Key Insights: The Innovator's Dilemma
Author: Clayton M. Christensen • Category: Business & Strategy
Christensen's puzzle is not why weak companies fail but why strong ones do. The firms he studied were well managed by every conventional measure, and several built working versions of the technology that displaced them. They lost anyway: the disciplines that made them excellent — serving demanding customers, protecting margins, funding only what the numbers justify — pointed them away from the threat.
5 Pillars of Success
1. Sustaining versus disruptive innovation A sustaining innovation improves what existing customers pay for: capacity, speed, reliability, features. Incumbents usually win those races: their processes and customers are built for them. A disruptive innovation is worse on those dimensions but cheaper, simpler or more convenient, so it first sells to buyers the leaders ignore. Christensen traces disk drives from 14 inches down through 8, 5.25, 3.5, 2.5 and 1.8. Incumbents led the sustaining advances; entrants led the early years of each smaller format. Seagate's engineers built a 3.5-inch drive; its customers wanted more capacity in the existing format, so a product the company could easily make was shelved.
2. Two trajectories and the shift in buying criteria Two curves move at different speeds: technology improves faster than customers can absorb, so products become good enough, then better than good enough. Once performance overshoots what most buyers can use, the basis of competition changes. In disk drives the sequence ran from capacity and functionality, to reliability and size, to convenience, then to price. Each shift changes which architecture and cost structure can compete. Managers keep improving the attribute that used to close the sale, while the attribute that now decides it is developed elsewhere.
3. Value networks and asymmetric motivation A value network — the customers, suppliers, competitors and cost structures a firm lives inside — decides which problems look worth solving. Demanding customers pull resources toward higher performance and higher margin, and cost structures belong to the network: a company built for the high end cannot earn acceptable margins at a low-end price, so the low end looks like bad business rather than an opportunity. Hydraulic excavators were weak but handy for small residential jobs like drainage and sewer work. Cable-shovel makers knew about hydraulics, but their contractor customers had no use for them. The winners came from the small-equipment network, whose buyers wanted compact, cheap machines.
4. Markets that do not exist cannot be analyzed A disruptive market cannot be sized, because the applications do not exist yet and the eventual customers cannot describe what they will want. Forecasts for the new disk drive formats were repeatedly wrong, and customer research mostly asks for more of what already sells. Christensen's alternative is discovery-based planning: write down the assumptions a plan rests on, test them cheaply and in sequence, and let strategy emerge from the tests. The electric vehicle case at the end is judged the same way: does the new trajectory close the gap faster than mainstream requirements rise?
5. Give the project a home that needs it Capabilities live in resources, processes and values, and are not equally portable. People and money move easily; processes fit the business you run; values — the margin and size thresholds for funding — are least flexible. A small market cannot satisfy a large company's growth needs: what would define a start-up is a rounding error there. Assign the project to a unit whose own customers need it, small enough that winning matters there. IBM's personal computer is the standard case: an autonomous Boca Raton unit that bought its microprocessor from Intel and its operating system from Microsoft, won the market quickly, and watched much of the profit pool migrate to those suppliers.
Key Takeaway
"Doing everything right according to traditional management principles is precisely why market leaders fail."
The failure is structural, not personal: customers, cost structure and funding criteria steer a well-run company away from the disruption that eventually matters.
Practical Task
1. **Test the label.** Decide whether the entrant is worse on what your best customers pay for. If it is, treat it as disruptive; if it is better, your normal process applies. 2. **Plot both rates.** Estimate how fast the entrant is improving versus how fast your customers' requirements rise. Act where the lines cross, not on today's gap. 3. **Map the value network.** List the customers, margins and cost assumptions the project would violate. A low-end price that cannot clear your cost structure is the mechanism to design around. 4. **Replace forecasts with tests.** Name the three assumptions the plan rests on, then design the cheapest experiment that could prove each wrong, within a spend you can afford to lose. 5. **Name the owner.** If winning a small, low-margin market would make nobody's targets, no mandate will save it: give it its own unit and thresholds. 6. **Schedule the review.** Disruption is a question about rates, so revisit the two trajectories quarterly instead of re-arguing the label.