Your finance team closes the books on China and the numbers don’t tie out. Sales says the brand sold ¥4.2M on Tmall during the month. The bank received ¥3.6M. Your ERP recorded something in between. Nobody can fully explain the ¥600K gap — and until they can, your China P&L is an estimate, not a statement of fact.
China marketplace settlement reconciliation is the process of matching what a platform like Tmall, JD, or Douyin says it owes you against the orders you actually fulfilled and the cash that actually landed in your account. It is the single most overlooked step in achieving accurate, gross-to-net P&L visibility for cross-border ecommerce — and the reason so many global brands operating in China never trust their own China numbers.
Why marketplace payouts never match your order revenue
Gross merchandise value (GMV) is a vanity metric. The amount a platform deposits in your account is GMV minus a stack of deductions that are applied across different timelines, in different reports, and often in different currencies. The gap between the two is where margin quietly disappears.
A typical Tmall or JD settlement deducts some combination of the following before a single yuan reaches your bank:
- Platform commission — a category-based take rate (often 2%–8%) on the order value.
- Payment & transaction fees — Alipay/WeChat Pay processing and settlement charges.
- Promotional co-funding — your share of coupons, full-reduction discounts, and campaign subsidies applied at checkout.
- Marketing & traffic costs — Zhitongche / JingDong Express ad spend deducted directly from settlement.
- Returns, refunds, and chargebacks — reversed in a later settlement period than the original sale.
- Logistics & warehousing — Cainiao or platform-fulfillment fees for bonded and domestic stock.
- Withholding and VAT adjustments — tax handled differently for cross-border (bonded) versus general-trade orders.
Each of these lives in a separate line of a settlement file, and the platforms do not net them into a clean “here is your revenue” figure. If you only watch the order dashboard, you are tracking a number that no one will ever pay you. This is the same timing-and-truth problem we unpacked in why your China P&L is always two weeks late.
The five components of a clean settlement reconciliation
A reconciliation that finance can actually sign off on resolves five distinct questions for every settlement period:
1. Order-to-settlement matching
Every settled line must trace back to a fulfilled order, and every fulfilled order must eventually appear in a settlement. Orphaned settlements (cash with no order) and orphaned orders (revenue with no payout) are both red flags that something was mis-mapped or lost in transfer.
2. Gross-to-net bridge
For each order you need an explicit walk from GMV down to net deposit: commission, fees, promo co-funding, ad deductions, and returns each shown as a discrete adjustment. Without the bridge, the gap is a mystery; with it, the gap is a report.
3. Timing alignment
Sales, settlement, and cash rarely fall in the same accounting period. A 618 order placed on June 18 may settle in early July and refund in August. Accrual-accurate revenue recognition requires mapping each event to the right period — a discipline ERPs like NetSuite and SAP are built around, but which break the moment China marketplace data arrives late or in the wrong shape.
4. Currency and cross-border treatment
Settlements are denominated in RMB, but your reporting currency is not. Cross-border (Tmall Global, bonded) and general-trade (domestic) orders carry different tax and FX implications, as the regulatory specialists at China Briefing have documented at length. Using a single blended FX rate is one of the most common ways China net revenue gets misstated.
5. Returns and refund reversals
Returns are not negative sales in the period they occur — they reverse a prior settlement. If your reconciliation can’t link a refund back to its original order and period, your margin will look healthier than it is, then lurch downward a month later.
Tmall, JD, and Douyin settle differently
There is no universal China settlement format. Each platform exports its own files, on its own cadence, with its own fee taxonomy:
- Tmall / Taobao — settlement via Alipay with T+1 to T+15 cycles depending on store type; bonded (Tmall Global) and domestic flagship stores report separately.
- JD (JD.com) — self-operated (1P) and marketplace (3P) models settle on entirely different logic; JD Logistics fees are itemized apart from commission.
- Douyin (TikTok China) e-commerce — live-commerce driven, with creator commissions and ad-spend deductions that can dwarf platform fees during a campaign.
Because the schemas differ, a reconciliation built for Tmall will silently mis-map JD or Douyin data. We break down these structural differences in our Tmall vs JD vs Douyin data comparison, and the harmonization challenge is the heart of getting China marketplace data ERP-ready.
The reconciliation workflow, step by step
- Ingest raw order, settlement, refund, and ad-spend exports from each platform — at the line-item level, not summary totals.
- Normalize every platform’s fee taxonomy into one common chart of deductions so a “commission” means the same thing everywhere.
- Match settlements to orders and refunds to their originating sales, flagging anything orphaned.
- Convert RMB to reporting currency using period-correct rates, segmented by cross-border versus domestic.
- Bridge GMV to net deposit with every deduction itemized and auditable.
- Post period-accurate net revenue and fees into the ERP, where they roll up into a P&L finance can defend.
Done well, this is what produces the unified P&L reporting that lets a global brand compare China margin against every other region on equal footing.
Why manual reconciliation breaks at scale
Most brands attempt this in spreadsheets. It works at one store and one platform. It collapses the moment you add a second marketplace, a flash-sale campaign, or a returns spike — because the volume of line items, the fee-schema differences, and the timing mismatches exceed what any analyst can match by hand before the close deadline.
The failure mode is predictable: reconciliation slips from monthly to quarterly to “when we have time,” estimates harden into reported numbers, and leadership makes inventory and marketing bets on a China P&L that is directionally wrong. Reliable multichannel order management is the upstream foundation, but order accuracy alone won’t close the settlement gap.
Automating settlement reconciliation with a unified data layer
Digate sits between the China marketplaces and your ERP as a managed data layer that performs this reconciliation continuously rather than at quarter-end. It pulls settlement, order, refund, and ad-spend data from Tmall, JD, Douyin, and Pinduoduo, normalizes each platform’s fee schema into a single model, matches every payout to its order, converts currency at period-correct rates, and posts a clean gross-to-net bridge into NetSuite, SAP, or your ERP of record.
The result is a China P&L that ties to the bank, line by line, the day the books close — not an estimate defended six weeks later.
Frequently asked questions
What is China marketplace settlement reconciliation?
It is the process of matching the payouts a platform like Tmall or JD reports against the orders you fulfilled and the cash you received, itemizing every deduction (commission, fees, promotions, ad spend, returns, and tax) so that net revenue posted to your ERP reflects what you were actually paid.
Why doesn’t Tmall GMV match my bank deposit?
Because GMV is gross order value before platform commission, payment fees, promotional co-funding, advertising deductions, logistics charges, and returns are removed. These deductions are reported across separate settlement lines and timelines, so the deposit is always materially lower than GMV.
How often should settlement reconciliation run?
For brands with meaningful China volume, continuously. Settlement, refund, and ad-spend events arrive throughout the period, so reconciling once at month-end forces error-prone catch-up. An automated data layer reconciles as data lands, keeping the P&L close-ready at all times.