PIM and product content
Product data quality: how to measure completeness and accuracy
Quick answer
Product data quality is measured on six dimensions: completeness, accuracy, consistency, validity, timeliness and uniqueness. Completeness is scored automatically per channel as the share of required attributes filled. Accuracy needs sampling against the physical product or source documents. Most PIM teams target 95 percent or higher completeness on mandatory attributes.
Key takeaways
- Measure product data quality on six dimensions: completeness, accuracy, consistency, validity, timeliness and uniqueness.
- Completeness should be scored per channel and locale, because a product can be complete for the webshop and incomplete for a retailer.
- Accuracy cannot be measured by software alone; a monthly sample of 50 to 100 products checked against the physical item or source is a practical method (editorial estimate).
- Typical targets are 95 percent or higher completeness on mandatory attributes and 98 percent or higher accuracy on regulated and logistic fields (editorial estimate).
Product data quality is measured on six dimensions: completeness, accuracy, consistency, validity, timeliness and uniqueness. Completeness and validity can be scored automatically in a PIM, while accuracy requires sampling against the physical product or its source documents. Teams that measure all six, per channel and per locale, can link data quality directly to listing rejections, returns and compliance risk.
The six dimensions of product data quality
The six dimensions below are mutually exclusive and together cover what can go wrong with a product record. They follow the data quality dimensions widely used in data management practice.
| Dimension | What it checks | How to measure | Example metric |
|---|---|---|---|
| Completeness | Required attributes are filled | Automated, per channel and locale | Percent of mandatory attributes filled |
| Accuracy | Values match the real product | Sample audit against product or source | Percent of sampled values correct |
| Consistency | Values agree across systems and variants | Automated comparison PIM vs ERP or channels | Number of mismatches per 1,000 records |
| Validity | Values follow format and business rules | Automated validation rules | Percent passing GTIN check digit, unit and list rules |
| Timeliness | Data is available when needed | Process timestamps | Working days from creation to publish |
| Uniqueness | Each product exists only once | Duplicate detection on GTIN, SKU, name | Number of duplicate records |
The so what: report all six, since a high completeness score hides wrong values. A catalogue can be 100 percent complete and still list incorrect weights on every product.
How to measure completeness per channel
Completeness should be calculated per channel and locale, because requirements differ by recipient. The formula is simple: filled required attributes divided by total required attributes for that channel, family and language. A product can score 100 percent for the own webshop and 70 percent for a marketplace that requires extra fields. Most PIM platforms, including Pimcore and Contentserv, calculate completeness natively. Distinguish two tiers: mandatory attributes that block publication, and recommended attributes that improve conversion, as described in product experience management.
How to measure accuracy: sampling and audits
Accuracy needs human verification, so run it as a recurring sample. As an editorial estimate, a monthly random sample of 50 to 100 products gives a reliable trend for most mid-market catalogues. Compare key attributes with the physical product, the packaging or the supplier's technical sheet, and record each value as correct or incorrect. Prioritise attributes with commercial or legal impact: dimensions and weights affect logistics and returns, while materials, allergens and warnings carry compliance risk. The GS1 Data Quality Framework, a best-practice guide aligned to the GDSN, helps suppliers assess and continuously improve the quality of master data shared with trading partners.
Setting targets and owners
Targets only work when each attribute group has an owner who can fix issues. The following process sets up a workable data quality routine in 6 to 10 weeks (editorial estimate).
- Classify attributes into mandatory, recommended and optional per channel.
- Assign a data owner to each attribute group, usually in product management, and data stewards for daily fixes.
- Set targets: 95 percent or higher completeness on mandatory fields and 98 percent or higher accuracy on regulated and logistic fields are common starting points (editorial estimate).
- Configure validation rules and publication blocks in the PIM.
- Publish a monthly dashboard per channel, category and owner.
This mirrors the ownership principles in DAM governance, applied to product records.
The business cost of poor product data
Poor product data shows up in three cost categories, which together form the business case for a data quality programme. Channel costs arise when retailers and marketplaces reject listings, delaying revenue, as covered in product content syndication. Customer costs arise when wrong sizes, colours or specifications cause returns and complaints. Compliance costs arise when mandatory information is missing: since 13 December 2024, the General Product Safety Regulation (EU) 2023/988 requires online listings to show manufacturer details, product identification and safety warnings. Validity rules based on GS1 standards, such as GTIN check digits, catch a share of these errors before publication.
What to do next
- Run a baseline on all six dimensions for your top 500 products.
- Configure channel-specific completeness scoring in your PIM.
- Start a monthly accuracy sample of 50 to 100 products, focused on high-impact attributes.
- Name an owner for every attribute group and publish targets.
- Link quality scores to rejection rates and return reasons to quantify the business case.
Frequently asked questions
What is a good product data completeness score?
For mandatory attributes, most teams aim for 95 percent or higher before publishing, and many PIM systems block publication below 100 percent on mandatory fields. For recommended attributes, 70 to 85 percent is a common target. Set targets per channel, since requirements differ by recipient.
How do you measure product data accuracy?
Take a random sample of products, compare key attributes with the physical product, packaging or supplier documentation, and calculate the share of correct values. Focus on attributes that cause returns or compliance risk, such as dimensions, weight, materials, allergens and safety warnings.
Who is responsible for product data quality?
Each attribute group should have a named data owner, usually in product management, with data stewards in e-commerce or content teams who fix issues. A central product data lead sets rules, monitors dashboards and reports quality scores to management.