Alibaba Data Scraping Services

A large share of listings have no price, and that is the finding

Consumer marketplaces always show a price. B2B wholesale often does not. A listing may show a range, a price at a stated MOQ, or nothing but a prompt to contact the supplier.

Why this cannot be papered over

• An estimated price is the least defensible field you could add, because the whole point of contact-supplier pricing is that it is negotiated.
• A range midpoint is not a price. It is arithmetic on two numbers the supplier chose for signalling.
• The no-price share is itself informative — categories where most listings hide price behave differently from those where price is published.
• Comparison must be restricted to listings with a comparable price state.

We deliver price_state as published_single, published_range, price_at_moq or contact_supplier, with price_min and price_max populated only where a range was actually published. Where the state is contact_supplier, price fields are null with a reason.

We also report the no_price_share per category, because it tells you how much of that category a price analysis can actually cover. That number is usually the first thing worth knowing and it is never in a vendor's sample.

MOQ tiers, and price only means something at a quantity

Where price is published it is typically tied to a minimum order quantity, often with a tier ladder. This is the same structure as our industrial and MRO service, and the same rule applies: a price without its quantity is not a price.

• MOQ varies enormously between suppliers for the same product, so the cheapest unit price may require ten times the order.
• Tier ladders mean unit price falls with volume, and the ladder shape differs by supplier.
• Sample availability and sample price are separate from bulk pricing and are a real sourcing field.
• Customisation options affect both price and MOQ.

We capture moq, price_tiers as an array, and sample availability where published. Comparing suppliers on unit price without MOQ produces a shortlist you cannot actually buy from.

Verification badges are captured, not endorsed

The platform displays supplier verification and assurance indicators. Those are useful signals and they are not our assessment.

• We capture the badges as published, with verification_verified_by_us set permanently false.
• We do not assess supplier legitimacy, capability or capacity. That requires audit, not extraction.
• We do not interpret response-rate indicators as reliability measures, since their computation is not published.
• Supplier location is captured as published, which is a sourcing field, not a verified fact.

The reason this matters more here than elsewhere: a sourcing decision made on a badge treated as our endorsement carries real commercial risk for the buyer. We record what the platform displayed on a date. Due diligence is a separate exercise and this dataset is an input to it, not a substitute.