The first dependable weight in our Yiwu warehouse often appears late. The buyer has chosen the product and may already have paid for it. A freight budget exists too. The carton then reaches us. We put it on a scale, and the number that looked settled often changes.

Read also: The Lines in Your Quote That Were Never Real

I run a sourcing check service in Yiwu. On 1688, the Chinese wholesale marketplace where many cross-border orders begin, the product-detail data can carry a shipping block for each SKU. The platform also labels the source of the package-size data. It may be entered by the merchant or returned from a fulfilment-centre measurement.

We read 120 top-ranked listings through 1688’s own read-only API on August 12, 2026. The sample covered the first ten results for twelve English searches in bulky home-goods categories. All 120 returned a shipping block, but one returned no weight field.

Of the 120 listings, 117 carried at least one SKU row labelled as merchant-entered. A fulfilment-centre measurement appeared in three listings. Those counts overlap because one laundry-basket listing contained both labels across its SKU rows. They are not neat, mutually exclusive buckets.

The dimensions were less complete. Only 45 listings returned a positive value for all three package dimensions. Another 74 returned zero or partial dimensions, and one omitted the dimension field. Without length, width and height, a buyer cannot check volumetric weight from the listing data alone.

That is important because carriers do not simply bill the grams shown on a product page. Air and express services commonly compare actual weight with volumetric weight and charge the larger figure. For a dense product, the scale may decide the bill. For a light storage basket or a plush item, the space occupied by the carton may matter much more.

The most extreme page we opened was for a kitchen organiser. It grouped 25 variants together. One row said 23 grams. Another said 1.236 kilograms. The package dimensions were incomplete, so there was no way to compare those figures with a volumetric calculation. The spread was 53.7 times. A laundry-basket page ranged from 180 grams to 3 kilograms across 18 rows.

Across the sample, 108 listings returned weight data for more than one SKU row, although some simply repeated the same figure. The problem is the mapping. If a buyer copies one number without matching it to the ordered variant, the arithmetic can be precise and still be about the wrong product.

The timing is not academic. On August 13, 2026, the US Court of International Trade decided Axle of Dearborn v. Department of Commerce and found that the president could rescind the de minimis exemption under IEEPA. China-origin goods had already lost duty-free de minimis treatment in May 2025. An appeal could change the interim legal position. Congress has set a separate statutory repeal date of July 1, 2027.

Weight does not determine every duty bill. Many tariffs are charged as a percentage of customs value, while some tariff lines use a quantity such as kilograms. Gross weight is still a required line-item field in the US Automated Commercial Environment. It also drives the freight bill for air and express shipments. The same number can matter when a forwarder prices the shipment and when the importer has to stand behind the customs record.

None of this proves that a merchant-entered weight is inaccurate. We did not buy and weigh these 120 products. The study records provenance and field completeness, not truthfulness. A seller may have measured the product carefully and then typed the result into the field. A fulfilment-centre measurement can also become stale after packaging changes.

An estimate can easily look like evidence. Buyers sometimes carry the listing figure into a freight budget. A quote prepared before the goods exist in packed form may have little else to work with.

Later documents can inherit the same number. Our sample stops at the product-detail response and did not follow these listings through a real customs entry, so it cannot say how often that chain occurs. It does show why the first figure should not be assumed to have come from a scale.

The field was designed first for marketplace logistics in China, not for a US customs entry. Domestic freight templates need a weight to produce a shipping charge. That purpose is different from establishing the packed gross weight of an export carton. The figure may be useful for an early estimate, but its job changes when an importer starts treating it as final.

In Yiwu, the practical distinction comes before the calculation. We look at the goods and ask whether the cargo is dense or light and bulky. With storage products and plush goods, the listed grams tell only half the story. The missing carton dimensions are often the more expensive omission.

That is why our warehouse puts the received carton on a scale and photographs the reading. We measure the carton as well. This is not process theatre. If we price international freight from a number we never checked and the scale later differs by a kilogram, the shortfall is ours.

The measurement can happen earlier without turning the order into an audit. Before the balance payment, the supplier can identify the exact SKU and provide the packed carton weight with its dimensions. The unit count belongs in the same reply. A photograph of the sealed carton on a scale is more useful than another screenshot of the listing. Mark any pre-packing freight figure as an estimate. Replace it with the warehouse measurement before the international shipment is booked.

For the first order, keep the listing figure beside the supplier’s packed figure. Add the warehouse result when the goods arrive. The difference becomes a correction factor for the next order. On a multi-SKU purchase, preserve the mapping by variant rather than keeping one number for the whole listing.

The full method and row-level dataset are published at https://supplymo.com/research/1688-weight-provenance. It is a non-random, one-day sample of platform-ranked listings, not a census of 1688 and not a finding of seller misconduct. It cannot predict a duty rate or a freight price. Those depend on the product, its classification, the route and the terms of the quote.

I would not use the listing figure for settlement. I would use what the warehouse scale shows after packing.

Liam Cai is the founder of Supplymo and runs sourcing and quality control in Yiwu, China, for overseas buyers. https://supplymo.com