Experience Curve Analysis
- Introduced by
- Bruce D. Henderson (Boston Consulting Group); learning-curve tradition (T. P. Wright; Louis Yelle)
- Year
- 1979
- Type
- Cost-projection pipeline linking cumulative volume to unit cost
- Family
- Strategic management / cost dynamics
What does it do?
Experience curve analysis describes and projects how the real unit cost of a product falls by a roughly constant percentage every time cumulative production volume doubles, and draws the strategic consequences for cost position and pricing.
The Boston Consulting Group, under Bruce Henderson, generalized the older manufacturing learning curve in the late 1960s and 1970s into the broader 'experience curve,' covering not just direct labor but all value-added costs, and made it the analytical backbone of its strategy advice — including the growth-share matrix's premise that the high-relative-share firm enjoys a cost advantage. Louis Yelle's 1979 Decision Sciences survey reviewed the underlying learning-curve literature and its mathematics, while Barry Hedley's 1977 article tied the experience-curve cost logic to portfolio strategy. The method fits a power-law relationship between unit cost and accumulated volume and uses the estimated learning rate to forecast costs and inform competitive strategy.
The intuition behind the method
The more times an organization makes something, the better and cheaper it gets at making it — and remarkably, this improvement tends to follow a stable rule: each doubling of total cumulative output knocks a fixed percentage off the real cost per unit. An '80% experience curve' means costs fall to 80% of their level every time cumulative volume doubles, a 20% reduction per doubling. The strategic punch line is about who has made the most. The firm with the largest cumulative volume has slid furthest down the curve and therefore has the lowest costs, which is why market-share leadership and cost leadership tend to travel together. Experience curve analysis estimates how steep a given product's curve is, projects where costs are heading, and helps a firm decide whether to price aggressively to build the volume that secures a durable cost advantage.
History
The experience curve descends from the manufacturing learning curve documented by T. P. Wright in 1936 for aircraft assembly, where each doubling of cumulative output reduced labor hours by a stable percentage. In the 1960s and early 1970s the Boston Consulting Group, led by Bruce Henderson, broadened this into the 'experience curve,' extending the regularity from direct labor to all value-added costs and elevating it into a general theory of competitive cost dynamics. BCG made it the foundation of much of its strategic advice, and Barry Hedley's 1977 Long Range Planning article wove the experience-curve cost logic into the growth-share portfolio framework. Louis Yelle's 1979 Decision Sciences article surveyed the accumulated learning-curve literature and its mathematics, consolidating the analytical tradition on which strategic experience-curve analysis rests.
When to use it
Use experience curve analysis when unit costs in a business decline predictably with accumulated production and the strategic question concerns cost position, pricing, or the value of market-share leadership. It fits products and processes — often in manufacturing or high-volume, technology-intensive industries — where there is a meaningful history of cumulative volume and real unit cost, and where learning, scale, and process improvement plausibly drive costs down. It is valuable for setting penetration prices to build volume, for forecasting cost trajectories, and for assessing whether a share leader has a defensible cost advantage. It is less reliable where costs are dominated by volatile input prices rather than learning, where products change too fast for cumulative experience to matter, or where the curve is extrapolated naively beyond the range supported by data.
Limitations and considerations
- The relationship is empirical, not guaranteed; cost declines require active management and are not automatic with volume.
- Extrapolating the curve beyond observed volumes is risky, since the rate can change as technology or processes mature.
- It conflates several distinct cost drivers — learning, scale, and innovation — into one cumulative-volume relationship.
- Cost data attributable to experience are hard to isolate from input-price changes, inflation, and accounting choices.
Source conditions do not replace independent scientific review or validation by an analysis engine.
Strengths
- Captures a robust empirical regularity — a roughly constant percentage cost decline per doubling of cumulative volume — in a single parameter.
- Connects cost dynamics directly to competitive strategy, explaining why share leadership often confers a cost advantage.
- Supports forward-looking pricing and capacity decisions by projecting future unit costs from accumulated experience.
- Generalizes the manufacturing learning curve to all value-added costs, broadening its strategic relevance.
Applications
- Forecasting future unit costs for a product as cumulative production accumulates.
- Setting penetration pricing to build the volume needed to secure a low-cost position.
- Assessing whether a market-share leader holds a defensible cost advantage from accumulated experience.
- Comparing the learning rates of products, processes, or competitors to gauge cost-improvement potential.
- Underpinning portfolio analysis where cumulative-volume cost advantages justify investing in share growth.
How it works
How it works
The analysis is driven not by time or annual output but by cumulative experience, so the first step is to accumulate total volume produced up to each point. Yelle's survey emphasizes that the learning and experience curves relate cost to this running total of units ever made, since it is accumulated practice — not the calendar — that drives improvement. Constructing the cumulative-volume series from period production data sets the independent variable on which the cost decline is measured and distinguishes the experience curve from ordinary time-based cost trends.
Formula
Q_t = \sum_{\tau\le t} q_\tau,\quad q_\tau=\text{units produced in period } \tauUnit cost is then modeled as a power-law function of cumulative volume: cost per unit equals the cost of the first unit times cumulative volume raised to a negative exponent. Yelle documents this canonical functional form from the learning-curve tradition, in which the exponent encodes how fast costs fall. The power law captures the central empirical regularity — a constant percentage cost reduction per doubling of output — and provides the model whose parameters the rest of the analysis estimates and interprets.
Formula
C_q = C_1\, q^{-b},\quad b=-\log_2(\phi),\ \ 0<\phi<1\ (\text{learning rate})Because the power law becomes linear in logarithms, the curve is estimated by taking logs of unit cost and cumulative volume and fitting a straight line, whose slope is the negative of the experience exponent. Yelle's review describes this log-linear estimation as the standard empirical procedure for learning and experience curves. Plotting log cost against log cumulative volume should yield a near-linear relationship if the experience curve holds, and the fitted slope gives a direct, data-based estimate of how steeply costs decline with accumulated experience.
Formula
\ln C_q = \ln C_1 - b\,\ln q + \varepsilon\ \ \Rightarrow\ \text{OLS slope } = -bFrom the estimated exponent the analyst derives the learning (or progress) rate — the ratio of unit cost after a doubling to unit cost before it. An exponent corresponding to a rate of 0.8, for example, means costs fall to 80% of their prior level with each doubling of cumulative volume, an 80% experience curve. Yelle's survey catalogs typical rates and their interpretation. This single number summarizes the steepness of the curve, makes products and processes comparable, and is the key parameter for projecting future costs.
Formula
\frac{C_{2q}}{C_q}=2^{-b}=\phi\ \ (\text{e.g. } \phi=0.8\Rightarrow 20\%\ \text{drop per doubling})Finally the fitted curve is used to project costs at future cumulative volumes and to draw strategic conclusions. Extrapolating the power law forecasts the unit cost the firm will reach as volume accumulates, informing pricing and capacity decisions. Hedley's portfolio logic builds directly on this: because cost falls with cumulative volume, the firm with the highest relative market share has typically produced the most and holds the lowest-cost position, which is why the experience curve underpins the strategic value of share leadership and the growth-share matrix's cash-generation premise. The projection thus links cost dynamics to competitive strategy.
Formula
\hat C(Q_{T+h})=C_1 Q_{T+h}^{-b};\ \ \text{leader's } RMS\Rightarrow \text{cumulative-volume cost edge}Common pitfalls
- Confusing the time-based cost trend with the experience curve, which is defined against cumulative volume, not the calendar.
- Failing to deflate costs, so that inflation or input-price swings masquerade as movement along the curve.
- Naively extrapolating a fitted curve far past the data range and assuming the learning rate stays constant indefinitely.
- Treating cost decline as automatic and under-investing in the process improvement that actually realizes it.
Frequently asked questions
What is the difference between a learning curve and an experience curve?
They share the same mathematics but differ in scope. The classic learning curve, traced to Wright's 1936 aircraft study and surveyed by Yelle, relates direct labor hours per unit to cumulative output, capturing worker and process learning. BCG's experience curve generalizes the idea to all value-added costs — not just labor but materials handling, overhead, marketing, and more — so it describes the total unit cost of a product as a function of cumulative volume. The experience curve is thus the broader, strategy-oriented extension of the narrower production learning curve.
What does an '80% experience curve' mean?
It refers to the learning rate. An 80% curve means that every time cumulative production volume doubles, the real unit cost falls to 80% of its previous level — a 20% reduction per doubling. The percentage is the progress ratio derived from the curve's exponent: a lower percentage indicates a steeper curve and faster cost decline. Estimating this rate from log-linear regression of cost on cumulative volume is the central quantitative output of experience curve analysis, since it summarizes how quickly costs improve with accumulated experience and supports cost projections.
Why does the experience curve support market-share leadership as a strategy?
Because costs fall with cumulative volume, the firm that has produced the most has descended furthest down the curve and holds the lowest-cost position. Market-share leadership generally means higher cumulative volume, so the leader tends to enjoy a structural cost advantage that is hard for smaller rivals to match. Hedley's portfolio logic builds on exactly this: it justifies pricing aggressively to gain share and investing in high-share businesses, and it underlies the growth-share matrix's premise that high relative share equals strong cash generation. The advantage is real only where the curve is steep and managed, not automatic.
Related methods
Sources
Yelle, L. E. (1979). The Learning Curve: Historical Review and Comprehensive Survey. Decision Sciences, 10(2), 302-328.
Open DOI referenceCopied.Hedley, B. (1977). Strategy and the 'Business Portfolio'. Long Range Planning, 10(1), 9-15.
Open DOI referenceCopied.