Learning curve
Also known as: experience curve, learning curve analysis
A learning curve describes how the time or cost required to produce a unit falls as workers repeat a task and gain experience. In cost accounting, it is modeled so that each doubling of cumulative output reduces the cumulative average time per unit by a fixed percentage.
A learning curve captures a simple observation about repetitive work: people get faster at a task the more times they do it. Early units take the longest, and improvement continues at a predictable — though decelerating — rate as cumulative output grows. Cost accountants use the relationship to forecast labor hours, set standards, and price contracts for products that are still early in their production run.
The standard model is the cumulative average-time learning curve, stated as a percentage. On an 80% learning curve, every time cumulative output doubles, the cumulative average time per unit drops to 80% of what it was. If the first unit takes 100 labor hours, then two units average 80 hours each (160 hours total), four units average 64 hours each (256 total), and eight units average 51.2 hours each (409.6 total). The incremental time for the second unit is only 60 hours, and for units five through eight together, 153.6 hours. A lower percentage means faster learning.
The practical uses follow directly from the math. A bid priced off first-unit cost will be far too high on a long production run, while a bid priced off a fully mature rate will lose money on the early units. Learning effects are strongest for labor-intensive, complex, repetitive processes and weakest for highly automated ones, and they flatten out once the task is mastered — assuming an unlimited learning effect is the classic forecasting error.
The CMA Part 1 exam tests learning curve analysis within forecasting techniques, and calculation questions are common: given a first-unit time and a curve percentage, compute the total or incremental hours at a stated output level. Know the cumulative average-time convention, since the alternative incremental unit-time model produces different numbers from the same percentage.
Key takeaways
- A learning curve models the decline in time or cost per unit as cumulative production experience grows.
- On an X% cumulative average-time curve, each doubling of output cuts the cumulative average time per unit to X% of its prior level.
- An 80% curve starting at 100 hours yields cumulative averages of 80 hours at two units and 64 hours at four units.
- Incremental unit time is found by differencing cumulative totals, not by applying the percentage directly.
- Learning effects are largest in labor-intensive, repetitive work and eventually flatten out.