Is Delivery Actually Making You Money? A Channel Profitability Test for Asian Restaurants
Ask an operator in Kuala Lumpur, Jakarta or Bangkok how delivery is performing and the answer usually arrives as a single number: monthly platform sales, or the percentage of total revenue that now comes through an app. Both figures are easy to pull from a dashboard. Neither answers the question that matters, which is whether the restaurant keeps more money at the end of the month because the channel exists.
The regional numbers explain why the question has become urgent. Research by Momentum Works put Southeast Asian food delivery platform sales at US$22.7 billion in 2025, growing 18 percent year on year, with every one of the six major markets recording double-digit growth and Thailand expanding fastest. Its East Asia work estimated a further US$38.6 billion across Japan, South Korea, Taiwan and Hong Kong, with South Korea accounting for roughly three-quarters of that total.
One detail in that research matters more to operators than the headline growth. The expansion is being driven by more users ordering more often, not by larger baskets — average order values have edged downwards as platforms compete on affordability. For a restaurant, more orders at lower values with a fixed commission structure is a very different business from the same revenue delivered in fewer, larger tickets.
Gross platform sales is the wrong number
A delivery order and a dine-in order of the same menu price do not carry the same cost. The delivery order pays a commission, absorbs packaging, often carries a merchant-funded discount, and consumes kitchen capacity during exactly the hours when the dining room needs it most. What a restaurant should be comparing across channels is contribution margin per order — what is left after every cost that only exists because that order exists.
The comparison below shows where the two channels diverge. The cost items are structural; the weighting differs by market, format and contract.
| Cost element | Dine-in order | Third-party delivery order |
|---|---|---|
| Food cost | Recipe cost | Recipe cost, plus portion variance for travel-safe items |
| Platform commission | None | Negotiated per merchant, not published |
| Packaging | Minimal | Container, lid, sealing, bag, utensils, sauce cups |
| Promotions | Operator-controlled | Often merchant-funded to maintain ranking and visibility |
| Payment and admin fees | Card or e-wallet fee | Settlement, integration and platform service charges |
| Beverage and dessert attachment | High | Typically lower |
| Kitchen capacity | Planned | Competes with peak dine-in tickets |
| Customer relationship | Owned | Held by the platform |
The four-number delivery test
Rather than debating whether platforms are fair, the more productive exercise is to run four calculations. Together they tell an operator whether delivery is a profit channel, a marketing channel, or a slow leak.
1. Contribution margin per delivery order
Take the platform menu price. Subtract commission, recipe cost, packaging, the merchant-funded share of any active promotion, and payment or service fees. What remains is what the order contributes towards rent, salaries and profit.
An illustration, using assumed figures for a mid-priced casual restaurant: a dish listed at RM35 in the dining room is listed at RM40 on the platform. Recipe cost is RM11. Commission at an assumed 25 percent is RM10. Packaging is RM1.80. The restaurant’s share of a running promotion is RM3. Contribution is RM14.20. The same dish sold at RM35 in the dining room, with no commission and no packaging, contributes RM24. The delivery order looks larger and earns roughly forty percent less. Operators should run this with their own contract terms, because the conclusion changes entirely at different commission levels and different average order values.
2. Incrementality
This is the number almost nobody calculates and the one that decides everything. Of the delivery orders received last month, how many came from customers who would not otherwise have walked in, phoned, or collected? An order that genuinely adds demand is worth accepting at a lower margin. An order that simply moves an existing customer onto a commissioned channel reduces profit on business the restaurant already had.
The practical test is to look at total covers and total revenue across all channels over twelve months, not delivery in isolation. If delivery sales rose while combined dine-in and takeaway sales fell by a similar amount, the channel is substituting, not growing. Postcode data, daypart patterns and repeat-customer overlap help refine the picture.
3. Capacity cost
Delivery orders arriving at 12:20pm and 7:40pm are not free. They occupy the same pass, the same wok station and the same staff as the highest-value tickets of the day. Measure ticket times on both channels during peak periods. If dine-in service degrades when delivery volume spikes, the true cost of the delivery order includes slower table turns, weaker reviews and lost drink attachment in the dining room.
Some operators solve this by restricting delivery availability during peak service, running a separate prep line, or limiting the delivery menu to items that hold well and cook fast. All three are legitimate answers. Ignoring the conflict is not.
4. Dependency
Calculate the share of total revenue arriving through platforms the restaurant does not control, and the share of customers whose contact details it will never own. Regional data makes clear that dependency is a choice rather than a given — food delivery penetration ranges from a low single-digit share in Japan to more than twenty percent in South Korea, despite broadly comparable income levels and urban density. Markets differ because operators and platforms pushed them in different directions.
A restaurant generating most of its orders through one aggregator has outsourced its demand, its pricing power and its customer relationships in a single decision. That may be an acceptable trade for a delivery-first brand with a cost base built for it. It is rarely acceptable for a full-service restaurant with a long lease and a dining room to fill.
Five mistakes that distort the answer
- Using dine-in food cost percentages on a platform menu. Marked-up delivery pricing flatters food cost percentage while contribution per order falls. Work in currency, not percentages.
- Treating promotional spend as marketing rather than cost of sale. If a discount is required to stay visible on the platform, it is a permanent channel cost.
- Excluding packaging. Container, sealing, bag and utensils are small individually and material across thousands of orders.
- Comparing this year’s delivery sales to last year’s delivery sales. The meaningful comparison is total business across all channels.
- Assuming published commission benchmarks apply. Major platforms in the region negotiate terms outlet by outlet and do not publish fixed tables. Model the contract actually signed.
What to do once the numbers are in
A negative or marginal delivery contribution has three credible responses, and the right one depends on why the number is weak.
- Re-engineer the delivery menu. Not every dish should travel. Build a shortlist of items with strong margin, short cook time and good hold quality, then price them for the channel rather than copying the dining room list. The same discipline applied to a dine-in menu applies here.
- Negotiate from evidence. Volume, growth rate, cancellation rate, preparation time and rating are all levers. Operators who arrive with twelve months of clean channel data negotiate better terms than those who arrive with a complaint.
- Build the direct channel deliberately. Own-website ordering, pickup incentives and membership do not replace aggregators, but they change the balance of power. A restaurant with a functioning direct channel can choose its level of platform dependency.
When a delivery milestone becomes a documented achievement
Operators who measure the channel properly usually discover something else: they are sitting on genuinely notable numbers. Orders fulfilled from a single kitchen in a day. Sustained ranking within a category in a defined market. Volume of a signature dish delivered over a year. These are the kinds of achievements that F&B businesses mention casually in a founder interview and never substantiate.
The distinction worth understanding is between a marketing claim and a documented one. A marketing claim asserts scale. A documented achievement states what was measured, over what period, under what conditions, and can be checked by someone outside the business. Where a milestone is specific, measurable and independently verifiable, it may qualify for record certification in Asia rather than remaining an unsupported line on a website.
Asia Record documents measurable achievements by businesses, organisations and individuals across the region, and its record holder directory already includes restaurant chains, café brands and food manufacturers. For an F&B operator whose delivery data shows a defensible regional first or largest, it is reasonable to assess whether the achievement meets the criteria and to apply for Asia Record recognition. Record recognition documents a business achievement; it does not substitute for food safety approvals, licensing or any other regulatory requirement.
The discipline underneath
Delivery in Asia is not going to shrink, and the operators who do best with it are not the ones who resent it. They are the ones who treat it as a distinct channel with its own P&L, its own menu, its own capacity plan and its own performance data — and who know, to the ringgit, baht or peso, what each order is worth. That number is the difference between a channel that funds the next outlet and one that quietly funds somebody else’s.
