Cycle Time Analytics
- Cycle time analytics
- Models change nature of discussion
- Based on numbers
- Scrum / Kanban don’t go to Cone of Uncertainty
- As uncertainty is reset at the end of each “sprint”
- What about no historical data
- Weibull distribution
- Just use min & max cycle times
- Has been within 7%
- Book: How to measure anything
- Book: The Principles of Product Development Flow
- LeanKit has this in it
- Using monte-Carlo simulation on measured cycle time and WIP
- Create simple model to run simulations to perform forecast
- Can give you a basis for diagnosis
- Central limit theory
- Book: The Flaw of Averages
- Forecast
- Estimate
- Within a stated uncertainty
- Estimate
- Estimate or #NoEstimate
- 70% chance you need to forecast
- Models change nature of discussion
changearc 