Improving Performance of Agile Teams using Statistical Analysis
summary
Uses Jira and Azure DevOps exports, starts with Excel then moves to R for multiple regression, and pulls ~10k to ~100k API calls to stitch histories. Models show adjusted R² around 0.6 for two portfolios, with bundle size, number of developers, and feature size as the largest contributors to change lead time. Teams build ETL scripts and verification baselines, find measurement debt blocking conclusions, and validate recommendations by remeasuring comparable features after interventions.
AI-written summary · about
chapters (17)
- 0:00Start
- 10:55Delivery Lead Time
- 21:18How Many Bugs Were Reported
- 22:09Statistical Analysis
- 22:45Dependent Variable
- 26:14Wait Time
- 30:43Feature Slicing
- 32:42Data Informed Transformation Loop
- 44:40Measurement Debt
- 47:42Data Extraction
- 50:33Data Availability
- 51:02Time Spent in Various Queues
- 51:30Time Logging
- 52:06Quality Challenges
- 53:41Data Entry Discipline
- 55:15Conclusion
- 58:33Security and Usability
related talks
41mRethinking Measurement for High-Performing Teams
35mGo Big with Supercharged Agile
40mAgile Software Development: Practice and Analytics
43mPredictable Delivery
44mThe cost of high-quality Agile metrics
44mEnhancing the Performance of Engineering Teams
28m05 how we do developer productivity panel
28mInformation to Inspiration: Morgan Stanley's Data-Driven Approach to Developer Productivity