Naming and Data Hygiene for a Product Sheet
Most sheets fail on naming, not on sourcing. “Adidas pants”, “adidas track pant”, “Adidas Tiro 23” and “tracksuit bottom” are four rows for one product family, and none of them can be filtered reliably.
The fix is a convention: brand, model, variant, colour, size range — in that order. If the listing does not name a model, use the closest descriptive phrase and put the seller’s title in a notes field. Duplicates are not a data problem; they are a naming problem.
A sheet with conventions can be checked in an afternoon. A sheet without them cannot be checked at all.