Factory-direct cuts the middleman margin — and quietly changes who controls the goods, how precise your contract is, and who you can recover from when the deal fails. A comparison from the enforcement side of the desk.
The first contract decision in any China sourcing project happens before anyone drafts a contract: who do you sign with? The trading company your industry contact recommends, or the factory you found yourself, with the promise that going direct cuts out the middleman's margin.
I won't argue about the margin. In most product categories, an intermediary takes a real spread — and buying direct from the production line is genuinely cheaper on a per-unit basis. The mistake is treating that as the whole analysis.
My office sits on the other end of these transactions. We get called when the shipment is late, when the quality claim is denied, when the deposit has vanished. And from where we sit, a trading company and a factory are not interchangeable counterparties — they are two different risk profiles wearing the same purchase order. Here is how I actually compare them, dimension by dimension, for clients deciding where to put their signature.
The factory-direct pitch sounds like pure arbitrage: same goods, lower price, no intermediary. What it actually does is move you one layer deeper into the Chinese corporate system — usually to an entity that is smaller, less internationally experienced, less regulated by its own customers' compliance departments, and much less used to being sued.
None of that is automatically bad. Some of the best suppliers I have worked with are single-entity factories run by engineers who have never lost a customer. But the savings and the risk come from the same place: you are deleting a layer whose full-time job was — among other things — absorbing commercial friction. A competent trading company issues clean documents, carries the quality claim on its own balance sheet before you have to fight the production line, and has enough asset diversity to be worth suing. Delete that layer and every one of those functions becomes your problem, governed by a contract the factory's office manager downloaded.
So the question is not "which is cheaper." It is: which entity, on these six dimensions, is the counterparty you want to be locked in a dispute with for the next two years? Because that is what a contract counterparty is.
Start where every verification starts: the business license. But read it like a credit analyst, not a tourist.
Chinese companies quote a subscribed registered capital — a number the shareholders promised to contribute on a schedule — which can be wildly different from paid-in capital, the money actually in the company. A factory with a subscribed figure of tens of millions of RMB and minimal paid-in capital is, financially, a much thinner entity than its license suggests. If that entity defaults, your judgment can only reach what it actually owns.
Scale matters for a second reason: operational depth equals dispute resilience. A mid-sized trading company with a real compliance function can produce a contract amendment in a day, knows what an original bill of lading is, and has dealt with a surveyor's report before. A small factory may produce excellent goods and then go completely silent for three weeks the first time you raise a formal quality claim — not out of bad faith, but because nobody in the building has ever handled one. In my experience, the deals that collapse into litigation are rarely the ones with bad products. They are the ones where a commercial hiccup met an organization that had no idea how to respond to it.
This is all checkable before you send a cent. If you have never pulled a Chinese corporate record yourself, start with the verification method I use in practice — it takes less than an hour and changes how you read everything else in this article.
The next dimension is the one almost no buyer checks: the entity's history as a defendant.
Chinese court records and enforcement lists are publicly searchable. Before signing, I want to know: Has this entity been sued before? For what — sales contract disputes, quality claims, labor, loans? How many times? Has it ever lost, and did it comply with the judgment, or did it end up on the list of judgment debtors? Is it currently subject to enforcement proceedings?
Reading that record is interpretive work. A factory with two historical quality disputes that settled is normal. A trading company that appears as a defendant in a dozen sales disputes across three years, with several judgments transferred to enforcement, is not "experienced" — it is a habitual defendant, and the pattern tells you the account manager who charms you on WeChat is not the person who decides what happens when a claim arrives. Conversely, an entirely clean record at a very small entity means only that nobody has bothered to sue it yet; it is not a credit rating.
One more pattern worth naming: when a supplier refuses to identify itself — will not produce a business license, won't tell you the legal name behind its factory gate or its payment account — that is no longer a diligence gap, it is a red flag of its own. I have had matters where the supplier's refusal to identify itself led to a police-verified identity check before we proceeded. Legitimate suppliers are proud of their licenses. They hang them on the wall.
This is the dimension buyers never think about until a dispute is live: in a Chinese trade, control of the physical goods and control of the transport documents can sit with different people — and neither of them has to be you.
Sign with a trading company, and the factory ships under the trading company's instructions; the bill of lading is typically consigned per the trading company's routing, freight is booked through its forwarder, and the export declaration names its customs registration. If a dispute erupts mid-shipment, the person who can divert the cargo, hold it at the port, or instruct telex release is your counterparty — which cuts both ways: a counterparty that controls the cargo has leverage, but a counterparty whose name is on the documents is also the entity your judgment can attach.
Sign factory-direct, and the goods sit on the factory's own floor until dispatch. For a claimant, that is actually attractive: physical assets in a fixed location are exactly what a court-ordered property preservation (asset freeze) can reach, and a factory's machines, inventory, and land-use rights tend to be harder to move than a trading company's bank balance. The offsetting risk: factories with no export history often hand the whole shipping file to a forwarder and never look at it again, and document discipline — originals, endorsements, clean receipts — is where disputes metastasize.
Whichever you choose, the contract should state who issues the documents, who consigns them, and who may vary that — before the first deposit moves.
Here is a generalization I will defend after years of reviewing both: factory contracts are drafted by people who do not expect to be sued, and it shows on every page.
A typical factory proforma — and I have reviewed hundreds — is a quotation sheet wearing a contract's clothes: product specs, unit price, delivery window, and almost nothing else. No governing law. No dispute resolution clause, or a one-liner that names no institution. Quality expressed as "good quality" or a reference to a sample nobody preserved. No agreed inspection standard or surveyor appointment mechanism. No liquidated damages for late delivery, no cap or floor on remedies, no language about documents, no force majeure definition beyond the phrase itself. The trading company versions are not poetry either, but they have usually been through disputes before — you can see it in the clauses that exist precisely because someone once lost money without them.
Recently I ran the full contract chain review for a Chinese state-owned trading group's first bulk-commodity import — 26,500 tonnes of mineral sands from an Australian mining supplier. The seller-side draft arrived with real structure, and it still took 55 annotated comments to close the gaps that mattered. If that is the gap count on a professional counterparty's paper, extrapolate to what a two-page factory proforma is hiding. The contract review service exists for exactly this reason — and if you want a sense of how deep that review goes, that mineral sands matter is written up in full.
Finally, run the worst-case tape. The contract failed, you have a judgment. Who is the debtor, and what can be enforced against?
Neither profile dominates. What matters is that you chose it deliberately — and structured the contract for it: a factory counterparty rewards early preservation strategy and tight performance obligations; a trading-company counterparty rewards corporate-structure diligence and payment-security terms. What punishes buyers is discovering, post-judgment, that they signed with the emptier of the two entities without ever comparing.
Compressed into the table I would actually use in a client meeting:
| Dimension | Trading company | Factory (direct) |
|---|---|---|
| Registered capital & scale | Often larger, diversified; still check paid-in vs subscribed | Often smaller, single-site; thin paid-in capital common |
| Litigation record | More entries but usually professionalized responses; check for habitual-defendant patterns | Fewer entries — may mean clean, or just never sued; watch for enforcement listings |
| Control of goods & documents | Controls routing, B/L, customs name; leverage and attachment target both | Goods on its own floor — good preservation target; weak document discipline |
| Contract precision | Dispute-tested clauses; negotiable but structured | Proforma quotations; gaps on governing law, quality standards, remedies |
| Recovery profile | Liquid assets, but possible shell structure | Tangible assets, but encumbered and owner-replaceable |
| Best when | Complex documents, multi-supplier consolidation, first orders, high-value goods | Stable specs, repeat production, close technical supervision, long-term relationship |
| Be careful when | Newly registered, refused verification, account ≠ contract name | High-value first order, complex documentation, no export history |
One honest closing note: plenty of buyers run fine for years with factory counterparties on thin paper, and plenty get burned by polished trading companies. The matrix raises your base rate; it does not replace verification of the specific entity in front of you. When the stakes justify it, that verification — corporate records, litigation history, beneficial owner, site reality — is a service we run end-to-end, and the working method is described on the supplier verification page.
This article is general information, not legal advice, and does not create an attorney–client relationship. Outcomes vary by case; nothing here is a guarantee of results.
Whether it is a factory or a trading company, the verification is the same hour of work: corporate record, litigation history, account match, site reality. We run it before the deposit moves — not after it disappears.
Verify my supplier