Your China team can close the books. The harder question comes six weeks later, when your external auditors sit down and ask a deceptively simple thing: “Show us how you know this Tmall revenue number is right.” For most Western brands, that is the moment the China P&L stops being a finance problem and becomes an audit-evidence problem. The reconciliation may be correct — but if you cannot walk an auditor from a platform settlement statement, through your reconciliation logic, to the exact journal entry in NetSuite or SAP, the number is unauditable. And unauditable revenue is a control deficiency, not just an inconvenience.
Auditing China marketplace revenue is the process of producing a complete, traceable evidence trail that links each recognized revenue and cost figure — gross merchandise value, platform commissions, refunds, VAT, and net settlement — from the original Tmall, JD, or Douyin settlement source, through documented reconciliation logic, to the specific ERP journal entries and group-consolidation lines, in a way that an external auditor or ICFR/SOX reviewer can independently re-perform. The goal is not a prettier report; it is defensible traceability at transaction level.
The short answer
- Auditors do not test your P&L number — they test whether you can re-derive it from source. The evidence trail, not the total, is what passes or fails.
- For China marketplaces the source of truth is the platform settlement statement reconciled to the bank deposit, not the seller-center order screen.
- Every recognized figure needs a documented path: settlement → reconciliation → gross-vs-net treatment → ERP journal → consolidation.
- The three deficiencies auditors flag most often are unexplained gross-to-net gaps, refund timing across period-end, and revenue booked without a matching VAT/fapiao trail.
- If your reconciliation lives in a manual spreadsheet a single analyst maintains, that is a segregation-of-duties and key-control finding waiting to happen — regardless of whether the number is right.
The rest of this guide covers what auditors actually ask for, the transaction-level evidence chain for Tmall, JD, and Douyin, the control failures that trigger findings, and an audit-readiness checklist you can run before year-end.
Why China marketplace revenue is uniquely hard to audit
A Western auditor testing US or EU ecommerce revenue has a comfortable path: an order system, a payment processor, and a bank statement that mostly agree. China marketplaces break that comfort in three specific ways.
1. The settlement statement is the primary record — and it is not in English
On Tmall, JD, and Douyin, the platform acts as escrow and settlement agent. The authoritative record of what you earned is the platform’s settlement statement (结算单), not the order list. Auditors need that statement, its mapping to the bank deposit, and a translation/field dictionary so they can read it. This is exactly the discipline behind settlement reconciliation: the deposit is net of dozens of deductions, and the audit trail has to explain every one.
2. Gross-to-net is a judgment, and judgments must be documented
Whether you recognize revenue gross or net depends on who controls the goods and who is the principal in the arrangement — a genuine accounting judgment under IFRS 15 and ASC 606. If you sell through a Chinese TP or distributor as principal vs. agent, the auditor will want your position paper, not just the number. Undocumented gross-vs-net treatment is one of the fastest routes to an audit adjustment.
3. Revenue, tax, and invoicing sit in separate systems
In China, the revenue event, the VAT liability, and the fapiao (发票) are often recorded in three different places — the marketplace, your ERP, and the Golden Tax / fapiao system. Auditors reconcile all three. If your VAT and fapiao reconciliation does not tie to recognized revenue, expect questions about both revenue completeness and tax exposure.
What external auditors actually request for China revenue
When a Big Four team scopes China marketplace revenue, the document request list is remarkably consistent. Prepare these before they ask:
- Settlement statements for each platform and entity across the period, with a field dictionary explaining every deduction.
- Bank statements showing settlement deposits, matched to the settlement records (proof of occurrence and cutoff).
- Reconciliation workpapers showing gross GMV bridged to net revenue and to the recognized ERP figure, with variances explained.
- Revenue recognition memo documenting the gross-vs-net and principal-vs-agent position for each channel.
- Refund and returns data with timing, to test that revenue reversals landed in the correct period.
- VAT/fapiao reconciliation tying output VAT and issued fapiao to recognized revenue.
- Journal-entry detail with a clear line from reconciliation output to the posted ERP entries and consolidation.
- Control evidence — who prepares, who reviews, and how the reconciliation is signed off each month.
Notice the pattern: almost none of it is the revenue figure itself. It is all traceability and control around that figure.
The transaction-level evidence chain, from settlement to journal
An audit-ready China revenue number can be re-performed by an independent tester at transaction level. That means the following chain has to hold for any sampled order:
- Order & settlement — the order appears in the platform export and on the settlement statement with a settlement (not order) date.
- Deduction decomposition — commission, service fee, promotion funding, ad spend, and refunds are each identified and classified, so gross GMV reconciles to net proceeds.
- Bank match — the net settlement batch matches a dated bank deposit, proving the revenue occurred.
- Revenue recognition — the gross or net amount is booked per the documented policy, with returns provisioned per returns reconciliation.
- Tax layer — output VAT and fapiao tie to the recognized figure via fapiao reconciliation.
- ERP & consolidation — the entry posts to the correct entity and rolls into the group P&L at the right FX rate.
If any link is manual, undocumented, or lives only in one analyst’s head, that is precisely where the auditor will drill — and where a control deficiency gets written up.
The control failures that trigger audit findings
These are the deficiencies that turn a clean revenue balance into a written-up finding, drawn from how brands actually get caught.
Spreadsheet-only reconciliation with no review
A single analyst maintaining the China reconciliation in an unversioned spreadsheet is a classic ICFR (internal control over financial reporting) weakness: no segregation of duties, no evidence of review, no audit trail of changes. Even if the number is perfect, the control around it is not. Poor data quality feeding the ERP makes this worse, because the input itself is unreliable.
Cutoff errors around refunds and settlement timing
Revenue recognized on order date but settled — or refunded — weeks later creates period-end cutoff risk. Auditors specifically test transactions around the month-end close boundary. Refunds that cross into the next period, or a Singles’ Day sale settling in December, are where cutoff findings cluster.
Unexplained gross-to-net variance
If you cannot explain why RMB 10m of GMV became RMB 7.2m of net revenue — line by line — the auditor cannot conclude the number is complete and accurate. An unexplained bridge is treated as a potential misstatement until proven otherwise.
Revenue without a matching tax and invoice trail
Recognized revenue that does not reconcile to output VAT and issued fapiao raises two flags at once: possible revenue overstatement and possible unrecorded tax liability. In China this pairing gets scrutiny from both the financial-statement auditor and, separately, the tax authority.
How to build controls auditors trust
Passing the audit is not about heroics at year-end; it is about designing the reconciliation so evidence is a byproduct of normal operation. The COSO internal control framework that most SOX programs are built on gives the shape:
- Automate the join. Pull settlement, order, commission, ad, and refund data through the API and reconcile systematically, so every match is logged and repeatable rather than hand-keyed.
- Version everything. Store settlement statements, reconciliation outputs, and journal support with immutable timestamps, so an auditor sees exactly what existed at close.
- Separate prepare and review. One person (or system) produces the reconciliation; a different, documented reviewer signs off — the segregation of duties ICFR requires.
- Make the bridge explicit. Every gross-to-net reconciliation should show each deduction category and a variance threshold that forces investigation.
- Keep a live policy memo. Document gross-vs-net and principal-vs-agent positions per channel and update it when platforms change fee structures.
This is where a purpose-built integration layer earns its keep: when the settlement-to-ERP pipeline is systematic rather than manual, the evidence trail is generated automatically, and audit prep collapses from weeks of spreadsheet archaeology to a query. It is the same reason brittle marketplace-to-ERP integrations fail audits — the data path cannot be re-performed.
Your China revenue audit-readiness checklist
Run this before the auditors arrive — ideally before every quarter-end, not just year-end:
- Can you produce the settlement statement and matching bank deposit for any sampled order in minutes?
- Does gross GMV bridge to net revenue with every deduction category explained?
- Is your gross-vs-net and principal-vs-agent position documented per channel and current?
- Do refunds and returns land in the correct period, with cutoff evidence?
- Does recognized revenue reconcile to output VAT and issued fapiao?
- Is there documented preparer/reviewer sign-off on every monthly reconciliation?
- Can an independent person re-perform the reconciliation from source without the original analyst?
- Do the ERP entries and consolidation tie to the reconciliation at the right FX rate?
Any “no” is a gap to close now — each one is a question an auditor is trained to ask.
How Digate fits
Digate connects Tmall, JD, Douyin, Pinduoduo, and WeChat directly to your Western ERP — NetSuite, SAP, Oracle Fusion, or Dynamics 365 — and reconciles settlement to revenue systematically, keeping the full transaction-level trail from platform statement to posted journal. That means the evidence auditors ask for is generated as a byproduct of the close, not reconstructed under deadline. Gross-to-net bridges are explicit, deductions are classified, refunds are tracked to period, and every figure traces back to source. Instead of a spreadsheet only one analyst understands, you get an auditable, repeatable path from settlement to group P&L — the difference between a China number you hope is right and one you can prove is.
Frequently asked questions
What evidence do auditors need for Tmall or JD revenue?
At minimum: the platform settlement statement, the matching bank deposit, a reconciliation bridging gross GMV to net recognized revenue, a documented gross-vs-net policy, refund/returns data with timing, a VAT/fapiao reconciliation, and journal-entry support showing the path into the ERP and consolidation. Auditors test whether they can independently re-derive the number from these sources.
Why is China marketplace revenue a SOX or ICFR risk?
Because the reconciliation is often manual, undocumented, and dependent on one person — which fails the segregation-of-duties and review requirements of internal control over financial reporting. The revenue can be accurate and still generate a control deficiency if the process around it cannot be evidenced or re-performed.
How do auditors test revenue cutoff on China marketplaces?
They sample transactions around period-end and check that revenue was recognized in the correct period based on control transfer, and that refunds settling later were provisioned correctly. Settlement and refund timing that crosses month-end is the highest-risk area — see China ecommerce month-end close for how to control it.
Does the settlement statement or the order report govern for audit?
The settlement statement, reconciled to the bank deposit, is the authoritative record of earned revenue, because it reflects the platform’s actual deductions and payout. The order report shows demand, not settled economics. Auditors anchor on settlement and treat unmatched orders as a completeness or occurrence question.
How far back should we keep China settlement evidence?
Retain settlement statements, reconciliations, and journal support for at least the statutory audit and tax retention period applicable to your entities — in practice several years. Because platforms do not retain granular statements indefinitely, exporting and archiving settlement data at close is itself an audit-readiness control.