Where the shipping weight on a 1688 listing comes from
Chargeable weight decides what a shipment costs, and it is computed from two inputs: a weight and a set of dimensions. We looked at what 120 top-ranked 1688 listings actually display for those two inputs, and where each figure says it came from. All 120 showed a shipping block. 117 of them carried a weight the merchant typed in themselves.
The one-line version, free to quote with attribution: 117 of 120 top-ranked 1688 listings source their shipping weight from the merchant's own entry, and only 3 carry a weight measured by a fulfilment centre.
Liam Cai · Founder, Supplymo
Published 15 August 2026 · Collected 2026-08-12 · Yiwu, China
n = 120 listings. Collected 2026-08-12. Non-random sample: 12 keyword searches, 10 platform-ranked rows each. All 120 listings displayed a shipping block.
117/120
Weight is merchant-entered
The platform labels the provenance of the figure. This is the label it showed.
3
Weights measured by a fulfilment centre
Out of 120. No listing carried a platform-estimated weight.
45/120
Complete dimensions shown
74 returned a dimensions field that was zero or partly filled.
108/120
Carry more than one weight
Different SKU rows on the same listing. Widest gap in the sample: 53.7 kg.
The number your freight quote is built on was typed in by hand
Volumetric weight is not an opinion. A forwarder takes length by width by height, divides by a volumetric divisor, compares the result to actual weight, and bills the larger of the two. That arithmetic is well understood and rarely disputed.
What gets less attention is where the two inputs come from. In this sample the weight on 117 of 120 listings was labelled as merchant-entered. Only 3 carried a weight that a fulfilment centre had measured. A merchant-entered weight is not automatically wrong. It is simply a figure that nobody outside the shop has checked, and it is the figure a buyer copies into a landed-cost estimate weeks before a carton exists.
The dimensions gap is wider than the weight gap
A weight at least existed on nearly every listing. Dimensions did not. Complete dimensions appeared on 45 of 120, and 74 returned a dimensions field that was present but zero or partly filled.
That matters more than it sounds, because for the categories we sampled the volumetric side usually wins. These are bulky, light home goods, the exact profile where a carton is charged on the space it occupies rather than what it weighs. When dimensions are missing, the volumetric half of the calculation cannot be done at all before the goods arrive at a warehouse, and the quote a buyer is working from is only half a quote.
One listing, several weights
108 of 120 listings carried more than one weight across their SKU rows. Often that is legitimate: a listing may run from a small size to a large one, and the weights should differ.
The practical problem is which figure travels. A buyer, an agent or an automated quoting tool reading such a listing picks one number, usually the first row or the cheapest variant. In this sample the widest gap between the lightest and heaviest SKU on a single listing was 53.7 kg. Pick the wrong end of that range and the freight estimate is not slightly off, it is a different order of magnitude.
What a buyer can do with this before paying
Ask which SKU the quoted weight belongs to, and ask for dimensions in writing when the listing does not show them. A supplier who ships the item regularly can answer both in minutes, and the answer is checkable later against the carton.
The wider point is about the shape of a quote, not about any individual seller. If a freight figure was derived from listing data, whoever is relying on it deserves to know which parts of it were measured and which parts were typed in.
What this study cannot tell you
Putting this before the method rather than after it is deliberate. A number without its boundary travels further than the boundary does.
- It does not say any displayed weight is wrong. Nothing was weighed. A merchant-entered weight can be exactly right, and a measured one can be stale.
- It does not say a platform or a seller has done anything wrong. Displaying a self-reported weight is normal practice across marketplaces.
- It is not a random sample. The rows come from the first page of 12 keyword searches, and the platform ranks that page. Highly ranked listings are, if anything, more likely to carry complete shipping data than an average listing.
- It covers bulky English-language home goods on the cross-border distribution channel, on one day. Dense, regulated or made-to-order categories may look different, and no trend is claimed from a single observation.
- The rows are listings, not distinct suppliers. Seller identity is deliberately not collected, so one shop may appear more than once.
How the sample was collected
- Source
- com.alibaba.fenxiao.crossborder:product.search.keywordQuery + product.search.queryProductDetail (read-only)
- Sampling frame
- 12 English keyword searches on the cross-border distribution channel, first results page only, platform-ranked rows. Keywords are bulky home goods, chosen because volumetric weight dominates their freight cost, so a wrong weight or missing dimensions is most expensive exactly here.
- Collection window
- Single day, 2026-08-12 in China
- Write operations
- None. No order, payment, enquiry, messaging or logistics endpoint was called.
The readings we tested before publishing
Each of these is a way the data could have been misread. We checked, and the check is published with its result, including where it came back inconclusive.
When dimensions are missing, is that an absent field or a value of zero we failed to distinguish?
Method. Record for every SKU row whether each of length, width and height was returned at all, separately from whether the returned value was positive. Publish the three-way split.
Result. Of 120 listings, 45 returned complete positive dimensions, 74 returned the fields with zero or partial values, and 1 did not return the fields at all.
Conclusion. The three states are reported separately. 'No dimensions shown' in this study always means the field itself was absent or incomplete, never a zero we misread.
Does the platform's AI weight-estimation machinery show up on these listings?
Method. The detail schema carries aiWeight and aiWeightAccuracy keys, so AI-estimated weights are counted separately from merchant-reported and measured ones wherever the platform labels them.
Result. Zero of 120 listings carried an AI-estimated weight label. The displayed weights are merchant self-reported on 117 listings and measured logistics returns on 3. The same pattern held in an earlier draw of the same frame.
Conclusion. The study reports provenance, not accuracy. A merchant-reported weight may be correct; the point is that nothing in the listing shows it was measured.
Is the weight-source label listing-level or copied shop-wide?
Method. Source labels vary across SKU rows within single listings and across listings under the same keyword.
Result. Distribution recorded per listing in the rows array for reader re-computation.
Conclusion. Treated as listing-level data as returned by the detail endpoint.
The stated limits, in full
All 8 of them travel inside the evidence file, so they stay attached to the data after it leaves this page.
- 1.This is not a random sample. Rows come from the first page of a keyword search, which the platform ranks. Highly ranked listings are, if anything, more likely to carry complete shipping data than an average listing.
- 2.This measures where the displayed weight and dimensions come from, not whether they are correct. No parcel was weighed. An AI-estimated weight can be accurate and a merchant-reported one can be wrong.
- 3.The source labels and the AI confidence figure are the platform's own. We did not audit how the platform estimates weights or computes confidence.
- 4.The keywords are bulky English-language home goods chosen because volumetric weight dominates their freight cost. Dense, regulated or made-to-order categories may show different provenance and are not represented.
- 5.Single day of collection. Platform fields can change, and no trend is claimed from one observation.
- 6.All figures describe the cross-border distribution channel of the API. Buyers browsing the domestic site directly may see a different presentation.
- 7.Nothing here says a platform or a seller has done anything wrong. A buyer quoting freight from listing data simply deserves to know which numbers were measured and which were modelled.
- 8.The rows are listings, not distinct suppliers. Seller identity is deliberately not collected, so the sample cannot be de-duplicated by shop and one seller may appear more than once.
Use this data
Everything below is free to reuse with attribution. If you are writing about this and need a cut of the data we have not published, ask and we will run it.
Quote it, or take the whole file
One line, free to quote with attribution
117 of 120 top-ranked 1688 listings source their shipping weight from the merchant's own entry, and only 3 carry a weight measured by a fulfilment centre.
How to cite this study
Cai, Liam. "Where the shipping weight on a 1688 listing comes from." Supplymo, 12 August 2026, https://supplymo.com/research/1688-weight-provenance.
The underlying data
120 rows, one per listing, no supplier identity. 8 stated limits travel inside the file. Free to reuse with attribution under CC BY 4.0.
Put this chart on your page
The chart updates when the study does, so an embed does not go stale on your side.
n = 120 listings. Collected 2026-08-12. Non-random sample: 12 keyword searches, 10 platform-ranked rows each. All 120 listings displayed a shipping block.
Embed code
<iframe src="https://supplymo.com/embed/research/1688-weight-provenance" width="100%" height="460" loading="lazy" style="border:1px solid #CFC4B5;border-radius:12px" title="Where the shipping weight comes from on 120 top-ranked 1688 listings — Supplymo"></iframe>This records the provenance a listing labels for its own weight figure. It does not verify whether any weight is accurate, and it is not a claim that any seller misstated anything.
Questions we get asked
Does this mean 1688 shipping weights are unreliable?
No. It means we can say where the figures come from, not whether they are accurate. 117 of 120 sampled listings labelled their weight as merchant-entered, which tells a buyer how much independent verification stands behind it: none, at the point of display.
Why do missing dimensions matter if the weight is shown?
Because chargeable weight is the larger of actual and volumetric weight, and volumetric weight needs all three dimensions. For bulky, light goods the volumetric side usually decides the bill. With dimensions absent on 75 of 120 listings, that half of the calculation cannot be completed from the listing alone.
How was the data collected?
Through 1688's own read-only keyword-search and product-detail API, in a single session on 2026-08-12. No orders, messages, inquiries or other write operations were involved. The full row-level data and the method are downloadable on this page.
Can I reuse the data?
Yes, under CC BY 4.0, including commercially. The CSV, the evidence JSON with the method and limits, and a copyable citation are all on this page. If you need a cut that is not published here, ask.