Why use?
Judgments make ranking changes reproducibly similar and help make regressions visible.
Glossary
Relevance Judgment
A relevance Judgment is a predefined assessment of how well a result fits a specific search question.
Levels
Work with clear levels such as relevant, partially relevant and irrelevant, tailored to your use case.
Context
The same product match can be judged differently per query type; for example, a SKU query requires stricter precision.
Maintenance
Capture owner and change reason so that the test set can grow with assortment and search behavior.