Technology S-Curve

Thinking Tools3.4 Facts About the FutureAlso known as S-Curve Analysis, Technology Evolution Analysis

Developed by Richard N. Foster (Innovation: The Attacker's Advantage, 1986), with roots in Genrich Altshuller's laws of technical evolution (Creativity as an Exact Science, 1979) and extended by Clayton M. Christensen (The Innovator's Dilemma, 1997).

Plots a technology's performance against effort or time to place it on its S-shaped curve, and spots the successor rising before the incumbent tops out.

A battery chemistry near its ceiling, and a successor climbing its own curve

Performance per dollar, with research spending steady so that years stand in for effort; year 5 is this year. The black curve is the current chemistry: it rose fast and is now flattening near its ceiling, so each year's work buys less. The colored curve is the successor, set by its ceiling, how fast it improves, and the year of its fastest gains. Where it crosses the black curve, the next dollar of research buys more on the new curve; that is the time to have moved. Illustrative.

200
0.7
year 7
Current chemistry: gain from year 5 to year 6
1.1
Successor: gain from year 5 to year 6
26.8
Successor in year 10
178
First year the successor leads (11: not within ten years)
year 7

Successor, performance per dollar by yearCurrent chemistry

0501001502001234567891005010015020012345678910

An example

A maker of cordless power tools spent most of its battery research budget improving the cell chemistry it had used for a decade. The engineering director plotted the team's results year by year against the same steady spending. The early years had each added five to seven points of performance per dollar; the last two had added under four points and then about two. The curve was flattening toward a ceiling set by the chemistry itself.

A newer chemistry was still well behind, at about 40 percent of the current performance, and sales saw no reason to look at it. Its curve was the problem: it was gaining more than 20 points a year and accelerating, with a ceiling the lab put at about twice the current one. On those estimates the newer chemistry would pass the current one in about two years.

The director moved a third of the battery budget to the new chemistry that quarter, kept a small team on the old one for cost reductions, and set a review of both curves every six months.

Why it works

A technology's performance does not improve at a constant rate for the same effort. Early on, gains come slowly while the basics are worked out; then they come fast as the approach matures; then they slow as the technology nears a physical or practical limit, so each further dollar buys less. Plotting results against effort shows where on that curve a technology sits, which a chart of this year's performance alone cannot.

The second insight is that successors start below. A new approach usually performs worse than the mature one when it appears, so it looks unpromising to anyone comparing today's numbers. What matters is the slope and the ceiling. Firms that wait for the newcomer to beat the incumbent on today's measures arrive after the attackers who moved when the old curve flattened.

The thinking behind it

Richard N. Foster Richard N. Foster, a McKinsey director, argued in Innovation: The Attacker's Advantage (1986) that a technology's performance improves along an S-curve, slowly, then quickly, then slowly again as it nears its limits, and that incumbents lose to attackers who move early to the next curve while the incumbents are still investing in the old one. It is a practitioner's synthesis drawn from McKinsey's client work.

Genrich Altshuller Genrich Altshuller, an engineer and inventor, created TRIZ, the theory of inventive problem solving, from the study of thousands of patents, set out in works including Creativity as an Exact Science (1979). TRIZ holds that technical systems evolve by knowable patterns, which is the root of the idea that a technology's path can be read and its successors anticipated before they arrive.

Clayton M. Christensen Clayton M. Christensen, a Harvard Business School professor, tested the S-curve against the history of disk drives in Exploring the Limits of the Technology S-Curve (1992), in Production and Operations Management, and found that the flattening came at different points for different firms in the same industry. The Innovator's Dilemma (1997) extended the argument into his theory of disruptive innovation, in which well-run incumbents lose to simpler, cheaper technologies that start in markets their best customers do not value.

How to run it

  1. Pick the performance measure. Choose the measure customers pay for, such as performance per dollar, energy per kilogram, or cost per unit.
  2. Plot it against effort. Chart results against cumulative research spending or engineering time, not just calendar years.
  3. Read the slope. Compare recent gains per unit of effort with earlier ones; shrinking gains mean the curve is flattening.
  4. Estimate the ceiling. Ask what physical or practical limit the approach is nearing, and how close the current results are to it.
  5. Find the challengers. List the newer approaches, however far behind, and plot their early results the same way.
  6. Project the crossing. Estimate when a challenger's curve passes the incumbent's, under low and high assumptions.
  7. Shift effort before the crossing. Move part of the budget to the new curve while the old still leads, and review both regularly.

Where it goes wrong

Years are plotted in place of effort. A curve that flattens because spending was cut looks like a technology at its limit; spot it when the slowdown coincides with a budget change. Plot results against cumulative effort.

The ceiling is guessed by the incumbent's champions. Teams invested in the current approach tend to place its limit far away. Have the ceiling estimated by people with no stake in either curve.

Every new approach is treated as the successor. Most challengers never reach the incumbent's performance. Track several, judge each on its early slope and its credible ceiling, and move money in steps as the evidence firms up.

When to use something else

  • Three Horizons. When the question is how to split effort across today's business, the next one, and longer bets.
  • Hype Cycle. When the noise around a new technology needs reading apart from its real progress.
  • Real Options Analysis. When the move to a successor can be staged as a series of small, reversible bets.

Where it sits in SOLVE eX

Go to the source

  • Innovation: The Attacker's Advantage, 1986
  • Creativity as an Exact Science, 1979
  • The Innovator's Dilemma, 1997

Back to the catalog