Retail Vendor Compliance for China Importers (2026): Routing Guides, ASNs and the Chargebacks That Eat Your Margin

The container clears customs, the pallets go into the DC, the PO closes, and six to twelve weeks later a debit memo lands against an invoice you had already counted as collected. No phone call. No defect claim. Just a compliance code and a number.

This is the least discussed margin leak in importing from China, because it happens after everyone has stopped paying attention. The factory shipped what you ordered. The forwarder delivered what it was paid to deliver. The product is fine. And you still lost money.

The cause is almost never the product. It is the gap between how a Chinese factory builds an export carton and how a US retail distribution center receives one — two different engineering problems, solved by two different people who never speak to each other.

1. The routing guide is the real contract

Most importers read the vendor agreement once, sign it, and then work from the purchase order forever after. That is backwards.

The PO governs quantity, price and delivery date. The vendor agreement — which incorporates a routing guide and a compliance manual by reference — governs how the goods must physically arrive: which carrier, which label, which pallet, which data transmission, which appointment window. Retailers typically reserve the right to update that guide with limited notice, and the version in force on your ship date is the one you are measured against.

This matters disproportionately when your supply chain is in China, for three reasons.

• The spec has to travel further. A domestic vendor gets a routing guide change and calls its own warehouse. You have to transmit it to a factory that may treat it as paperwork rather than a production requirement.
• The factory optimises for export, not for receiving. Its carton is designed for container fill, strength and its own pallet stack — goals that frequently conflict with the DC's.
• Nobody reconciles the two documents. The PO says 24 pieces per carton, the factory's export carton holds 20, and someone in the middle decides 20 is close enough.

The structural problem is that a chargeback is unilateral. The retailer debits your remittance, sends a code, and puts the burden on you to dispute inside a window. There is no negotiation at the moment it happens, which is why the only effective control point is upstream.

2. Where the money actually goes

Chargeback schedules differ by retailer. The taxonomy does not — a handful of failure types account for most of the money.

Two things stand out. Only one row is about the product itself — everything else is packaging, data and logistics, decisions that are cheap to control at the factory and impossible to control at the dock. And several rows are data problems rather than physical ones: a trailer can arrive perfectly built and still be charged back because the ASN said 48 pallets and the DC counted 47.

3. The three fields that cause most of it

PO number

Retailers use fixed-format PO numbers, and that exact string has to appear on the carton label, the packing list, the BOL and the ASN. A transposed digit does not read as a typo at the DC — it reads as freight with no home, which is a receiving exception and therefore a charge. The same applies to the short DC code inside the ship-to block: one wrong character sends a trailer to the wrong state, and the freight, redelivery and missed-window charges all attach to the vendor.

GTIN / UPC

This is the field most often quietly wrong, because it looks like a formality. A GTIN has to come from a GS1 company prefix that you license — you cannot generate one, and a factory cannot generate one for you. When a supplier offers to "use the same barcode as last time", what is on offer is either a barcode licensed to someone else or a recycled code; both fail at the scanner and, on marketplaces, suppress the listing. See our guide to sourcing for Walmart Marketplace.

SSCC

The serial shipping container code is an 18-digit number, one per carton, encoded in a GS1-128 label, and it must match what the ASN says is in that carton. Factories cannot invent these: the number is built from your GS1 prefix plus a serial reference, and uniqueness must hold across every shipment you will ever make.

You have two options: generate a label file per shipment — a PDF with the correct SSCC range, PO, item and quantity pre-printed — and send it to the factory to apply, or specify a compliant on-demand label printer at the factory and audit its output. What does not work is emailing a barcode specification and hoping.

4. Case pack: the most expensive number nobody checks

Case pack is not a packaging preference. It follows from the retailer's planogram, shelf depth and DC split logic. If the guide says 12 units per case and you ship 24, the DC must break every case before it can allocate to stores — labour that is charged back, and repeated non-compliance escalates from a fee to a rejected PO.

Here is why it goes wrong so often with Chinese suppliers. The factory already owns a die-line: its standard export carton holds a certain number of units, uses a certain board grade, stacks a certain way. Changing the case pack means a new carton size, which means a new die, a new pallet build and a different number of cartons per container — and that last consequence changes your freight, as set out in container loading and carton design. So the factory resists, the importer doesn't push, and the compromise lands on the DC floor.

5. Pallet geometry: two optimisations pulling apart

A container rewards volume; a distribution center rewards standardisation. The factory will build the pallet that suits the container unless told otherwise.

The standard US grocery pallet is 48 × 40 inches, and most retail DCs are built around it. Guides typically specify the footprint, a maximum overall height including the pallet, no overhang, a pallet grade, and a stacking configuration expressed as TI (cartons per tier) and HI (tiers per pallet). That TI/HI pair also appears on the ASN and the pallet label, so a mismatch between what was declared and what was built is a data failure on top of a physical one.

• Overhang is the classic failure. The carton count divides neatly into a 48 × 40 footprint, so the factory builds it — and the product overhangs by a couple of centimetres. Trailers and racking are unforgiving about that.
• Height is the second. Filling a 40HQ rewards tall stacks; a DC with a 60-inch receiving spec does not care what the container rewarded.
• Mixed SKUs are often prohibited outright, because a mixed pallet cannot be put away without being broken down. For small multi-SKU orders that usually argues for a stateside 3PL.
• Wood packaging: solid-wood pallets need the ISPM 15 treatment mark to enter the US, and many DCs reject damaged or unmatched pallets anyway.

6. Routing, appointments and the last mile inside the DC

Retail guides generally specify one of two arrangements. Under collect, the retailer's carrier picks up and you must submit a routing request through a portal before a cutoff, then use the assigned carrier. Under prepaid and add, you arrange delivery but still have to meet the DC's appointment and documentation requirements.

Both punish the same mistake: treating delivery as "the forwarder will handle it". A forwarder's door-delivery quote normally means a truck arrives at some point on some day, while a retail DC runs on booked appointments and will turn a trailer away — early or late — and charge for the privilege.

The China-specific risk is timing. Ocean transit has variance and the appointment must be booked against a date that will move, which is how a physically perfect shipment still generates fees: the vessel arrived two days late, nobody rebooked, the trailer missed its slot, and the charges are coded as a vendor compliance failure. Treat the arrival notice — not the sailing date — as the trigger for booking.

7. Push the spec upstream: nine clauses for the factory PO

Everything above is a purchase order problem wearing a logistics costume. These clauses convert a retail routing guide into instructions a Chinese factory can execute.

That last clause is the one that changes behaviour, because it moves the cost to the party who controls the decision. It only works if the requirement is specific enough to be auditable — you cannot back-charge for failing to meet "retail standards". For the wider set of commercial terms worth locking down before payment, see purchase contract terms before you pay.

8. The audit that happens before the cartons are sealed

Pre-shipment inspection usually checks the product. For a retail-bound order it must also check the shipment as a receivable object — while cartons can still be reprinted.

1. Scan five case labels and confirm each SSCC is unique and present in the ASN file.
2. Verify the GTIN scans and matches the item master, on the unit and on the case.
3. Count units in three cartons by hand against the declared case pack.
4. Measure one finished pallet: footprint, overall height, overhang on all four sides.
5. Confirm TI/HI matches the ASN declaration, not just the physical build.
6. Photograph the label panel of a sealed carton and the pallet from two sides.
7. Reconcile total cartons on the packing list against what is staged for loading, and confirm the ship-to and DC code character by character against the PO.

9. When the debit memo lands

Assume it will land at least once. What decides whether it costs you money is the file you can produce, and how fast. First, respect the dispute window — vendor agreements set one, usually in weeks from the memo date, and a missed window converts a disputable charge into a final one. Second, build the file from documents that are hard to argue with: the signed BOL with the appointment number, proof of delivery, the ASN transmission record plus its functional acknowledgment, label photographs from origin, the packing list and the loading photo set.

Then separate the causes properly. A carrier that delivered late against a confirmed appointment is a carrier claim, not a vendor failure. A DC counting 47 pallets against an ASN of 47 is a receiving error. A factory that printed the wrong SSCC is your error — pay it and fix the process, because disputing a charge you caused costs credibility you need for the ones you didn't.

Finally, track chargebacks by code for two consecutive quarters. The distribution is usually brutal: a handful of codes account for most of the money, and two of them often point at a single factory or forwarder. That is fixable — a vague sense that "chargebacks are a cost of doing business" is not.

The Bottom Line

A retail chargeback is not a penalty for a bad product. It is a penalty for an information gap between a routing guide in Arkansas and a packing line in Guangdong — and it is tempting to hand that problem to whoever manages freight. That is the wrong owner. Every failure in this guide is decided at the factory, in the sample and PO stage, by someone choosing a carton size, a label printer and a pallet pattern.

Put the case pack, the carton dimensions, the label file, the SSCC range and the pallet build into the purchase order. Make the obligation auditable. Back-charge the failures. Everything else in vendor compliance is detail around those five decisions — and all five are still free to make before the die-line exists.

— Jack Zhang, ChinaBizGuides