قياس الأثر دون خداع النفس

في هذا الدرس

measure real effect with honest metrics.

الموضوعات المخطط لها

• Vanity metrics (lines generated, suggestions accepted) say activity, not value.
• Use delivery metrics (lead time, deployment frequency), quality metrics (change failure rate, defect rate, rework), and developer experience (survey, time on toil).
• Measure review load: more code can mean more review time.
• Beware novelty effect, selection bias (enthusiasts self-select), and short windows.
• Report uncertainty and negative findings.