Developer productivity is waste - Michael Coté - NDC Oslo 2024

Developer productivity gains often come from factors outside of individual developer activities, particularly in the “outer loop” of software delivery

Key metrics to track include DORA metrics (lead time, deployment frequency, change failure rate, MTTR) rather than individual developer metrics like lines of code

Measuring individual developer productivity is generally ineffective and can be counterproductive. Focus should be on team and organizational outcomes

Main productivity improvements come from:

• Build and deployment automation
• Standardizing platforms and reducing complexity
• Reducing cognitive load and interruptions
• Streamlining the end-to-end software delivery process

Developer surveys and flow state measurements are better indicators than pure activity metrics

Platform standardization and reducing the number of different technologies/tools has a major impact on overall productivity

Focus should be on business outcomes and value delivered rather than developer activity measurements

Management often wants productivity metrics to make staffing and budget decisions, but this approach is typically flawed

AI and automation tools should augment developers rather than try to replace them

Improving the “outer loop” (CI/CD, deployment, platforms) typically yields better results than optimizing individual developer workflows