Wright's Learning Curve (Unit Time)

T. P. Wright noticed in 1936 that every time cumulative aircraft output doubled, the labour hours per unit fell to a consistent fraction of what they were. That fraction names the curve: an 80% curve means unit 2 takes 80% of unit 1, unit 4 takes 80% of unit 2, and unit 8 takes 80% of unit 4. Starting from 100 hours, that is 80, then 64, then 51.2 hours for the eighth unit. The closed form $T_n = T_1 n^{\log b/\log 2}$ does the same walk in one step.

The rate is not a constant of nature; it depends on how much of the work is human. Assembly-heavy aircraft and shipbuilding work sit near 80%, mixed machining and assembly nearer 85 to 90%, and highly automated processes above 95% because a robot does not learn. Quoting a first-of-a-kind job at the eventual steady-state rate is the mistake that eats the margin, and quoting it at the first-unit rate is the mistake that loses the bid.

Two cautions. This is the unit formulation, giving the time for the nth unit specifically, not the cumulative average form that Boeing popularised, and the two give different numbers from the same rate. And the curve flattens: real learning stalls once the process is tooled and the operators are practised, so extrapolating an 80% curve out to unit 10,000 predicts labour hours that no factory has ever achieved.