Beyond Estimates (Estimates and "NoEstimates") - Let's Explore the Possibilities
summary
Argues that time/cost estimates are guesses biased by anchoring and planning fallacy and often drive bad decisions; cites a canceled project that cost about $75M and a practice since 2003 of dropping estimates and using one-week experiments and tiny increments instead. Uses work time vs cycle time distinctions, applies an 80/20 then 20/20 = 4% observation (20% of features give 80% value, do 20% of that) and reports early delivery of a small slice in ~3 weeks that changed priorities and kept the project moving.
AI-written summary · about
chapters (10)
- 0:00Introduction to Beyond Estimates
- 1:53Why estimates don’t work
- 7:52Cycle of continuous no improvements
- 10:20What is an estimate really
- 14:58Purpose of estimates
- 22:19Bias and noise in estimates
- 27:12Domains where estimates apply
- 32:53Practical no-estimates approach
- 37:07Delivering value without estimates
- 40:00Q&A
