The four verification gates, one by one

Each gate answers one question, and the questions are not interchangeable. Registration asks whether the entity exists. Main business asks whether that entity makes the thing you are buying. Export records ask whether it has actually shipped. The interview asks what it can produce on a schedule you can plan around.

A factory can pass the first and fail the second, and that is the most common pattern in Chinese sourcing: a perfectly real company that is not a maker. The order matters for cost rather than for rigour — registration is a public lookup and takes minutes, so running it first removes the wrong candidates before anyone books a call. The interview runs last because it is the only gate that needs the factory's own time, and spending that on a company which has already failed an earlier gate wastes goodwill worth keeping.

Gate 1 — Business registration

What we match: the legal name the factory trades under, its unified social credit code, whether the registration is still active, the year it was established, and the insured employee count. That last figure earns its place because it is the hardest number on the licence to inflate — a payroll record is a weaker thing to fake than a claimed capacity.

What the gate is for: removing entities that do not exist as described. Shell companies, deregistered businesses and trading desks presenting themselves as makers all end here, and they end cheaply, before any call is booked and before the factory is told we are looking.

What disqualifies a record: the quoted name maps to no registered entity, or maps to one that is deregistered, revoked or flagged as operating abnormally. A near-match is not a match. If the invoice would be issued by a different company from the one we verified, the record does not go up, because the buyer would be contracting with an entity nobody checked.

Gate 2 — Main business evidence

A registered company is not necessarily a maker of the thing you are buying. This gate asks whether fishing tackle is what the factory actually does, using three sources that have to agree with each other: the registered business scope, the product lines the factory publishes itself, and the certifications it holds.

The failure it catches is specific and common. A company with a broad trading scope can hold a valid licence, pass gate 1, and still buy every item it sells from someone else. That is not fraud — plenty of legitimate exporters work this way — but it changes who carries the quality risk on your order, and it changes what a factory visit would actually show you.

What disqualifies a record: a registered scope that does not cover fishing tackle; product lines where tackle appears as a side category among unrelated goods; certifications naming a different legal entity from the one quoted. Where the picture is mixed, the record states the split rather than resolving it in the factory's favour.

Gate 3 — Customs export records

Customs records answer a question a factory cannot credibly answer about itself: has this entity actually shipped, and to where. We read two things — destination markets and shipment frequency — because together they separate a factory that exports from one that says it can.

This gate also produces the most useful negative signal on this page. A factory with no export history is not automatically unsuitable: it may be a capable maker that has only sold domestically, and for a buyer with no export requirement that is fine. But it means the first order carries risks a shipping history would already have retired, and the buyer should price that in.

What disqualifies a record: no export history under the verified entity, or declared markets that contradict the shipping record. Where history exists but is thin, the record is published with that stated plainly instead of smoothed over. A short record is more useful to a buyer than a rounded one.

Gate 4 — Direct interview

Everything above can be assembled from documents. The interview is the only gate that cannot be copied from a website, which is why it is the one we will not skip. It is a structured call rather than a conversation, covering five things: monthly capacity in units, MOQ per colourway or size, lead time for a repeat order and for a first order on new tooling, the price basis (FOB port or EXW factory), and the material actually used.

Answers have to arrive as numbers. "It depends" is a legitimate answer to some questions and a failing answer to these. A factory that cannot state its own capacity does not have a capacity you can plan a product launch around. Samples can be requested at this stage, and the approved sample is what the order is later inspected against — which is why the sample matters more than the quotation.

What disqualifies a record: capacity, MOQ or lead time that cannot be stated as numbers; a price basis that will not be put in writing; refusal of the interview itself. A refusal is treated exactly as a failure, and the reason is recorded internally rather than published.

Why a failed record is absent, not pending

A pending record is still a recommendation. Publishing a factory as "under review" tells a buyer it is worth a look, and that is a claim we have not earned the right to make. Absence is the honest state for a record that has not passed, and it is the only state that cannot be misread.

It is also a decision the build makes rather than a status anyone maintains by hand. A factory that fails any gate does not appear, and the number of withheld records is published on the directory page, so the gap is counted rather than hidden. Records are re-checked on a 90-day cycle, so a record that has gone stale is visible as stale instead of quietly left standing.

What the gates cost, and how long they take

Verification runs per factory, not per product, and the cost is our time. The factory pays nothing, which matters more than it sounds: a process that charges the supplier is a process the supplier can shop around for.

End to end, a factory that passes takes roughly one working week. The time is not spent on paperwork: gates 1 and 3 wait on registers and customs data we do not control, and gate 4 waits on a call that has to fit the factory's own schedule across a time difference.

The cost sits in gate 4. A public lookup and a records check are minutes of work; a structured interview that produces numbers worth publishing takes an hour of prepared questions and a follow-up pass to confirm nothing was rounded. That is why this directory publishes fewer factories than the platforms list — and why the gap between the two numbers is a feature of the method rather than a shortfall in it.

Where this sits in the process

This page is one topic out of the verification standard. The overview is How we verify a Chinese supplier, and the other topics on this page are listed there.