An online store can look polished while still having serious problems in its product data. Products may be missing , images or technical specifications. The same item may exist several times under slightly different names. As the catalogue grows, small inconsistencies can quickly become operational problems.
A is a structured review of product information. The goal is not simply to count empty fields, but to understand whether the data is complete, consistent and useful across the store and the channels connected to it.
What counts as product data?
Product data includes the information that identifies, describes and presents a product: titles, , , , brand, category, price, availability, dimensions, materials, technical specifications, descriptions, images and variants.
For a store with 30 products, much of this can be checked manually. With 5,000 or 50,000 products from multiple suppliers, the task changes completely.
Missing and inconsistent information
A product may have a title and price but lack an identifier, brand, dimensions or important technical attributes. In other cases the information exists only in some categories or from some suppliers.
Consistency matters just as much. The colour black may appear as Black, black, Sort and BLK. A person understands the intention; filters, exports and automated rules may treat them as different values.
standardises those values so equivalent information is handled consistently.
Identifiers and duplicate products
Reliable identifiers help determine whether two records describe the same product.
GTIN is a global trade item identifier. SKU is usually an internal stock identifier, while MPN is the manufacturer’s own part or product number.
An audit can detect missing or invalid identifiers, duplicated identifiers and inconsistent use of SKU or MPN.
Large catalogues also often contain duplicates. One supplier may call an item Bosch GSR 18V-55 Professional, while another uses GSR18V55 Bosch Drill. Name-only matching may create two separate products. Better matching combines signals such as GTIN, SKU, brand and MPN.
Categories, filters and variants
Catalogue quality is also about structure.
Uncategorised products, overlapping categories, inconsistent units and poorly governed attributes affect navigation and filtering. If a Width filter uses millimetres in one category and centimetres in another, a data problem becomes a user-experience problem.
The same applies to each . Size, colour or capacity variants should be linked correctly and carry the right price, availability, identifier and image.
Images and descriptions
Images and product copy are part of the dataset too.
An audit can reveal products without images, very small assets, repeated images or categories with inconsistent presentation.
Descriptions should provide the information customers actually need to evaluate a product. A sentence such as Professional quality drill is not very useful if the purchase depends on capacity, compatibility, weight or intended use.
Product data rarely stays in one system
Product information may start in an or supplier file, move into a , appear in the online store and then be distributed through a to Google or a marketplace.
The website may also use so search engines and other systems can interpret the product information.
When those systems disagree on price, availability, identifiers or names, consistency breaks down. Google Merchant Center documents that missing, inaccurate or conflicting product data can lead to limited eligibility, display problems or product disapprovals.
What should an audit deliver?
A useful audit should end with concrete findings, not a vague statement that the data can be improved.
It can show how many products are missing GTIN, brand or technical specifications, how many possible duplicates exist, and which categories have the most serious problems.
The findings can then be prioritised according to risk: critical data problems, issues that affect findability and filtering, and lower-priority improvements.
When is a Product Data Audit useful?
It is particularly useful when the catalogue has grown quickly, suppliers send different spreadsheets, teams spend too much time correcting information manually, filters are unreliable, or the store is preparing for a migration.
It can also be valuable before implementing a PIM, before a major redesign or when Merchant Center reports many product-level issues.
Not every online store needs a redesign first. Sometimes the most important problem is not the interface but the information behind it.