The Experience Curve
The experience curve is an idea developed by the Boston Consulting Group in the mid-1960s. Working with a leading manufacturer of semiconductors, the consultants noticed that the company’s unit cost of manufacturing fell by about 25% for each doubling of the volume that it produced. This relationship it called the experience curve: the more experience a firm has in producing a particular product, the lower are its costs. Bruce Henderson, the founder of bcg, put it as follows: Costs characteristically decline by 20–30% in real terms each time accumulated experience doubles. This means that when infiation is factored out, costs should always decline.
The decline is fast if growth is fast and slow if growth is slow. There is no fundamental economic law that can predict the existence of the experience curve, even though the curve has been shown to apply to all industries across the board. Its truth has been proven inductively, not deductively.
By itself, the curve is not particularly earth shattering. Even when bcg first expounded the relationship, it had been known since the second world war that it applied to direct labour costs. Less labour was needed for a given output depending on the experience of that labour. In aircraft production, for instance, labour input decreased by some 10–15% for every doubling of that labour’s experience.
The strategic implications of the experience curve came closer to shattering earth. For if costs fell (fairly predictably) with experience, and if experience was closely related to market share (as it seemed it must be), then the competitor with the biggest market share was going to have a big cost advantage over its rivals. qed: being market leader is a valuable asset that a firm relinquishes at its peril.
This was the logic underpinning the idea of the growth share matrix. It justified allocating financial resources to those businesses (out of a firm’s portfolio of businesses) that were (or were going to be) market leaders in their particular sectors. To do this, of course, implied starvation for those businesses that were not and never would be. Over time, managers came to see the experience curve as being too imprecise to help them much with specific business plans.
Once the strategic implications of the general principle had been taken on board, there seemed little to be gained from pursuing it any further. Inconveniently, different products had curves of a different slope and different sources of cost reduction. They did not, for instance, all have the same downward gradient as the semiconductor industry. A study by the Rand Corporation found that “a doubling in the number of reactors results in a 5% reduction in both construction time and capital cost”. Part of the explanation for this discrepancy was that different products provided different opportunities to gain experience. Large products (such as nuclear reactors) are inherently bound to be produced in smaller volumes than small products (such as semiconductors).
It is not easy for a firm to double the volume of production of something that it takes over five years to build, and where the total market may never be more than a few hundred units. In theory, the experience curve should make it difficult for new entrants to challenge firms with a substantial market share. In practice, new firms enter old industries all the time, and before long many of them become major players in their markets.
This is often because they have found ways of bypassing what might seem like the remorseless inevitability of the curve and its slope. For example, experience can be gained not only first-hand, by actually doing the production and finding out for yourself, but also second-hand, by reading about it and by being trained by people who do have experience. Furthermore, firms can leapfrog over the experience curve by means of innovation and invention. All the experience in the world in making black and white televisions is worthless if everyone wants to buy colour sets.