How to use it
- 1. Set the curveFirst-unit hours and the learning rate: 80% means every doubling of volume cuts unit time to 80%.
- 2. Toggle log–logThe power curve becomes a straight line - that’s how you fit a real rate from factory data.
- 3. Use the solverType a target unit cost and get the exact unit number where you cross it - pricing below today’s cost, legally.
- 4. Compare two factoriesA slower starter with a steeper rate eventually wins. Find the crossover unit.
Wright’s learning curve
Your factory
Rival factory (dashed)
Price-crossing solver
–
Time for unit N (hr)
–
Total run hours
–
Avg hours/unit
–
Cost of unit N
Unit hours vs unit number. In log–log the curve is a straight line with slope log2(rate) - the fingerprint of learning.
The doubling table: same percentage drop from 1→2 as from 128→256. Early units teach cheaply; later lessons cost thousands of units.
Try these experiments
- At 80%, check the table: unit 2 = 80 hr, unit 4 = 64, unit 8 = 51.2. Every row is ×0.8 - the whole model in one column.
- Set the rival slower at start (140 hr) but at 72%: find the unit where its dashed curve dives under yours. Experience compounds like interest.
- Use the solver to price forward: if the market pays $2,000/unit, how many units until you profit? That’s the logic behind aggressive launch pricing.
About this tool
The concept
Wright (1936) noticed aircraft labour fell a fixed percentage with every doubling of cumulative output. The effect drives capacity planning, make-vs-buy and launch pricing.
How it works
Tn = T1·nb with b = log(rate)/log 2. Cumulative time is summed exactly; the solver inverts the power law for the target-cost unit.
Winning strategy
Bid on tomorrow’s cost, not today’s: with a real learning rate, early losses are tuition. But beware - learning stalls when products or crews change.
Background: Experience curve · Wright (1936), Journal of the Aeronautical Sciences