Metrics We Retired (And What We Use Instead)

Pass rate as a headline metric encourages silent deletion of tests. We retired it in internal dashboards and replaced it with journey completion rates measured against explicit oracles. The Performance Testing Fundamentals cohort sees this contrast early: throughput without context misleads. Second, we talk about lead time to fix flakes. Not glamorous, but it predicts trust in automation better than green percentages. We show how to visualize it without blaming individuals — focus on system incentives. Third, we acknowledge limits. Some metrics need product analytics collaboration; QA alone cannot own them. The post closes with a candid list of metrics we still find tempting but refuse to publish.