The cheapest risk management in sourcing is a document check that costs less than a dinner. The most expensive mistake in sourcing is wiring five figures to a factory that looked real online. Between those two sentences sits a market of verification services whose 2026 pricing runs from $199 for a database-backed legal-entity pull to $2,500 for a premium on-site audit from a global firm — a range wide enough that the real question is not whether to verify but which tier your order size actually requires. This piece walks the seven checks that stand between a supplier profile and a wire transfer, with current pricing for each, so the verification budget scales with the exposure instead of being an afterthought.
Start with the pricing map, because it calibrates everything else. At the base sits document-based verification: a report pulled from China's National Enterprise Credit Information Publicity System — the GSXT — that confirms the legal entity exists, matches the factory's claimed name, and surfaces registration details, for around $199. The middle tier is the man-day inspection market: on-site services billed at $150 to $350 per man-day, where a trained inspector spends a day at the facility checking what the brochure claims. The top tier is the premium audit from the global compliance firms — SGS, Bureau Veritas — running to $2,500 or more for multi-day, multi-standard assessments. A useful rule of thumb follows from the spread: verification spend should run roughly half a percent to two percent of the wire value, which makes almost every tier decision self-evident once you commit to having a tier at all.
Check one is legal-entity verification, the $199 baseline that no order should skip. The GSXT pull confirms the registered legal name, the unified social credit code, registered capital, ownership structure, and — critically — whether the entity status is normal, revoked, or in abnormal-operation listings. This is where the classic sourcing scams die: the trading company posing as a factory, the shuttered operation still quoting from a website, the middleman with a name one character different from the real manufacturer. The check takes days, costs less than two hundred dollars, and catches the fraud patterns that a video call and a polished Alibaba storefront never will. If your order justifies a wire, it justifies this check; there is no order size below this threshold.
Check two is financial health. Registered capital tells you the commitment floor, but the signals that matter are operational: payment histories from other buyers, litigation records, credit reports from the verification bureaus that aggregate Chinese-court and administrative data. The patterns worth flagging are specific — a factory with active litigation from other suppliers, sudden registered-capital reductions mid-negotiation, or ownership transfers that coincide with your quote getting suspiciously better. The 2026 environment makes this check cheaper than it used to be: the document-verification services bundle credit-style summaries with the GSXT pull, so the marginal cost of looking at financial signals on top of legal existence is small. Treat unexpected financial opacity as an answer, not a gap.
Check three is the capacity audit, and its unit of account is the man-day. For $150 to $350 per man-day, an inspection firm puts someone inside the factory to verify what the sales pitch claims: production lines actually running, the stated workforce actually present, the machinery list matching reality, monthly output plausibly supporting your order timeline. The market has clear structure: QIMA anchors the tech-forward end at $299 per man-day, backed by a $250 million investment and a platform serving over 30,000 brands; TradeAider offers half-man-day inspections from $99, the cost-accessible entry for first orders that do not justify a full day; and the global compliance firms occupy the premium tier for complex manufacturing. A one-day audit covers the essentials — existence, capacity, conditions; a two-day audit adds production-process tracing and quality-system depth. Match the days to the order's risk, not to the calendar's convenience.
Check four is certification authenticity. ISO 9001, CE marking, UL listings, industry-specific credentials — these are the table stakes of supplier legitimacy, and certificate forgery is a known, persistent pattern in the category. The check is procedural rather than clever: collect the certificate numbers, verify each against the issuing body's public database, confirm the named scope covers your product category, and check the validity window. Forged and expired certificates cluster at the same suppliers that fail other checks, which is exactly why the seven-check ladder works — fraud is rarely sophisticated enough to pass a coordinated verification stack even when it defeats any single check.
Check five is intellectual-property protection, the check that only matters until the day it matters completely. The moving parts are three: the NDA with scope that covers your designs rather than generic language; mold and tooling ownership — who paid, who owns, where the tooling physically lives, and the contractual mechanism for retrieving it; and the golden sample, the sealed reference unit both parties sign, which becomes the binding quality standard for every subsequent shipment. Tooling ownership is the classic trap: buyers pay for molds, assume ownership, and discover at the relationship's end that the factory holds the physical asset and the contract is silent on return. The clause costs nothing in negotiation and everything in its absence.
Check six is the sample program as a system rather than an event. The pre-production sample is not a courtesy; it is the contract's reference exhibit. The discipline is in the sequence: golden sample sealed and countersigned before the production order confirms, pre-production sample checked against it, and tolerance bands written into the specification — dimensions, materials, finish, packaging — so that "close enough" never becomes the operative quality standard. When disputes arise months later, the question that decides them is what was signed, not what was said. A sample program with paper behind it converts the inspection standards of check seven from opinion into measurement.
Check seven is the inspection cadence, the ongoing rhythm that carries verification from pre-wire diligence into production reality. The standard structure for a serious order is three touches: a during-production check that catches problems while correction is still cheap, a pre-shipment inspection against the AQL sampling standard before the goods leave the factory, and container-loading supervision when the order value justifies it. The 2026 service standard includes fast scheduling and 48-hour report turnaround, with photo documentation and defect classification — capabilities that QIMA-style platforms have made the default. The tier structure maps to risk: the compliance-heavy firms fit complex, regulated manufacturing; the e-commerce-focused specialists — QIMA, V-Trust, Testcoo, and peers — fit pre-shipment and during-production checks for importer-scale orders. Pick the inspector for the failure mode you fear, not the logo you recognize.
For repeat orders, the ladder compresses into a lightweight cadence rather than a full re-run. The entity check refreshes annually unless ownership changed; certifications re-verify on their validity windows; the sample program persists as the standing reference; and inspection cadence scales with order value and the previous three shipments' scores — a supplier with three consecutive clean pre-shipment reports earns during-production checks on sampling, while a first defect resets the full three-touch rhythm. The supplier file accumulates this history, and the verification budget for a proven partner drops well below the half-percent floor. That is the compounding payoff of doing the ladder properly the first time: diligence paid once keeps discounting every order after it.
Stack the seven checks into a ladder and the logic of the verification budget emerges. A $15,000 first order justifies the $199 document check, a bundled credit summary, and possibly a half-day inspection — a few hundred dollars of diligence against a five-figure exposure. A $150,000 program order justifies the full stack: entity, financials, a two-day capacity audit, certificate verification, an IP package, a sealed sample program, and the three-touch inspection cadence — still under one percent of wire value. The teams that get this wrong do not usually skip verification out of arrogance; they skip it because nobody assigned the ladder, and each check individually seemed optional. The fix is making the ladder a precondition of the wire rather than a judgment call in the moment.
The discipline generalizes beyond fraud protection. Every check on the list — entity, financials, capacity, certifications, IP, samples, inspections — produces artifacts that outlive the transaction: a verified supplier record, a signed golden sample, inspection reports that become the baseline for the next order. Run the ladder on every new supplier and the record accumulates into the thing procurement teams actually run on: a supplier file where trust is documented rather than remembered. In a market where the difference between a factory and a facade is a $199 database pull, the seven-check ladder is not bureaucracy — it is the cheapest insurance in international sourcing, priced by the man-day and paid for the first time it catches what a website could not.
The seven checks share one design principle: every artifact they produce is reusable. The entity report, the credit summary, the audit certificate, the sealed golden sample, the inspection history — each attaches to the supplier record rather than the transaction, so the second order starts with a verified file instead of a blank page. Verification priced per wire is an expense; verification built as a library is an asset. The ladder, run consistently, is how a sourcing operation stops re-purchasing the same certainty every quarter.