Agile metrics for predicting the future
summary
Uses just start and end dates and small samples (5–11) to forecast work, not story points; computes a lead time distribution to say e.g. 88% of past stories finish within 9 days and 50% within 4 days, measures flow efficiency (often 5–15%), and counts throughput per sprint. Runs a Monte Carlo simulation on historical throughput to give probabilistic ranges (example: 50% chance in 4 sprints, 85% by 5), surfaces risks, and produces an MVP (Ofcom compliance) that meets the deadline.
AI-written summary · about
chapters (18)
- 0:00Intro
- 5:03Key lessons
- 5:48About Mattia
- 6:18Book
- 6:34Business questions
- 7:16Story points
- 8:35Too much work in progress
- 10:56Unexpected events
- 11:24Smaller is better
- 11:46Data and tools
- 13:43Lead time
- 14:05Kanban example
- 15:35Histogram
- 17:00Conversations
- 20:46throughput
- 26:44probabilistic forecasting
- 32:23continuous forecasting
- 38:10recap
related talks
36mThe Development Metrics You Should Use (but Don’t)
30mSoftware Estimation Made Easy
57mThe Science of not estimating
16mPractically predictable – Forecasting delivery when nothing is certai
41mYour Project Behaves Like a Hurricane--Forecast It Like One
46mEstimates and complexity: a tale in 42 and a half minutes
43mPredictable Delivery
44mPredictably Agile