Modern Software Engineering on YouTube · Open on YouTube
Design docs, RFCs and ADRs · Proficient · Lesson 81 of 98 · 17 min video, then about 2 hours of practice
Estimation and planning under uncertainty
After this you can give an estimate as a range with stated assumptions, break a project into milestones that each deliver value, and re-plan honestly when reality diverges.
Why it matters
Estimates are always wrong; the skill is making them useful. Break the work down until each piece is understood, estimate in ranges, state the assumptions and the biggest unknowns, and do the riskiest part first so the estimate improves quickly. The Shape Up idea of an appetite (how much time is this worth?) often beats estimation altogether: fix the time and vary the scope. Trap: padding silently. State the uncertainty openly instead; it builds far more trust.
How to use this resource
A thoughtful take on why estimates fail and how to make them useful.