Master the Art of Estimation

Study Notes

The author believes that effective estimation techniques should capture both time and uncertainty.
The author's technique involves breaking down work into smaller tasks based on complexity and using labels to indicate time estimates.
The author also estimates uncertainty by applying a multiplier to the expected time.
The specific labels and multipliers used are arbitrary but should be consistent over time.
The author recommends aiming for low uncertainty in order to produce a more accurate estimate.
If there are externalities, such as other projects waiting on the current one, a wide range of estimates may be too broad.
Refining the estimate may involve breaking down extra-large tasks into smaller ones or reducing uncertainty on highly uncertain tasks.
Time-boxing can be used to invest time in scoping or de-risking in order to produce a better estimate.
Tracking actual time can be used to refine and train estimation skills over time.
Other estimation techniques include PERT, Monte Carlo, and Wideband Delphi.