Original research

What a Chinese wholesale marketplace actually shows a buyer

We sit on the supply side, in Yiwu. These studies measure what is on the page at the point a supplier is chosen, using the marketplace's own read-only API.

Journalists and researchers are welcome to use the figures. Every study publishes its raw evidence file and states what its sample cannot support.

Liam Cai · Founder, Supplymo

Yiwu, China

Published studies

Studies

Each one links to its own method, evidence file and limits.

4 August 2026

A supplier score of 5.0 can sit on top of a 1.0

38 of 120high headline, low component

In a Supplymo sample of 120 top-ranked 1688 listings, 65 showed the maximum 5.0 supplier score and 21 of those carried a component score at or below 2.5.

120 listings · 12 categories · 10 stated limits

Read the method and limits

4 August 2026

The listing you find first is rarely the cheapest

41 of 48had a cheaper visual match

In a Supplymo pilot of 48 top-ranked 1688 listings, 41 had a lower-priced retained visual match on the first image-search page.

48 seeds · 12 categories · 8 stated limits

Read the method and limits

How these are made

Three rules every study here follows

Read-only, and reproducible

Every sample is collected through the marketplace's own public API using search and detail endpoints only. No order, payment, enquiry or messaging endpoint is ever called. The sampling script is in the repository.

The evidence file is published

Each study ships a sanitised evidence file carrying every row, the full method and the limits. No supplier identity exists at any stage: offer IDs, seller IDs, shop names, titles and prices are never written to it.

The limits are stated up front

These are non-random samples taken from ranked search results on a single day. Each study says so in its own words, near the top, rather than in a footnote.

For journalists

Using these figures

The figures are free to quote with attribution to Supplymo and a link to the study. Each study page carries a downloadable evidence file if you want to check a number or run your own cut. If something is unclear, or you need a figure the study does not publish, write to us and we will say plainly whether the sample can answer it.