For a long time, Amazon's data quality check (IDQ) was based primarily on attribute completeness. But with the introduction of CDQv3 (Composite Data Quality) in March 2026, Amazon has shifted its focus: away from the mere presence of data, toward content accuracy and logical consistency. For brands, this means a fully filled-out backend no longer guarantees visibility if the information doesn't semantically add up.
Technical background: How the CDQ's AI checks your data
Unlike the old IDQ system, CDQv3 uses modern Large Language Models (LLMs) and computer vision algorithms. The system no longer runs a simple checklist review, but a semantic matching analysis:
Cross-Modal Matching
The AI compares whether the text in the title matches the attributes in the "Product Overview" and the image content.
Category Benchmarks
The system knows the industry standards for each category. If your details (e.g. unusual units of measurement or contradictory material information) deviate from the standard, this is flagged as a defect.
NLP (Natural Language Processing)
Amazon analyzes titles and bullet points for redundancy and keyword spamming to ensure readability for customers.
The scoring scheme: Weighting and grades
To simplify the management of your catalog, Amazon summarizes quality in a single metric. The weighting clearly shows that the focus is on the elements most visible to customers:
→ The quality grades
Your success is now measured in grades. Your goal should always be Grade A, to secure the trust of customers and the Amazon algorithm.
Grade A (85%+)
High quality. Your data builds trust and maximizes visibility.
Grade B (70–85%)
Good foundation, but with optimization potential.
Grade C/D (0–70%)
Action needed! Your product data has defects that are dragging down your conversion rate.
Grade U (Ungraded)
Warning! This indicates policy violations. Immediate product suppression is at risk.
Input: your levers (weighting)
Output: your data score (grades)
Risk management: Egregious vs. non-egregious defects
Amazon precisely distinguishes how severely an error disrupts the customer journey. This determines how quickly and harshly it is penalized:
Policy Violations
Title errors (such as keyword spam, impermissible special characters, or incorrect length) are now often classified directly as a policy violation. The result is usually immediate search suppression of the listing.
Egregious Defects
A defect is considered critical if the top 6 attributes are factually incorrect, or if the completeness of the top 10 attributes is below 50%. Serious errors in variation families also fall into this category. This leads to an immediate drop in the score and carries a high risk of product suppression.
Non-Egregious Defects
These include non-standardized formatting or minor deviations in capitalization. These errors reduce visibility more gradually ("organically") and should be fixed as part of regular catalog maintenance.
Business impact: Data quality as a profitability lever
Maintaining your CDQ scores is not a technical exercise for its own sake — it has a direct impact on your profit margin. High data quality has been shown to lead to:
Lower advertising costs
A higher conversion rate from accurate data improves your Ad Rank. You pay less for the same ad placement.
Reduced return rates
Since the AI ensures that title, images, and attributes form a coherent whole, the likelihood of mispurchases due to misunderstandings decreases.
Future-proofing
Amazon will roll out further Product Type Classification updates in 2026. Only brands with a clean CDQ score will be placed correctly in the automated category structure updates.
→ Xingu tip: Make your content Rufus-ready too!
Have you heard of Amazon Rufus? The AI chat assistant only recommends products when it understands the content and the content is tailored to customer needs. In our article on Amazon Rufus, you can use the checklist to check whether your listing is Rufus-ready, so the AI showcases you as a top recommendation in the future.
Conclusion: The switch to CDQv3 is a signal for vendors and sellers to treat their catalog as a dynamic asset. Those who optimize proactively secure an unfair competitive advantage over brands that merely react to warning notices.
FAQ – Frequently asked questions about CDQ
Yes, absolutely. CDQ is a catalog-wide metric. While sellers often see the score directly in Seller Central, for vendors it's the basis for catalog health and heavily influences Buy Box stability and visibility.
IDQ primarily checked quantity (Is the field filled in?). CDQ checks quality (Is the content correct?).
Grade U means "Policy Violation." The product is often no longer findable by customers (search suppressed) or is completely blocked until the error is fixed.