Shopify Product Catalogue: A Data Quality Checklist
Catalogue quality affects how confidently customers can choose a product and how reliably a team can maintain the store. Before a large import or automated update, establish which product information is correct and which system is allowed to change it.

Choose an authoritative source for each field
A live store export, a supplier spreadsheet and an internal stock system may contain different versions of a product. Do not assume the newest file is correct for every field. Record which source controls price, stock, descriptions, images and identifiers.
Use a stable matching rule. If you match on SKU, first check for blanks, duplicates, leading zeros and spaces. Treat identifiers as text where necessary. A failed match should go into a review list rather than creating a new product automatically.
Audit a representative sample
Start with products that expose the catalogue’s difficult cases: several variants, long specifications, discontinued lines and items with incomplete images. Record the issue and the proposed correction so the same decision can be applied consistently.
- Product names identify the item clearly.
- Variant options distinguish purchasable choices.
- Specifications come from an approved source.
- Images belong to the correct item and are permitted for use.
- Descriptions do not invent compatibility or warranty claims.
- Collections and filters use consistent values.
Treat imports as controlled changes
Shopify documents its product CSV format and the behaviour of imports. Read those requirements before editing a file, particularly when updating existing products. Do not assume an omitted field and a blank field have the same effect.
Keep the original export and work on a copy. Test a small batch, review the resulting products and only then expand the update. Compare counts before and after the import and investigate unexpected additions or changes. A clear exception report is often more useful than a message saying the import completed.
Build a maintainable catalogue workflow
Record what was changed, when it changed and which source supplied the value. For automated updates, define whether unavailable source data means “leave unchanged” or “set empty”. Decide what should happen if a feed stops arriving.
Prioritise errors that prevent a correct purchase before cosmetic rewrites. A missing variant identifier or misleading specification can cause more trouble than an inconsistent sentence. Once the underlying information is reliable, improve product descriptions and internal links around the questions customers actually ask.
Catalogue Relay is an ARTiFICIAL MEDIA project in development. Its final features and availability are not yet confirmed; this checklist describes a general catalogue approach rather than a promise about that product.