Data validation rules on product attributes catch bad values at the point of entry, before they spread to every channel. A weight in the wrong unit, a GTIN with a missing digit or a title over the marketplace limit all slip in quietly. In SKULaunch you can set validation rules on attributes, depending on your set-up, so bad values are caught on import and in the grid rather than discovered in a rejected feed.
The problem
Most catalogues validate too late. Data is accepted as typed or imported, and problems surface when a marketplace rejects a feed, a filter shows a nonsense value, or a customer spots a 400 kg kettle. By then the error has spread to every channel, and finding the cause takes longer than the fix.
The usual workaround is a spreadsheet of checks someone runs before a big upload, or a developer's script that only one person understands. Neither runs when a buyer edits a product on a Friday afternoon, and neither tells a supplier what was wrong with their file.
Before and after
An illustrative example of rules on a single category in a small appliances catalogue.
Before: what got through
Weight "4000" with no unit on a kettle. GTIN with 12 digits on an EAN-13 field. Title of 240 characters for a channel with a 150 limit.
After: with rules in SKULaunch
- Weight must be a number between 0.1 and 10 kg for kettles
- GTIN must be 8, 12, 13 or 14 digits with a valid check digit
- Title limited to the channel's maximum length
- Failing values flagged at import and in the grid
How it works in SKULaunch
- Set types first. Attribute types such as number, dimension, range or select already restrict what can be entered.
- Add policies. Define rules such as required, allowed ranges, lengths or formats on attributes, scoped to the categories where they apply.
- Validate on the way in. Imports and edits are checked against the rules.
- Find and fix. Filter for products failing a rule and correct them in bulk.
The main types of data validation rules for product attributes
Almost every product data rule falls into one of these types. Which ones you can apply in SKULaunch depends on the attribute type and your set-up.
- Required. The value must be present. Brand, title, GTIN and the attributes behind your filters are the usual starting set.
- Type. The value must be the right kind of data: a number, a date, a yes or no. "Approx 2" is not a number.
- Format. The value must match a pattern, such as a GTIN with a valid check digit or a supplier code in your agreed structure.
- Range. Numbers must fall between plausible limits for the category. A kettle does not weigh 4,000 kg.
- Allowed values. The value must come from a controlled list, so colour is "Grey" and never "GRY".
- Length. Text must fit the shortest limit of the channels it goes to.
A further type, conditional rules, makes one attribute required only when another has a certain value: if Cordless is Yes, then Base included must be filled. These are worth writing down even if you start with simpler rules, because they capture how your buyers actually think about a product.
A worked rule set for one category
An illustrative set of rules for kettles:
- Required: brand, title, GTIN, capacity, power, colour, main image
- Capacity: number in litres, between 0.3 and 3
- Power: number in watts, between 500 and 3,500
- Weight: number in kg, between 0.1 and 10
- Colour: from the colour list only
- Title: no longer than 150 characters
- GTIN: 8, 12, 13 or 14 digits with a valid check digit
Seven rules, each tied to a real failure: a rejected listing, a broken filter, or a product that looks wrong to a customer. That is the test for any rule. If you cannot name the failure it prevents, leave it out.
Validate what downstream systems reject
Start with the rules your channels already enforce: required fields, character limits, identifier formats. Those are the errors that cost you listings. Then add plausibility ranges for the attributes customers rely on, such as weights and dimensions, where a decimal in the wrong place causes the most confusion. Units matter here too. A range check only works if every value is in the same unit, so normalise units of measure before you set limits.
Common mistakes
- Ranges so tight they reject real products, so people learn to override every warning.
- Global rules that should apply to one category, such as a weight limit meant for kettles hitting garden furniture.
- Validating only at export, when the person who typed the value has moved on.
- Rules nobody owns, so nobody updates them when a channel changes its limits.
- Blocking whole imports for one bad value, which sends data back into spreadsheets.
What changes
Errors are caught at entry, feeds stop bouncing, and the catalogue stays plausible.
Common questions
What are data validation rules for product data?
They are checks on attribute values, such as required, format, range, allowed values and length, applied when data is entered or imported.
What is the difference between validation and completeness?
Completeness asks whether a value is there. Validation asks whether it is right. You need both. See how to measure completeness by category and channel.
Can rules differ by category?
Yes. Rules can be scoped to the categories where they apply.
Do rules block imports?
Failing values are flagged so you can decide to fix or override.
How do I validate GTINs and barcodes?
Check the length and the check digit. See how to validate GTINs and barcodes across the catalogue.
Is this available for every attribute type?
Validation depends on the attribute type and your set-up. Confirm the rules you need with us before you plan around them.
More on product data quality, category-specific attributes, and why incomplete product data slips through.

.avif)