An online store can look professional while having substantial product-data problems.
Items may lack , images or technical specifications. The same product may appear several times under different names. As the catalogue grows, these inconsistencies quickly become operational problems.
A systematically reviews product information. Its purpose is to establish whether data are complete, consistent and usable in the store and its connected channels.
Product data are more than names and prices
Product data include all information identifying, describing and presenting a product.
This may include product names, , GTIN, , brand, category, price, availability, dimensions, material, technical properties, descriptions, images and variants.
A small catalogue can often be checked manually. With thousands of products from several suppliers, quality assurance becomes a discipline of its own.
Missing information is only one kind of error
Some products lack information. Others have information entered in several ways.
The colour black may appear as Black, black, Sort and BLK. A person understands the meaning. Filters, exports and automation may treat these as four different values.
standardises such values so information with the same meaning is treated consistently by the system.
Identifiers make matching possible
Reliable identifiers make it easier to determine whether two datasets describe the same product.
GTIN is a global trade-item identifier. SKU is normally the business's internal item number, while MPN is the manufacturer's part or product number.
A review can identify missing GTINs, invalid identifiers, the same identifier on multiple items and inconsistent SKU or MPN use.
Duplicates are related. Two suppliers may describe the same product differently. Matching should therefore use more signals than the name alone.
Categories and filters reveal data quality
Product data become visible through categories, filters and variants.
Products without categories, almost identical category names and inconsistent attributes weaken navigation. If a width filter uses millimetres in one category and centimetres in another, the problem is both technical and editorial.
The same applies to . Size, colour and capacity must connect to the correct main product while carrying the correct price, availability, identifier and image.
Images and descriptions are data too
Product images and text belong to the same quality system.
An audit can identify missing images, very small images, the wrong reused image or categories with entirely different image standards.
Descriptions can have the same problem. A generic phrase such as 'professional quality drill' does little when customers need capacity, compatibility, weight and usage information.
The goal is not the longest possible description, but having the information affecting the choice.
Errors follow the product further
Product information is rarely used only in the store.
Data may travel from or supplier files into a , then the store and a for Google or marketplaces. The website may also use to make information machine-readable.
Different prices, names, availability or identifiers across systems create consistency problems.
Google Merchant Center documents that incomplete or conflicting product data can cause limited visibility, incorrect display or rejection.
The result should be a prioritised error overview
A Product Data Audit should not end with a general recommendation to improve the data.
It should show specific findings: how many products lack identifiers, which categories lack technical properties, how many possible duplicates exist and where the main problems lie.
Findings can then be prioritised by consequence:
- errors that stop distribution or identify the wrong product
- errors weakening search, filtering and comparison
- missing information primarily reducing content quality
When a review adds most value
A Product Data Audit is particularly useful when a catalogue has grown quickly, suppliers provide different Excel files, staff spend substantial time on manual corrections or filters work poorly.
It is also relevant before implementing a PIM, migrating platforms, a larger redesign or when Merchant Center reports many product errors.
Not every store first needs a new design. Some first need better control of the information the design will display.
