Advice on Measuring Software Developer Productivity: A Practical Framework

McKinsey recently proposed a sound framework that combines DORA (Google) and SPACE (Microsoft) metrics which I use to monitor our team's productivity. These metrics are rational and show the impact of software development on the organization. I believe software development outcomes are always a shared responsibility of the team, rather than an individual.

Key Team Performance Metrics

• End-User Complaints: Low number of end-user complaints not only brings a positive impact on a business but also reflects a team's understanding of crucial priorities.
• Issue Resolution Time (Including Deployment): Swift issue resolution highlights an efficient development process and high code/design quality.
• Time for Changes/Enhancements: A team's ability to effectively plan, design, and implement features is crucial. A lagging pace might indicate technical debt or design concerns.
• Change Failure Rate: This metric tells us about software stability and maturity level
• Meeting Own Estimates: Such skill requires a high level of technical and business understanding of the project

Evaluating Individual Contributions

Once a competent team is established, assessments can rely on feedback from teammates and direct managers. During individual evaluations, I consider:

• Work Attitude: Positive and proactive attitudes improve collaboration and problem-solving.
• Self-Learning: Active learning indicates innovation and a growth mindset.
• Collaboration: Willingness to support teammates bolsters productivity.
• Domain Understanding: An engineer cannot be fully efficient without understanding the project domain.
• Code Readability: Clean code accelerates development and is a mandatory part of software quality.
• QC Interaction: Numerous rejected merge requests slow down the process and bring too much work to the QC team. Developers bear the responsibility for the quality of their work.
• Process/project improvement efforts: Good engineers always try to decrease routine and automate as much as possible.

Further Resources

For a deeper understanding, refer to McKinsey's article "Yes, You Can Measure Software Developer Productivity". Remember any metrics and frameworks should be adapted to your company.