing — listed in the Uber Eats app, drivers fulfilled by Uber:

• Uber Delivery: 30% commission per order
• Self-delivery: 16% commission per order (you handle the driver)
• Pick-up: 6% with validated in-store pricing, 10% otherwise (rose from 6%/9% in March 2026)

Webshop pricing — Uber Eats provides a "commission-free" branded online ordering page that uses their delivery network:

• Uber Delivery: 25% per order
• Self-delivery: 6% per order
• Pick-up: 6%

Tiered marketplace plans — most Australian restaurants on Marketplace sit in one of three tiers. After the March 2026 rate change:

• Lite: 20% (raised from 15%)
• Plus: 25% standard, 30% on orders from Uber One members
• Premium: 30%

Plus, there's a $500 (ex GST) activation fee per first location, $350 for subsequent locations, and a $200 (ex GST) damage fee per device for lost or damaged hardware.

None of this includes Sponsored Listings, Uber's pay-per-click advertising layer. Without sponsored placement, your tile sinks below competitors that buy visibility. With it, you pay for every click on top of the commission.

In practice, an Australian restaurant on Marketplace with Uber Delivery and some sponsored placement is looking at an effective rate north of 30% on delivery orders. That figure is what to plug into the math, not the headline percentage.

What DoorDash actually charges Australian restaurants in 2026

DoorDash's Australian commission page lists three partnership plans, the same structure used in DoorDash's other markets:

• Basic: 15% delivery commission, 6% pickup. Cheapest, lowest visibility.
• Plus: 25% delivery commission, 6% pickup. Includes DashPass listing (DoorDash's loyalty subscribers tend to order more often).
• Premier: 30% delivery commission, 6% pickup. Adds automatic ads and a "20-order guarantee" that refunds your commission if you receive 20 or fewer orders in a month.

Most operators on DoorDash gravitate to Plus or Premier because Basic-tier restaurants struggle for visibility. The Premier 20-order guarantee is useful protection for new venues, but if you're already hitting volume, you're paying 30% to stay competitive.

The structural picture is similar to Uber Eats: a "cheaper" tier exists, but the real-world choice for an active restaurant lands at 25–30% on delivery orders.

The Menulog closure and why it matters

Menulog operated in Australia for 20 years before its parent company, Just Eat Takeaway, shut down Australian operations on 26 November 2025. Customers were redirected to Uber Eats. About 120 staff were made redundant.

The competition picture is now narrower. The Conversation and the ABC both flagged the same concern: with one fewer marketplace competing for restaurants, the remaining two have more pricing power. The March 2026 Uber Eats fee hikes, announced four months after Menulog's exit, are consistent with that read.

For Australian operators, the practical takeaway is simple. Marketplace commissions are not heading down. Whatever the rate today, plan as though it will be higher in 18 months.

The hidden costs of marketplaces

The headline commission percentage isn't the whole story. A few line items that don't show up on the standard fee table:

Sponsored placement. Without paid ads, your visibility in the app decays as competitors buy higher tiles. The cost varies, but a restaurant doing meaningful volume in a busy postcode often spends $300–800 a month on Uber Eats Ads alone, on top of the commission.

Uber One surcharge. On the Plus tier, orders from Uber One members trigger a 5% premium, bringing those orders to 30%. Uber One penetration is growing. The effective rate on a "25%" plan creeps up over time without any new contract.

Discount control. Marketplaces frequently run promotions on your prices to lift conversion. You typically share the discount cost. The customer sees "50% off your favourites" but the restaurant pays for the discount alongside the commission.

Customer data lock-in. You don't receive the customer's email, phone number, or order history. You can't run a loyalty programme on someone who came in through the marketplace because you don't know who they are. Every order is anonymous from the restaurant's perspective.

Brand dilution. Your restaurant is one tile among 200. The marketplace is the brand the customer remembers, not yours. When they're hungry again, they open the Uber Eats app, not your website.

One-off fees. Activation ($500 first location, $350 subsequent), damage fees ($200 per device), late-onboarding service charges. Real numbers that add to the all-in cost.

Add it up and the "30% commission" framing understates what marketplace dependency costs an Australian restaurant.

The honest costs of running your own website

A direct-ordering channel is not free. The trade-offs are real and worth naming.

Payment processing. Online card payments in Australia run roughly 1.7–2.2% (plus a fixed per-transaction fee, often around 30c) depending on your provider.

Software / platform cost. A purpose-built restaurant webstore charges either a flat SaaS fee or a smaller commission. Platforms like Oddle Shop sit around 3% all-in, with the software and customer-data layer included. SaaS-only options vary widely.

Delivery fulfilment. You either run your own drivers (best for venues with consistent volume) or plug into a per-delivery network. Uber Direct, DoorDash Drive, and Australian on-demand services like Sherpa typically charge AU$8–12 per metro drop. The cost is per-order, not per-revenue, which is why higher average order values favour direct ordering.

Marketing. Marketplaces drive demand for you. Your own website doesn't. You need to fill the funnel through email, SMS, social, Google, SEO, and walk-in conversion. This is real budget, but the cost-per-acquired-guest is recoverable because the relationship belongs to you. Once a guest orders directly twice, you've effectively replaced the marketplace's role for that customer at a fraction of the cost.

Setup. A day or two for a standard webstore. Longer if you're building from scratch or integrating POS systems.

Be honest: direct ordering only makes sense if you have, or can build, demand. For a brand-new restaurant with no audience and no email list, marketplaces are doing real work. The decision changes once volume exists.

A $100 order, two channels, line by line

Concrete math. A $100 order on Uber Eats Marketplace with Uber Delivery, then the same $100 order on your own website with Uber Direct fulfilment.

Uber Eats Marketplace (Uber Delivery):

• Customer pays $100
• Uber Eats commission (30%): –$30
• Sponsored Listings attribution (estimated, $5 per converted click): –$5
• Restaurant nets: $65 before food cost, labour, packaging

Own website (Oddle Shop + Uber Direct delivery):

• Customer pays $100
• Software fee (~3%): –$3
• Payment processing (~2%): –$2
• Uber Direct delivery fee (per-drop): –$10
• Restaurant nets: $85 before food cost, labour, packaging

Per-order retained: $20. Across 100 orders per month, that's $2,000. Across 500 orders per month ($50,000 in revenue): $10,000 per month, $120,000 per year.

The own-website cost line scales linearly with orders. The marketplace cost scales linearly with revenue. As your average order value grows, the gap widens further.

A few notes on the assumptions:

• Larger orders close the gap from the delivery side (the $10 Uber Direct drop is fixed regardless of order size, so a $150 order on direct nets you $40 more than the same order on marketplace).
• If you run your own delivery (in-house drivers), the per-order delivery cost drops further once you reach a threshold of volume — typically around 30 deliveries a day.
• The marketplace number gets worse if you're paying for Sponsored Listings aggressively. The own-website number gets better as your repeat rate climbs and your acquisition cost amortises.

When marketplaces still make sense

The math doesn't say "marketplaces are bad." It says "marketplaces are expensive customer acquisition." Used that way, they have a legitimate role.

New restaurants with no audience. If you've just opened, the marketplace is the cheapest way to get strangers to try your food. Pay the 30% on those first orders. Treat them as marketing spend, not as the sales channel.

Cuisines that diners search for in-app. Some categories — Japanese, burgers, pizza, Thai — have strong pull demand inside the marketplace. People go looking for them by category, not by restaurant name. Showing up there matters.

Off-peak hours. Marketplaces can fill kitchens during quiet windows your direct channel doesn't reach.

Geographic expansion. Testing a new suburb where you have no brand presence yet. Marketplaces give you a soft launch.

The framing that works: marketplaces are an acquisition channel, not a dependency. The minute a customer becomes a regular, your job is to move them to direct ordering, where the unit economics work.

The hybrid model: running both without cannibalising

Most Australian operators end up running both marketplaces and a direct channel. Done well, that's the right answer. Done badly, you spend on both and convert neither.

A few practical rules that work in practice:

• Match or beat marketplace prices on your own website. If your direct channel is more expensive than Uber Eats, your regulars will keep ordering through Uber. Match the price, or price the marketplace 10% higher to recover commission while staying competitive.
• Bag inserts on every marketplace order. A flyer with a 15% discount code for the next direct order. The cost is trivial, the conversion is real. This is the cheapest customer-acquisition channel a restaurant has.
• Loyalty programme on the direct side, not the marketplace. Every $1 in repeat behaviour you can drive to direct ordering is a multiplier on lifetime value. Customer Intelligence tools that join order history across direct ordering, reservations, and walk-ins make this far easier than running it manually.
• Track repeat rate by channel. If the same customer has ordered through Uber Eats three times in a month, that's three commissions paid on a regular you could have converted. Email them. Offer them something direct.
• Don't run aggressive promotions on the marketplace. You pay the discount and the commission. Save the deep discounts for your own list.

Choosing a food delivery platform isn't a binary decision. It's a portfolio. Most successful operators we work with end up at roughly 60–70% direct, 30–40% marketplace by revenue within 12–18 months of taking direct ordering seriously.

A decision framework for Australian operators

Quick rules of thumb, indexed to monthly delivery revenue. These are rough thresholds, not laws.

Under $5,000/month in delivery revenue. Marketplaces only. You don't have enough volume to justify the time spent setting up and marketing a direct channel. Use marketplaces to find your audience.

$5,000–$15,000/month. Start your own webstore. Run marketplaces in parallel for discovery. Begin building your email list. The math is already meaningfully in your favour on the orders you can shift to direct.

$15,000–$40,000/month. Direct ordering is mathematically better. Every $10,000 in monthly delivery revenue routed through your own website instead of Uber Eats Marketplace saves roughly $2,500–3,000 in commission and platform fees. Use marketplaces as a customer-acquisition channel only.

$40,000+/month. The marketplace commission line is costing you a full-time staff member's salary every month. Sometimes two. At this volume, the question is no longer "should we have direct ordering" but "how aggressively can we move existing customers off marketplaces?"

A back-of-envelope way to size it: every $10,000/month in delivery revenue on Marketplace Uber Eats with Uber Delivery costs around $3,000 in commission. The same $10,000 through your own webstore costs around $500. The breakeven on setup happens fast — usually within one to two months for any restaurant doing meaningful volume.

Where Oddle Shop fits

Oddle Shop is a high-converting webstore built specifically for restaurants. Not a generic e-commerce template, not a marketplace skin. A storefront purpose-built for menu, ordering, scheduling, dietary preferences, and the rhythms of a busy service.

Pricing is transparent: a single commission rate, no activation fee, no per-device damage charge, no Sponsored Listings layer that the platform forces you into to stay visible. The customer data — email, phone, order history, average order value, day-part patterns — stays with the restaurant, ready to plug into Oddle's Customer Intelligence and Marketing products for loyalty and re-engagement.

Delivery integrates with Uber Direct and DoorDash Drive on the fulfilment side, so you keep the driver network without the marketplace commission. The on-premise side — Oddle Reserve for reservations, Oddle Terminal for in-restaurant payments — connects to the same guest record. Every channel feeds one customer profile.

If you'd like to see the numbers run against your own restaurant's revenue, book a demo and we'll walk through it line by line.

The point, plainly

Marketplaces are not the enemy. They are an expensive customer-acquisition channel that earns its place for new venues and discovery. They are not, for most Australian operators above $15,000/month in delivery revenue, the right place to do volume.

The Menulog closure shrunk the marketplace pool to two. The March 2026 Uber Eats rate change pushed prices up, not down. Whichever way the next 18 months go, the operators who own their own channel will be the ones who keep margin on the table.

Do the math against your own numbers. A worked example is good. Your worked example is better.

Related guides

• The ultimate guide to finding the best restaurant reservation system in Australia — the AU companion piece on the reservation side of the operator stack.
• The ultimate guide to online ordering systems for restaurants — the pillar guide on how restaurant online ordering should work in 2026.
• Choosing a food delivery platform: is the marketplace really the best? — the global view on marketplace vs direct, useful background to this AU-specific breakdown.