Two things get said about 1688 constantly, and both are out of date. The first is that a foreign buyer cannot use it. The second is that you need an agent, full stop, as though that settled anything. What is actually true is that 1688 now has two separate entrances, they behave like different platforms, and which one you go through determines your catalogue, your paper trail and who is legally responsible for getting the goods out of China.
Two entrances, and they are not the same platform
The cross-border channel is built for buyers outside mainland China. It presents a curated slice of the marketplace with an English interface, and — the part that matters most — it supports payment from outside China without a mainland bank account. 1688 documents the mechanics itself in its cross-border trade guide.
The main site is the domestic wholesale market. The full catalogue lives there, it is in Chinese throughout, and every mechanism behind it — payment, dispute handling, after-sales, the customer service channel — assumes a Chinese company with a Chinese bank account and a delivery address in mainland China.
The difference between the two is not convenience. It is catalogue size, and the filter is supplier willingness: a company appears in the cross-border channel because it decided it wanted overseas orders. That is the same selection effect that makes Alibaba what it is, applied one layer down — which is worth holding onto, because escaping that filter is usually the whole reason someone goes to 1688 in the first place. The comparison between the three channels sets out what each one selects for.
So the practical position is this. If what you want is generic and available through the cross-border channel, use it and most of the friction below disappears. If it is not — and for anything even slightly specialised it usually is not — you are buying from the domestic market, and the rest of this applies.
Reading a listing properly
Most of the money wasted on 1688 is wasted on the product page, before anyone has been contacted. Four things are routinely misread.
The price you can see is usually not your price. Domestic listings quote in bands, and the number shown most prominently is generally the best band — the one for the largest quantity. Next to it sits the minimum order. A listing showing a very low unit price at a thousand pieces and a much higher one at ten is not being deceptive; that is simply how the domestic market quotes. Read the band table, not the headline number.
The unit is the most expensive thing to get wrong. Listings are priced per piece, per carton, per set, per dozen or per pair, and these appear interchangeably across listings within the same category. A price that looks like a remarkable find is very often quoted per carton. This single misreading accounts for more wasted enquiries than anything else on the platform, and checking it costs nothing.
Stock status changes what you are agreeing to. Goods described as in stock ship from inventory. Made-to-order and pre-sale listings do not — the latter is a commitment against a production run that has not started. The lead times behind these are not comparable, and a listing that reads like stock can turn out to be a pre-sale.
Freight and the tax invoice are usually excluded. Domestic prices are commonly quoted ex-workshop, with domestic freight to your consolidation address on top. Separately, many domestic quotes exclude a VAT invoice. That matters more than it sounds: a VAT invoice is what an exporter needs in order to claim the export VAT refund, so asking for one after a price has been agreed is a renegotiation rather than a formality.
The signals on 1688 worth actually reading
1688 shows something Alibaba largely does not: a domestic transaction record. Recent sales volume and the repeat-purchase rate describe what Chinese buyers — who can read every word on the page, and who will complain in their own language — actually did. That is considerably harder to manufacture than a badge, and it is the most useful number on the listing.
It is not, however, a quality signal in the sense an overseas buyer wants. High volume with a high repeat rate tells you the product reliably meets the domestic market's expectations at the domestic market's price. Whether those expectations match your market's is a different question, and the answer is often no, for reasons covered further down.
For navigating without Chinese, the site's image search works better than translated keywords. A photograph returns visually similar listings across the catalogue, which sidesteps the fact that machine-translated category names on 1688 are frequently wrong in ways that are very hard to notice from the outside.
How the money actually moves
There are three routes, and they differ mainly in what trail they leave behind.
- Official cross-border payment. Through the cross-border channel, a company outside mainland China can pay without holding a local bank account. The money moves through the platform, which means there is a platform record of the order and a platform mechanism if something goes wrong.
- An agent or forwarder paying in local currency on your behalf. Normal practice, and there is nothing wrong with it. It does change who your counterparty is: the contract and the paper trail are now between you and the agent, not you and the workshop. That is a trade rather than a problem, as long as you know you made it.
- Paying an individual directly. Personal transfers for a company order leave essentially no usable trail — no invoice, no company name, nothing establishing what was bought from whom. This is the one route with no recovery path at all. What each payment method actually protects works through why.
There is a quieter cost to the domestic route that is easy to miss. On 1688 you often never establish the legal entity behind a storefront. Many are operated by trading companies or individual resellers, and the domestic market has no particular reason to make that legible to you. If your order needs a named, accountable counterparty — because of certification, because of liability, or simply because you intend to reorder — that absence is the real limitation, not the language.
Getting the goods out of China
This is the part 1688 does not do, and the part that catches people out.
A domestic seller sells domestically. They ship to a Chinese address and their obligation ends there. Exporting requires a Chinese entity holding import-export rights to make a customs declaration in its own name. Your supplier is frequently not that entity and has no interest in becoming one.
So somebody has to be the exporter of record. In practice that is an agent or a freight forwarder: goods from several 1688 sellers arrive at a consolidation warehouse, are checked and repacked, and leave on a single declaration. Whoever performs that role is also the party who needs the tax invoices mentioned earlier — which is why the invoice question belongs in the price negotiation and not after it.
The consequence for your own arithmetic is that a 1688 unit price is not comparable to an Alibaba FOB price. The 1688 number is ex-workshop. To reach the line an FOB quote sits on, add domestic freight, consolidation and repacking, export packing where the domestic packing will not survive a sea voyage, the declaration itself, and the agent's fee. Which Incoterm you buy on then decides who handles everything beyond that point, and DDP or FOB is where that choice gets made.
Why the domestic price is lower, and what is missing from it
Part of the gap is genuine. A domestic listing carries no English-speaking sales staff, no export documentation overhead, no trade-show budget and no platform commission aimed at overseas buyers. That is real cost the domestic seller does not have and is not charging you for.
The rest of the gap is made of things that are absent rather than free. Export-grade packing, compliance testing for markets outside China, and the quality control layer an export-focused supplier runs by default are often missing from the domestic price because the domestic product never needed them.
Compliance is the one to take seriously. A product made for the Chinese domestic market may never have been tested against the rules that apply where you intend to sell it — not because anyone cut a corner, but because it was never intended to leave the country. That is a scope mismatch rather than a defect, and it lands entirely on the importer. If your category is regulated, treat testing as a line item on any 1688 order rather than hoping a certificate turns up. And when one does turn up, verify it with the issuing laboratory rather than reading the logo; which check stops which failure covers the rest of that ground.
When 1688 is the right tool, and when it is not
It is genuinely excellent for one category of order: an existing, generic, unregulated product, in a quantity too small to interest an export supplier, where the specification is simply whatever the seller already makes. Nothing beats it on price for that, and the friction is worth absorbing.
It is the wrong tool when the order needs an accountable counterparty. Custom products with tooling, where ownership of the mould has to be written down. Certification-dependent categories. Anything where you intend to hold a supplier to a specification across repeat orders, or where a defect classification has to be agreed before production starts. All of those need a named company that knows it is exporting and understands what it has signed up to. Running that through a domestic storefront is not cheaper — it is the same work, done later, by you.
If you would rather not run the export side yourself, that is the work our supplier sourcing and verification service takes on. To see what a 1688 listing costs delivered, the landed cost tool gives an estimate before you talk to anyone.
