When worse beats better

Few terms in management have been worn as smooth as disruption. It now means little more than new and somewhat aggressive, which is a pity, because the theory it came from makes a precise and unsettling claim: that well-run firms lose markets not despite good management but because of it. The claim only works if the words are used carefully, and the place to start is the distinction the word disruption was coined to mark.

Most innovation, including most brilliant innovation, is sustaining. A sustaining innovation improves a product along the dimensions of performance that existing mainstream customers already value: sharper images, faster processors, safer cars, longer range, lower fuel burn. The improvement can be incremental or heroic. Each new generation of airliner is a technical achievement of the highest order, and it is aimed squarely at the customers the manufacturer already serves, judged by the metrics those customers already use. Radical and disruptive are not synonyms, and the difference is not a pedantic one. A breathtaking leap aimed at your best customers is still sustaining.

The empirical record on sustaining competition is unambiguous and often forgotten. Incumbents usually win sustaining battles, even radical ones. Clayton Christensen's study of the disk drive industry, the unglamorous laboratory in which the theory was built, found established firms leading almost every sustaining transition, however technologically demanding. They have the customers, the capital, the engineers and the motivation. When the game is making the existing thing better for the existing buyer, the incumbent is structurally favoured to win it.

The disruptive pattern, first sketched by Joseph Bower and Christensen, rests on an observation about trajectories. Technologies tend to improve faster than customers' ability to absorb the improvement, so products eventually overshoot what the mainstream actually needs. That overshoot opens space beneath the market: room for a product that is worse on the metrics the mainstream prizes but better on a different axis entirely, cheaper, simpler, smaller, more convenient, more accessible.

A disruptive innovation enters through that opening. It does not challenge the incumbent's best products head-on; it begins where the incumbent is not looking, and there are two classic doors. Low-end disruption recruits the incumbent's least profitable customers, the ones being over-served and overcharged. Steel minimills began with reinforcing bar, the bottom of the market, and the integrated mills were briefly more profitable for retreating upmarket and conceding it. Then the minimills' quality improved, tier by tier, and the retreat had nowhere left to go. New-market disruption competes with non-consumption instead, serving people who previously consumed nothing at all. The transistor radio sounded terrible by the standards of a living-room set, and it was not competing with living-room sets; it was competing with silence, for teenagers who could never have afforded the alternative. The personal computer played the same trick on the minicomputer.

In both variants the foothold looks like a toy or a bad business, and that is the camouflage. The entrant's product then improves along the conventional metrics, riding the same steep trajectory, until one day it is good enough for the mainstream, at a cost structure the incumbent cannot match.

The uncomfortable heart of the theory is that the incumbent's failure is produced by its strengths. Inside any well-managed firm sits a resource allocation process designed to do exactly what good management requires: consult the best customers, compare gross margins, size the addressable market, and promote the managers who back winners. A disruptive opportunity fails every one of those tests. The best customers do not want it. The margins are worse. The market is small or, in the case of non-consumption, literally unmeasurable, since no one can size a market that does not yet exist. And no ambitious manager volunteers to champion a low-margin toy when the alternative is a flagship line extension.

So the proposal is not rejected by stupidity or complacency. It is filtered out by the same machinery that makes the firm well run, decision after defensible decision. Meanwhile the asymmetry of motivation does its quiet work: the margins the incumbent is structured to flee are the margins the entrant is delighted to win. By the time the trajectories cross and the entrant's product satisfies the mainstream, the newcomer has scale, a hardened cost structure and improving technology, and the incumbent's response is a crash programme fighting its own accounting. That is the innovator's dilemma, properly stated: every individual choice was rational, and the sum of them was fatal.

The definitional discipline matters because the prescriptions differ. Not every startup is disruptive, and not every breakthrough is. Christensen himself argued that Uber was not, in the strict sense, a disruption of the taxi industry: it began by serving mainstream customers with a better and often cheaper service, a sustaining improvement delivered by an entrant. Tesla entered from the top of the market, selling performance to the affluent before working downwards, which is the opposite of a low-end foothold. These may be excellent businesses and genuine threats, but they travel a different road, and an incumbent watching only the low end will be blindsided by them precisely because they do not fit the disruptive script.

A second confusion is subtler. Rebecca Henderson and Kim Clark showed that incumbents also stumble over architectural innovation: changes in how familiar components connect, with no new science and no low-end attack anywhere in sight. The knowledge of an architecture gets embedded in a firm's structure and communication channels, so when the architecture shifts, the organisation keeps answering questions nobody is asking. Filing such failures under disruption obscures a different disease with a different treatment. Disruption is a story about customers and money. Architectural failure is a story about organisational knowledge. A firm can die of either while inoculated against the other.

The prescriptions follow from the mechanism. Because the filter that rejects disruptive bets is the resource allocation process itself, the one response with a consistent record is structural: place the disruptive venture in a separate unit with its own cost structure, its own metrics and its own customers, free to be excited by margins the parent finds embarrassing, as IBM did when it built the PC business in Boca Raton, far from the mainframe culture. The separate unit must find its own foothold customers rather than borrowing the parent's, because borrowed customers reimpose the old filter through the back door.

Watching the right indicators matters as much as structure. Disruption announces itself in places conventional dashboards ignore: share among non-consumers rather than share among current customers, adoption at the low end, the widening gap between what products offer and what buyers actually use, customers declining upgrades because last year's version was already more than enough. The jobs-to-be-done lens, asking what progress a customer is hiring a product to make, is the most practical tool for spotting non-consumption, since people without a product still have the job.

And none of this licenses neglecting the sustaining game. The firm must keep winning the present while it options the future, funding both kinds of work under different rules and resisting the temptation to judge either by the other's metrics. That balance, exploration and exploitation held in deliberate tension, is the same ambidexterity that runs through every serious account of organisational renewal, wearing different clothes.

The comparison is not a contest in which disruption is the glamorous winner. Sustaining innovation pays today's bills, and most of the time it is exactly the right thing to do, which is precisely what makes disruption so hard to see from inside: the warning signs are indistinguishable from prudence. Sustaining and disruptive innovation are different games, scored by different customers on different metrics, and they punish different mistakes. The firms that endure are not the ones that pick the fashionable game. They are the ones that can tell, before the trajectory lines cross, which game they have been invited to play.