Every claim on this site carries a date and a source. This page explains where those sources come from and what we store.
Method and sources
Last reviewed 22 October 2026
The rule we work to
- Read the original, not the summary. Where a rule comes from a Chinese government release, we read that release — not an English article describing it.
- Date every claim. A fact without a date is a rumour. Each claim on this site is stamped with when we read it.
- Archive what we read. Source pages are saved with the date they were captured, because links rot and pages get rewritten. The archive is the asset.
- Do not publish numbers that move. Tax rates, exchange rates, prices and visa lists change. We tell you where the current figure is published and how to check it, instead of printing a number that will be wrong in a month.
- State what we could not verify. Every page ends with a list of open items. If that list is long, the page is weak — that is useful for you to know.
Source types, ranked
| Rank | Source | How we use it |
|---|---|---|
| 1 | Chinese government releases — tax administration, immigration, commerce, customs, airport authorities | Primary. Read in Chinese, quoted with the issuing body and date. |
| 2 | State media in Chinese and English — Xinhua, CCTV, People's Daily | For official statistics and for how the phenomenon is being framed. Attributed as state media, not treated as neutral. |
| 3 | Platform help pages and operator documentation | For how payment, refund or shipping systems behave in practice. |
| 4 | Independent reporting and traveller accounts | For experience, never for rules. Flagged as anecdote. |
| — | Other English guides | Not used as a source. If our text restates one, it is a mistake we want reported. |
Archive
Captured source pages are stored under /sources/, one directory per content page, named by the date of capture. Each content page links to the capture it relied on.
Every capture is the page as it read that day: the URL, the byte count of the raw response, and a SHA-256 of it, so a reader can tell whether the file they are looking at is the one that was fetched. The pages that could not be fetched are kept too, as a file that says what the error was, and the content page that relies on them says so in its own "What we did not verify" section rather than implying we read something we did not.
Browse the archive from the source index. It is excluded from the sitemap and from crawling, because it is an appendix rather than part of the site.
Where a page could not be captured, the content page says so rather than implying we read something we did not.
Updates
Pages carry a "last reviewed" date. Where a rule changed, the page records both the old and new position with the date of the change, so you can tell whether a guide you read earlier is still current.
Contact
Corrections and source suggestions: hello@chinahaul.best.
That address forwards to our editor's inbox rather than being a mailbox in its own right. Two consequences, stated plainly: a message with an attachment will not arrive, and any reply we send comes from the editor's own address, not from chinahaul.best. If you want confirmation that something reached us, ask in your first message and we will quote back the date we logged it.
Limits of this method
- We are not on the ground. We do not visit markets and we do not shoot photographs. Anything that would require standing in a shop is out of scope until a local partner is involved.
- Chinese primary documents are not always published in English, and some are only available as PDFs behind government portals. Where a capture failed, we say so.
- Rules are interpreted, not invented. Where the official text is ambiguous, we quote it and say it is ambiguous rather than picking the reading that makes a cleaner sentence.