When AI drafts variants and reports quickly, test counts can rise easily. Experiment quality asks whether results are trustworthy and reusable in future decisions.
The demo's TESTS line climbs rapidly while LEARNINGS moves only after evidence checks. Count a learning when the hypothesis, data quality, and result review hold up.
Not every test needs to win. A surprising result has value if it was measured reliably and can inform the next decision.
When to use
Use it when AI increases test volume without improving decisions.