Framework Agnostic Capacity Planning at Scale
summary
Framework-agnostic capacity planning uses Jira and Azour DevOps data, Monte Carlo simulations (about 4,000 runs) and percentile choices (default 85th) to convert story throughput into feature capacity for ~100 teams. Teams adopt right sizing (e.g. reducing feature child items from 26 to ~5), which cuts average cycle time from 40 to 23 days in one case, and use continual reforcasting to produce probabilistic delivery windows rather than single deterministic dates. Probabilistic forecasting drives plans, not blind automation.
AI-written summary · about
chapters (28)
- 0:00Introduction
- 1:27Why framework agnostic matters
- 3:13Problems with traditional planning
- 7:03Quarterly planning challenges
- 9:31Industry thought leaders
- 10:33Framework agnostic explained
- 11:15Breaking features down
- 12:19Right sizing with data
- 15:05Right sizing in action
- 18:55Forecasting to confidence
- 21:02Monte Carlo simulation
- 23:05Validating forecasts
- 24:29Continuously reforcasting
- 27:18Stop starting start finishing
- 28:02Scaling across teams
- 31:31Tips for implementation
- 32:55Summary
- 34:06Q&A
- 34:27Q&A fresh teams
- 35:09Q&A trial period
- 36:13Q&A planning events
- 36:54Q&A enterprise architecture
- 37:57Q&A client forecasting
- 38:38Q&A short vs long-term
- 39:01Q&A conservative stakeholders
- 39:43Q&A team changes
- 40:26Q&A breakdown struggles
- 41:07Closing
related talks
47mThe Only 2 Things needed to Scale Agile
58mScaling Agile in Organizations
51mMastering Trade-offs for Scaled Agile Success
43mPredictable Delivery
45mProject Release Planning: Building an Iterative Framework.
59mSet Your Teams Up for Improved and Enduring Velocity
47mAccounting for Agile software
57mThe Science of not estimating