Experiments
Test commercial ranking before exceptions become standard.
Use a clear hypothesis and protect existing strong search routes.
Hypothesis
Describe which query group and which commercial signal are being modified.
Baseline
Capture current ranking, CTR and conversion.
Guardrails
Protect exact matches and relevance evaluation.
Segments
Analyze device, store and query type where relevant.
Decision
Document why the change is being rolled out, modified or reversed.