Who funds the discount inside a 30 minute order

How much does a 30 minute delivery actually cost the merchant who fills the order? As of November 12, 2026, the public record offers a range rather than a number. Alibaba, JD.com and Meituan have together pledged close to 200 billion yuan to subsidise one hour delivery, and merchants who track their own numbers put the effective cost of an instant retail order at 18 to 25 percent once delivery fees, mandatory promotions and platform coupons are counted. The distance between those two figures is what a brand's channel team has to price before the next campaign, because the subsidy is announced by the platform while the cost is booked by the merchant.

What a subsidy total cannot show

Subsidy figures are the easiest number to publish and the hardest to read, because they measure intent rather than outcome. A pledge of 200 billion yuan across three platforms says how much each company is prepared to lose, not what a merchant pays. The two numbers connect only through the promotion calendar, which is where the transfer happens: a platform that funds a discount from its own budget still needs the merchant to hold the shelf price, and the merchant's contribution is booked as a promotion rather than as a service fee, so it never appears in the commission line a brand negotiates.

The order split that decides who can subsidise

Endurance depends on how much of the delivery market a platform already holds. By the last quarter of 2025 the instant retail order split ran at roughly 45.2 percent for Taobao Shangou, 45.0 percent for Meituan, 8.4 percent for JD.com and 1.5 percent for Douyin, and the three leaders had burned an estimated 150 billion yuan or more pushing daily meal orders from around 80 to 90 million to a peak above 200 million. Meituan, with the most share to defend, swung to a quarterly loss of about 3 billion US dollars as the fighting peaked in early 2026.

Two conclusions follow. A platform that already owns couriers and stores can subsidise an order more cheaply than one that has to rent both, which is how JD.com entered late and still holds a visible position in the order split. And the cash position, not the quarterly loss, sets the endurance: net cash of 400, 144 and 110 billion yuan is the reason none of the three stopped when losses peaked, and it is also the reason a smaller retailer cannot answer a subsidy with a subsidy.

Two costs that never reach the invoice

The merchant side of the arithmetic is harder to see because it is spread across several lines. A university study cited by Yicai found combined takeout and dine in profits for merchants fell about 1.7 percent in the early phase of the subsidy war and as much as 8.9 percent during its most aggressive stretch. Restaurants that opted out of the promotions lost customers almost immediately, while those that opted in kept volume and watched margin shrink, which is the definition of a cost that cannot be refused.

For a consumer brand the same mechanism shows up in three places: the discount depth required to keep a product listing live during a campaign, the buffer stock held to avoid a stockout at peak, and the pick labour added to a retail store that was never laid out for assembling orders. None of the three appears as a commission, and all three reduce the contribution margin per unit that a category manager is measured on at the end of the quarter.

Where the money reappears in a merchant's numbers

The regulator moved before the merchant arithmetic was settled. China's State Administration for Market Regulation fined seven major platforms a combined 3.6 billion yuan on 17 April 2026, officially for ghost delivery and food safety violations, then followed in June 2026 with draft rules banning open ended large scale subsidies and requiring platforms to give merchants seven days notice before launching a promotion. Both platform chief executives went on record afterwards saying the industry needed to return to what they called rational competition.

What the rules change is not the level of discount but its predictability. A seven day notice requirement turns a subsidy from a surprise into a line a merchant can plan against, and a ban on open ended subsidies pushes competition toward capacity, where cost is fixed and payback depends on volume per node. That shift favours platforms with dense store networks, and it favours brands that can commit volume to a smaller number of stores instead of spreading thin across a listing that never sells through. The earlier BXT Data note on node coverage sets out why completion rate, not store count, is the number to audit.

Why the category mix matters more than the total

Instant retail is growing at roughly two and a half times the rate of conventional e commerce, and projections put the segment above 2 trillion yuan in sales by 2030, which makes it the fastest growing channel most consumer brands can reach. The growth is uneven inside the category. Emergency purchases such as medicine, bottled water and one missing ingredient carry high frequency and low basket value, while planned household purchases carry the opposite profile and are the ones that justify holding buffer stock inside a store.

That is why a platform level growth figure says very little about a specific brand. A brand whose range sits in the emergency basket competes on availability at the moment of need and is exposed to every stockout. A brand whose range sits in the planned basket competes on the breadth a store can hold, and its ceiling is set by shelf space rather than by rider supply. Reading category growth without splitting it this way is how a brand ends up funding promotions in a channel where it can never hold enough range to earn them back.

What a brand should settle before the next campaign

The useful preparation is not another discount plan. It is a written definition of what the channel may do to the brand's margin, agreed before the platform's promotion calendar arrives, plus the four numbers that make the trade visible after the fact, when the invoice and the campaign report disagree. The fourth of those is a price monitoring routine, because the gap between shelf price, campaign price and settled price on the same twenty SKUs is exactly where a merchant funded discount hides.

FAQ

• Is a 30 minute promise a distribution decision or a marketing one?

  It is a distribution decision with a marketing surface. Whether an order can be completed inside the window depends on store range, staffing and stock, and only after that does the promise become something a campaign can advertise.

• Why do platforms report order share instead of category growth?

  Order share is stable and auditable within a quarter, while category growth moves with the promotion calendar. The first is a capital story that supports a subsidy budget, the second is an operating story that a brand can act on, so the two are reported for different audiences.

• Does the subsidy ban remove the cost from merchants?

  No. It removes the unpredictability. A merchant that knows a promotion is coming seven days ahead can plan stock and staffing, but the price contribution and the buffer stock are still funded by the supplier of the goods.

• What should a small brand without a store network do?

  Pick the two or three stores in each city that reliably complete orders and concentrate range and stock there. Coverage across many listed but unstocked stores costs money and produces stockouts that the platform records against the brand.

Sources

• The Economic Times on Reuters reporting: Chinese ecommerce giants make expensive bets on fast deliveries
• Verified Market Research: Chinese e-commerce giants escalate the instant retail price war
• Marketing to China: order split, merchant cost and the April 2026 platform fines
• The Edge Malaysia on Reuters: China Singles Day sales rise 4.8% to US267b