What it means
The pattern was first measured in aircraft manufacturing and has since been observed in electronics, construction, surgery and software delivery. The rate is expressed as a percentage, so an 80% curve means that when cumulative output doubles, the average time per unit falls to 80% of its previous level.
A lower percentage means faster learning. The reasons behind it are ordinary rather than mysterious.
People get quicker with repetition, teams remove awkward steps, tooling improves, scrap falls and suppliers become familiar with the specification. None of it happens automatically, but it happens reliably enough to plan around.
The commercial uses are pricing, bidding and capacity planning. A contractor pricing a run of twenty identical units from the first unit's cost will be far too expensive to win the work, while one that ignores learning entirely on a one-off project will underprice badly.
Estimators therefore need a defensible rate before they quote. It also underpins competitive strategy.
A producer that reaches high cumulative volume first sits on a lower cost base than a later entrant, which is one argument for pricing aggressively to gain share early. The advantage weakens when technology changes and accumulated experience becomes less relevant.
Two cautions are worth remembering. Learning flattens, so the gain between the first and second unit is far larger than the gain between the hundredth and two hundredth, and beyond some point the curve is nearly flat.
It also depends on continuity, because staff turnover, long gaps between production runs and design changes all reset part of the accumulated knowledge.
In practice
Real-world examples.
Example
A solar installer prices its first commercial rooftop project at 640 labour hours. By its eighth project of similar design the crew averages about 330 hours, and the company rebuilds its quoting model around demonstrated experience rather than the original estimate.
Example
A medical device firm bidding to supply 500 units warns its board that the first 50 will be produced at a loss. Its bid assumes a 90% curve, and the board approves the price only after seeing the cumulative cost projection rather than the first-unit cost.
Example
A software consultancy tracks hours per implementation of the same platform across clients. The twelfth implementation takes roughly half the hours of the first, so the firm moves from time-and-materials pricing to fixed-fee pricing and keeps the difference as margin.
Think of it
“The learning curve is getting better and cheaper with practice-experience making you more efficient.
Formula
Calculation
Formula: cumulative average time per unit after X units = time for the first unit x X raised to the power b, where b = log(learning rate) / log(2). For practical work the shortcut is easier: every doubling of cumulative units multiplies the cumulative average time per unit by the learning rate.
Worked example. A shipyard builds a new class of patrol boat. The first hull takes 100,000 labour hours and the yard assumes an 80% learning curve based on its own history with similar vessels.
At 2 cumulative units the average is 100,000 x 0.80 = 80,000 hours per unit. At 4 units it is 80,000 x 0.80 = 64,000 hours, and at 8 units it is 64,000 x 0.80 = 51,200 hours. A contract for 8 boats therefore needs 8 x 51,200 = 409,600 labour hours in total, not 8 x 100,000 = 800,000 hours. At a labour rate of $60 per hour the difference between those two estimates is $23,424,000, which is why the assumed rate is negotiated as carefully as the price itself.Case study
Seen in the real world.
Kestrel Modular is an illustrative builder of prefabricated classrooms invented for this entry. It won a tender for twelve identical buildings and priced all twelve using the labour hours from the prototype, which had taken 2,400 hours. A rival won a similar tender the following month at a price Kestrel's finance director thought was reckless.
Reviewing the loss, the team looked back at an earlier repeat order and found their own records implied roughly an 85% curve. Applying that rate, the twelfth building would have taken close to 1,300 hours rather than 2,400, and they could have bid around 20% lower while keeping the same margin across the contract as a whole.
The fictional company rebuilt its estimating template to require an explicit learning rate assumption on any order of more than three identical units, backed by historical hours where they existed. It won the next two multi-unit tenders and delivered both close to the estimated hours.
Watch out
Common mistakes.
- Assuming the curve applies to one-off, non-repeating work. Learning needs repetition of something similar before it has anything to build on.
- Confusing the learning rate with the saving. An 80% curve means the average falls to 80% of the previous level, which is a 20% reduction, not an 80% one.
- Extending the curve indefinitely into the future, when in reality the gains flatten out once a process is mature.
Questions
People also ask.
Where does the learning rate come from?
Ideally from the company's own historical hours on similar work, because published rates vary widely by industry and process.
Does the curve apply to material costs?
Only indirectly, through lower scrap and better purchasing, since the physical quantity of material per unit changes far less than labour time.
What resets the curve?
Significant design changes, new equipment, long gaps between production runs and high staff turnover all destroy part of the accumulated experience.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%