Best Practices to Measure Software Engineering Performance

Cyberdime
Published: February 16, 2023

Your board is clearly after something—and their ask for lines of code measures is simply a symptom of their desire to better understand how engineering is performing. 

We’ve implemented a balanced scorecard that provides a holistic view of engineering without focusing on a single, poor approximation of developer ‘effort.’ Further, these metrics provide the needed insight to empower and unblock your team. We have found time and time again that productivity gaps are caused by systemic issues, not individual performance.  

Productivity: Focused on team throughput

  • What is each team’s throughput? 

  • How much time is spent writing and rewriting code? 

  • How efficiently is new code written?

Predictability: Focused on sprints & releasing

Reliability: Focused on how collaboration can improve reliability

  • How do quality issues trend over quarters?

  • How often are defects impacting customers?

  • How thoroughly are PRs being reviewed?

  • What percentage of PRs have no reviews?

Delivery: Focused on DevOps

  • How often do teams deploy code?

  • What is the lead time to make a change?

  • When incidents or defects occur, how long does it take to recover?

  • How often do deployments degrade service?

Source: www.pluralsight.com