ERP Checklist for Foreign Companies Operating in China

How to use this checklist

A foreign-invested enterprise in China carries requirements that a domestic-only business does not: it must satisfy Chinese tax and statutory rules and, at the same time, report into a foreign parent under a different accounting framework and a data-transfer regime. That dual obligation is what makes ERP selection in China harder than it looks. This checklist gathers the seven areas that decide whether a system will work, each with the questions that expose the difference between a real capability and a marketing claim. Use it before you shortlist, and again before you sign. No single system wins on every line; the point is to know which trade-offs you are accepting.

1. Invoicing and the fapiao lifecycle

The invoice is the most common point of failure. Verify that issuance, verification, reconciliation, and archive are handled as an integration, not a manual step.

• Does the system issue VAT special invoices through the tax infrastructure without re-keying?
• Are inbound fapiao verified against the counterparty's record before payment?
• Is there an automatic monthly reconciliation between ledger VAT and the invoice pool?
• Does it support fully digitalized e-fapiao, and is the signed digital file archived?

For the detail behind these, see our fapiao guide.

2. Golden Tax and tax filing

Confirm the pipeline, not the promise. The test is whether a sale can travel from order to filed declaration without leaving the system.

• Is the tax interface native, middleware, or manual export-import?
• Who maintains the connector when formats change, and with what response time?
• Can the system produce declaration-ready data and import filing status back?

See our Golden Tax integration comparison for how to tell the three approaches apart.

3. Data residency and cross-border transfer

Decide the data architecture deliberately. The parent needs consolidated numbers; the regulator constrains bulk movement of personal data.

• Can the system be deployed in a China region, and does it support a domestic database?
• Can it classify data and minimize what crosses the border?
• Does it log transfers and support the documentation a filing requires?
• What will the foreign parent actually be able to see, at what level of detail?

See our data residency guide.

4. Payroll and social insurance

Payroll is a compliance process in China, not just a calculation. Confirm the components and the upkeep.

• Does it handle social insurance and the housing fund on locally defined bases?
• Does it apply the cumulative individual income tax withholding method correctly?
• Who maintains city-specific bases and rules, and how are changes delivered?

See our China payroll guide.

5. Statutory reporting

The reporting calendar draws on the same books. Require that the outputs come from the system, not from spreadsheets.

• Can it produce statutory financial statements in the required format from the books?
• Does it support the annual corporate income tax settlement with stored reconciliations?
• Are tax, statutory, and invoiced figures consistent from one source of truth?

See our statutory reporting guide.

6. Consolidation and the statutory-to-group bridge

The parent consolidates under a different framework. The bridge between Chinese statutory profit and group profit should be a report, not a year-end project.

• Can the system carry a statutory view and a group view from the same transactions?
• Does it produce the reconciliation bridge from stored data?
• How are exchange rates stored and applied for consolidation?

See our HQ consolidation guide.

7. Transfer pricing and related parties

If the China entity transacts with affiliates, ensure the data is complete, consistent, and traceable.

• Are related parties tagged so their transactions can be extracted automatically?
• Can the system report related-party transactions by type and counterparty?
• Does each disclosed figure trace back to a source document?

See our intercompany transfer pricing guide.

The decision underneath the checklist

Once you have the answers, one strategic choice remains: global template or China-native system, or a combination. The pattern that works for many foreign-invested enterprises is a global system for group reporting and consolidation, with a China-native layer for statutory compliance, invoicing, and tax — connected by an owned reconciliation. The pattern that fails is a global template bent to fit China through one-off customizations nobody maintains, or a local system with no path to group consolidation. Whichever you choose, the seven areas above are the ones to verify with a live demonstration rather than a slide deck, because every claim in them can be tested. For the export side of the same compliance stack, see our export tax rebate ERP guide.

• Invoicing: integration, verification, reconciliation, e-fapiao archive.
• Tax: native pipeline, maintained connector, declaration data in and status back.
• Data: China-region deployment, classification, minimized transfer, transfer logging.
• Payroll: social insurance, housing fund, cumulative IIT, maintained local rules.
• Statutory reporting: statements from the books, annual settlement supported, cross-filing consistency.
• Consolidation: dual views from one transaction set, stored bridge, stored rates.
• Transfer pricing: tagged related parties, reportable flows, traceable figures.

How to score vendors against this checklist

A checklist is only useful if it changes a decision, so give it scoring weight rather than treating it as background reading. The simplest approach that works is to score each of the seven areas on two axes: capability and evidence. Capability asks whether the function exists; evidence asks whether the vendor can demonstrate it with your data. A vendor that scores high on capability and low on evidence is selling a claim, and should be marked down accordingly. In practice, the areas where vendors converge on capability — basic tax and payroll — matter less for differentiation than the areas where they diverge, which are usually invoicing integration, data architecture, and the statutory-to-group bridge.

Weight the seven areas to reflect your own business. A China entity that reports into a foreign parent should weight consolidation and data residency heavily. A China-only trading company should weight invoicing and tax. A manufacturer with affiliates should weight transfer pricing and payroll. There is no universal weighting, and applying a generic one is how organisations end up choosing a system that is excellent at the wrong things.

A realistic implementation timeline

China ERP projects fail on sequencing more often than on software. A timeline that respects the dependencies looks roughly like this, and each phase should finish before the next is relied upon.

• Phase one — discovery and gap assessment: document the actual invoicing, tax, data, and reporting flows, and list the gaps with owners.
• Phase two — core configuration: the ledger, the chart of accounts, and the statutory framework, validated with a trial close.
• Phase three — compliance integration: invoicing, tax, and reporting connected, tested in parallel against the existing process.
• Phase four — consolidation and cut-over: the statutory-to-group bridge running, the manual routes retired, and the controls documented.

The temptation is to compress phases two and three and move quickly to cut-over. Resist it. The compliance integration is where the risk concentrates, and it is far cheaper to spend an extra few weeks proving the pipeline than to discover a gap in the first month of live operation, when the manual process has already been dismantled.

Where the budget actually goes

Teams comparing a global system against a China-native one frequently compare licence fees and conclude wrongly. The China-specific costs sit outside the licence. When you build the business case, include the connector build, the recurring maintenance tied to tax-format changes, the internal finance time to own reconciliation, and the cost of the parallel period during implementation. For a global system, these lines are often larger than expected because the localization is a project; for a China-native system, some of them are absorbed into the product. A comparison that leaves them out will favour the global option for the wrong reason, and the error tends to surface a year or two later in the form of unexpected maintenance and manual effort.

The five questions that expose a gap

If you take nothing else from this checklist, take these five questions. They are fast to ask and hard to answer vaguely, and each one targets an area where marketing claims most often outrun reality.

• Show me a sale travelling from order to issued fapiao without leaving the system — and then show me a red-letter invoice.
• Where does my data physically sit, and what exactly reaches our overseas parent?
• When the tax authority changes a format, who fixes the connector, how fast, and at whose cost?
• Show me the statutory-to-group bridge as a repeatable report, not a year-end exercise.
• Who owns the China localization after go-live, and what happens when that person is unavailable?

Any vendor with a genuinely China-mature product answers these in specifics. Any vendor without one answers them in adjectives. That difference, more than any feature matrix, is what this checklist exists to surface.

A quick readiness self-assessment

Before you evaluate any vendor, it is worth assessing your own organisation, because the system can only be as good as the process it replaces. Score each statement honestly; the areas where you answer no are where the implementation effort — rather than the software choice — will determine the outcome.

• Customer and vendor master data is accurate and regularly verified, including tax registration details.
• There is a documented owner for China compliance, not an informal one.
• The statutory close runs to a schedule, with reconciliation between ledger VAT and the invoice pool performed each period.
• The differences between the local and group frameworks are known and listed, not discovered at year-end.
• Related parties are identified and their transactions can be extracted without manual reconstruction.

A 'no' is not a reason to postpone the project; it is a reason to include a process workstream alongside the system workstream. Teams that buy software to fix a process problem usually end up with both. Teams that fix the process and choose software that supports it tend to have a quiet implementation, which is the best outcome available.

If you are already live and worried

Many readers will not be selecting a system at all — they are running one and suspect the China localization is thinner than it should be. The same seven areas work as a diagnostic, and the fastest way to use them is to test the highest-risk flow first: invoicing. Ask for a red-letter invoice to be processed end to end, and for the monthly reconciliation between ledger VAT and the invoice pool. If those two are automatic and auditable, the foundation is probably sound and the remaining gaps are fixable within the existing platform. If they are manual, the localization's limits are structural rather than cosmetic, and the honest options are to invest in the connector or to add a complementary local system — but the diagnosis comes first, and it takes an afternoon rather than a project.

Whichever situation you are in, the underlying principle is the same. In China, compliance is not a module you switch on at the end; it is a set of flows that either run through the system or run alongside it in spreadsheets. Every question in this checklist is really asking which of those two worlds you are in, and the answer determines both the cost of running the business and the exposure you carry. That is why it is worth asking the questions properly, before and after you buy.