Choose a data sample that tests all eight marketplace sources
Build a source-by-field coverage checklist before accepting a collection from eight Chinese marketplaces.
A tidy sample can show that a file is readable while leaving important source types untested. Before expanding a data collection, choose examples that expose the decisions your team needs to make.
This guide covers 1688, Alibaba, Taobao, JD, Pinduoduo, Weidian, Goofish and Yiwugo. A Facebook vendor post advertises several families of 1688 data. That is a useful prompt to ask what each proposed field actually means; the advertisement does not establish its availability or accuracy.
Choose cases by source and data shape
Start with the source list and the fields you will use. For each source, select a representative listing and a difficult case where the required information may be missing, conditional or nested. Record how the examples were selected.
Use the website’s fictional pack to plan the checklist: quantity tiers on 1688 and Alibaba; options on Taobao, JD and Weidian; offer conditions on Pinduoduo; stated condition on Goofish; and shop context on Yiwugo. These illustrate different structures, not guaranteed fields on every live listing.
Count the records you actually tested
Consider an original fictional pilot of sixteen links, two per source. If twelve return records and four remain unresolved, the review should show all sixteen outcomes. A file containing only the twelve successes hides the coverage gap.
For each required field, record whether it was observed, unavailable in the source or unresolved in collection. Report the denominator: twelve completed records and sixteen attempted links answer different questions. Do not present this small, deliberately selected pilot as an estimate of the whole marketplace.
Agree what an unresolved case changes
Write down which gaps block the intended task and which can remain visible. If variant dimensions are essential to catalogue preparation, a missing dimension needs an owner and a decision before expansion. Optional shop context may be acceptable when clearly marked unavailable.
Ask the reviewer to record accepted, needs clarification or outside the agreed scope, with the reason. Confirm revisions against the same cases. PandaBuyNow’s public files are fictional structure examples; project coverage, sample selection and acceptance criteria require an agreed project sample. This review concerns data delivery and does not test physical product quality.