Chinese crude imports are published monthly by customs, in volume and by origin, and they are among the more reliable series in the whole non-OECD dataset. They are the single most watched Chinese oil number for good reason: they are large, they are measured at the border, and they are published on a regular schedule.
The limitation is not accuracy but interpretation. An import volume tells you barrels crossed the border. It does not tell you whether they went into a refinery, into commercial storage, or into strategic reserve, and those three have completely different implications for future demand. A high import month driven by reserve accumulation and one driven by strong refinery runs look identical in the customs series.
Timing adds a second layer. Imports reflect cargoes that were bought weeks or months earlier, at prices and arbitrage economics that no longer apply. A strong import month is evidence about the market as it was when those cargoes were fixed, not about the market today. Reading the print as a current demand signal misdates it systematically.
Where the series is genuinely strong is in the medium-term trend and in the composition. Year-on-year import growth is a reasonable read on structural demand direction, and shifts in the origin mix reveal changing arbitrage and contractual relationships that are hard to observe any other way. Those are slower claims than a monthly print supports, and they are the ones the data can carry.