Hidden Economics

Learning Curves and the Cost That Falls with Cumulative Volume

Measure your own learning rate from job costing data you already have, then use it to price the work you have not done yet — and see exactly why the cost of the hundredth unit should never be estimated from the cost of the first.

  • Advanced
  • 13 min total
  • 15 chapters

What decision this helps you make: Whether to bid below current cost on volume you expect to learn on, how much early share is worth, and when to stop believing that cost will keep falling.

What this topic is

A learning curve is the observed regularity that unit cost falls by a roughly constant percentage every time cumulative output doubles. Not output per month — output ever. It was first measured on aircraft assembly in the 1930s, has since been documented across manufacturing, construction, energy technologies and services, and it is the reason the tenth installation costs so much less than the first while nothing visible has changed about the work.

Why it matters

Costs that fall with accumulated experience change three things at once. They make early volume worth more than its own margin, because it buys a permanently lower cost base. They make pricing a forward-looking exercise rather than a markup on today's cost. And they mean that a competitor who got ahead early may be structurally cheaper than you in a way no amount of efficiency effort closes. Getting the rate right turns all three from arguments into arithmetic.

Who should learn it

Anyone who quotes repeat work, anyone deciding whether to buy early share, and anyone who has been told that costs will come down with volume and wants to know by how much and on what evidence.

What you will understand

  • The difference between learning, which depends on cumulative output, and scale, which depends on output per period
  • How to measure your own learning rate from records you already have, and how to use it to price forward
  • Why the curve flattens, why doublings get harder, and why learning leaks to competitors
  • What the experience-curve strategy of the 1970s got right, what it got badly wrong, and which half survives

Prerequisites

Common misconception

"We will get cheaper as we grow." That sentence merges two entirely different effects and the merger is where the money is lost. Economies of scale depend on how much you make per period — spread fixed costs over more units this month and unit cost falls, and it rises straight back if volume drops. Learning depends on how much you have ever made, and it does not reverse when this month is slow. One is a function of your current run rate; the other is a function of your history. A business that assumes it has the second when it only has the first prices for a cost position it will never reach.