Critiquing Common Engineering Productivity Metrics
Alex challenges traditional metrics like commit frequency and advocates for measuring system stability and delivery speed.
The discussion focuses on output evaluation at when Alex critiques conventional productivity metrics used by management. Alex asserts that tracking raw commit counts, pull request volume, or lines of code written creates perverse incentives that encourage developers to break changes into artificially tiny, meaningless fragments.
Smac shares examples at where teams gamed metric dashboards to meet arbitrary quarterly targets without improving actual product capabilities. Smac emphasizes that quantitative tracking often captures activity rather than meaningful outcome or value creation.
Alex proposes focusing on delivery health indicators at , such as deployment frequency, lead time for changes, and mean time to recovery after incidents. Alex argues these operational indicators provide a far more accurate reflection of engineering capability without distorting developer behavior.
Smac concurs at , concluding that qualitative assessments and team health surveys must complement high-level operational metrics to provide a full picture of engineering organization performance.