Meesho Data Scraping Services

A product has a price distribution, not a price

Meesho's model lets resellers list a supplier's product with their own margin added. The result is that the same product appears at multiple prices simultaneously, set by different sellers with different margin expectations.

Why averaging destroys the signal

• The spread is the finding. A wide spread indicates a product where resellers see margin headroom; a narrow one indicates competition has compressed it.
• The floor matters most for competitive comparison, since that is the price a shopper can actually find.
• The mode indicates where most resellers have settled, which is a better read on the working market price than the mean.
• Outliers are informative rather than noise — an unusually low listing often signals inventory clearance.

We deliver price_min, price_max, price_median, listing_count and the full distribution where scope allows, rather than a single figure. For anyone benchmarking against Meesho, the floor and the median answer different questions and both are needed.

Supplier and reseller are different parties

A Meesho listing involves a supplier who holds the product and a reseller who is selling it onward. Where both are visible, they are commercially different actors and conflating them makes seller analysis meaningless.

• Supplier identity tells you who is actually sourcing the product, which is what a brand cares about for channel control.
• Reseller identity tells you who is marketing it, which is a different question entirely.
• One supplier can sit behind many resellers, so counting listings as sellers overstates the number of parties involved by a large factor.

We capture supplier_name and reseller_name separately where the platform displays them, plus distinct_suppliers per product. For brand protection work, one unauthorised supplier behind forty reseller listings is one problem, not forty — and only separated data shows that.

Very low price points break naive unit pricing

Meesho operates at price points substantially below most marketplaces, and this creates arithmetic problems that do not arise elsewhere.

• Rounding matters proportionally more. A one-rupee difference on a fifty-rupee item is a two percent price gap, which at higher price points would be noise.
• Shipping is a large share of total cost, so item price alone describes less of the transaction than usual.
• Pack sizes are often small or unstated, and computing unit price from an unparsed pack produces figures that are wrong by multiples.
• Free-shipping thresholds shift effective cost sharply at these levels.

We compute unit price only where pack architecture parses confidently, and return null with a reason otherwise rather than estimating. Shipping is captured separately, and we retain price at full precision rather than rounding — at these price points, rounding to the nearest rupee changes competitive conclusions.