Issue :

A category page displays the same performance score metric across multiple catalogs in the Discovery dashboard. This may appear unexpected when catalogs represent different locales.
 

Environment :

This is expected product behaviour and can be seen in staging and production.

Cause :

Category performance scores are calculated at the account level, not at the individual catalog or site level.
As a result, the same query–product pair will return the same performance score across all catalogs that belong to the same account.

 

Sample Example :

Consider Category ID: 123456789
The same category is queried across multiple catalogs/sites:

• example_catalog_fr
• example_catalog_de
• example_catalog_global

For each catalog/site, the category returns identical performance metrics for the same products:
 

"product1": {
 "totalScore": 95.18,
 "performance": 98.79,
 "relevance": 14.27
},
"product2": {
 "totalScore": 94.97,
 "performance": 98.73,
 "relevance": 14.27
},
"product3": {
 "totalScore": 94.83,
 "performance": 98.69,
 "relevance": 14.27
}

Explanation :

To ensure accuracy and consistency, category performance scoring is performed at the account level.
From a technical perspective:

• When a product exists in multiple catalogs or views, interaction signals from all views are aggregated into a single dataset
• Individual views may not always have sufficient data to generate reliable metrics
• Aggregation helps mitigate data sparsity and improves model stability

This approach ensures that performance metrics are based on the most complete and reliable data available.
 

Conclusion:

If a category shows the same performance score across multiple catalogs, this is by design and does not indicate a data issue.