How can we measure engineering productivity without creating a surveillance culture?

Avoid simplistic metrics such as lines of code or number of pull requests; they are easy to game and do not correlate well with value. Instead, focus on flow metrics like lead time from idea to production, frequency of deployments, and how often changes cause incidents or rollbacks. Combine these numbers with qualitative signals such as stakeholder satisfaction, clarity of roadmaps, and the team’s own sense of progress. Use metrics to ask better questions, not to rank individuals or assign blame. When you see a slowdown, look for systemic issues such as unclear priorities, brittle tests, or overworked reviewers before blaming people.