A market leader watches a clumsy, cheap, clearly inferior product appear at the bottom of its market. It does the sensible thing: ignores it, and keeps serving the customers who pay the most. A few years later that same product has climbed the ladder and eaten the leader's lunch. The maddening part is that nobody was asleep. The incumbent failed because it listened to its best customers and protected its best margins, not in spite of doing so.

The quick version

  • The dilemma: the practices that make a company successful, listening to top customers, chasing the fattest margins, are exactly what make it vulnerable to a cheaper, simpler newcomer.
  • Why: a "disruptive" product starts off worse on the measures the mainstream cares about, so the leader rationally declines to chase it, until it gets good enough and moves upmarket.
  • The move: don't just ask "is this good enough for our customers today?" Ask "who would this be good enough for, and where is it improving faster than the market needs?"
  • The honest caveat: disruption explains some failures cleanly and many others not at all. Use it as a lens, not a law, and never as an excuse to ship something bad and call it strategy.

The idea in depth

The theory comes from Clayton Christensen, a Harvard Business School professor, in his 1997 book The Innovator's Dilemma. Christensen studied the hard-disk-drive industry, a brutal, fast-moving market where generation after generation of dominant firms was wiped out, and found a pattern that repeated across steel, retail and excavators. The incumbents weren't beaten by better technology aimed at their best customers. They were beaten from below.

His central distinction is between two kinds of innovation. Sustaining innovations make an existing product better along the dimensions mainstream customers already value, a faster chip, a sharper screen, a roomier car. Big, well-run companies are superb at these, because their whole machine is tuned to serve demanding customers. Disruptive innovations are different: they start out worse on the metrics the mainstream cares about, but they're cheaper, simpler or more convenient, and they appeal to a fringe, the customers the incumbent is happy to lose, or non-customers who weren't being served at all. The dilemma is that responding to a disruptor looks, on every spreadsheet the incumbent trusts, like a bad idea: lower margins, smaller market, unhappy core customers. So the move is to separate the question "does this serve our best customers?" from the question "is this improving faster than the market's needs are rising?" The second question is the one that kills you.

Why doing everything right is the trap

Christensen's sharpest claim is that the failure is structural, not personal. The same disciplines that earn a leader a promotion, allocate capital to the highest-return projects, listen to your largest accounts, protect gross margin, are precisely the reflexes that steer resources away from a low-margin, low-end product with no proven market. A good manager should turn that project down on the evidence in front of them. Christensen, Michael Raynor and Rory McDonald revisited the theory in What Is Disruptive Innovation? (Harvard Business Review, December 2015), partly to push back on how loosely the word "disruption" had spread. Their tidy example: Uber, they argue, is not a classic disruptor, it didn't start at the low end or among non-consumers, it went straight at mainstream taxi demand. The label, they warned, had become a synonym for "any newcomer that shakes things up," which drains it of any predictive use.

That precision matters for a leader, because the mechanism suggests a counter-move the incumbent's own org chart resists. If your existing business will always (correctly) starve a low-end bet, the answer is often to run that bet somewhere the existing scorecard doesn't apply, a separate unit with its own customers, its own cost base and its own definition of "a win." The fix is structural, not motivational: you can't exhort a high-margin business into loving a low-margin one. You give the small thing a place to grow where it isn't strangled by the big thing's metrics. The honest limitation is that "spin it out" is easy to say and brutal to do well, most corporate skunkworks fail for reasons that have nothing to do with disruption theory, from politics to starved budgets to a parent that reabsorbs the unit the moment it looks promising.

flowchart TD
    A(["Disruptive entrant starts
cheap, simple, 'good enough'
for an overlooked fringe"]) --> B(["Incumbent ignores it,
rationally: low margin,
small market, not our customers"])
    B --> C(["Entrant improves fast,
climbs upmarket"])
    C --> D(["It becomes good enough
for the mainstream"])
    D --> E(["Incumbent's best customers
defect, too late to respond"])
The disruption sequence: the incumbent's rational choices at each step add up to its undoing. Leaders Loop

Where the theory breaks down

Disruption is one of the most-quoted ideas in business, which means it's also one of the most abused, and it has serious critics worth taking seriously. The historian Jill Lepore argued in "The Disruption Machine" (The New Yorker, 2014) that the theory is built on hand-picked cases and reads success backwards, that for every disk-drive firm disruption explains, there are tidy counter-stories where the "disrupted" incumbent did just fine. More pointedly, Andrew King and Baljir Baatartogtokh tested it empirically in "How Useful Is the Theory of Disruptive Innovation?" (MIT Sloan Management Review, 2015): they surveyed industry experts on the 77 cases Christensen used across his books and found that only seven of them, about 9%, actually fit all four elements of the theory. Disruption, they concluded, is a useful idea poorly defined and over-applied.

The practical reading isn't "ignore Christensen." It's that disruption is a pattern that sometimes happens, not a prophecy. Treat it as a diagnostic question, not a forecast. When a cheap, limited entrant appears, ask honestly whether it's improving faster than your customers' needs are, and resist both the reflex to dismiss it and the trendier reflex to panic and call everything disruption. The line a leader should remember: not everything that beats you came from below, and not everything from below will beat you.

A worked example

The figures here are illustrative, but the shape is true to the pattern. Picture a profitable B2B software firm, call it Meridian, selling a powerful analytics suite at roughly £40,000 a year to large enterprises. Its account managers report that customers want more: deeper integrations, more configurability, an AI add-on. Meridian builds them. Margins stay healthy; the flagship customers are delighted. This is textbook sustaining innovation, done well.

Meanwhile a two-person startup ships a stripped-down tool at £40 a month. It does maybe a fifth of what Meridian's product does, and Meridian's sales team laughs it off, rightly, by their numbers: no enterprise would touch it, and chasing that market would mean a product worth a hundredth of the revenue per customer. So Meridian declines. But the cheap tool isn't aimed at Meridian's customers at all; it's serving the thousands of small firms that were never going to pay £40,000, the non-consumers. With that wide, forgiving base, it improves quickly, and within three years it's "good enough" for mid-market buyers, then for the easier enterprise use-cases. Now Meridian's renewals start slipping, and the response that would have been cheap in year one is expensive and late in year four.

What should Meridian have done? Not panic-bought the startup, and not bolted a cheap tier onto a business whose every incentive would smother it. The defensible move is the structural one: stand up a small, separate team with permission to win at the low end on its own economics, measured on adoption and learning, not on quarterly margin, and treat it as cheap insurance against a threat the core business is built to misjudge. As McKinsey's Three Horizons model frames the same instinct: protect today's business and seed tomorrow's at the same time, because the org that's brilliant at one is usually structurally bad at the other.

flowchart LR
    M(["Meridian core:
chase top customers,
protect £40k margin"]) -->|"correct, but
blind to the low end"| R(["Renewals slip
3–4 years later"])
    S(["Separate low-end unit:
win on its own economics,
measured on adoption"]) -->|"cheap insurance"| P(["Option to move upmarket
before the entrant does"])
Two responses to the same threat: the core business is structurally unable to take the low-end bet, so a leader gives it a home of its own. Leaders Loop

Frequently asked questions

Is "disruptive" just a fancy word for "really innovative"?

No, and conflating the two is the most common mistake. In Christensen's strict sense, disruption is a specific path: start at the low end or among non-consumers with something cheaper and simpler, then climb upmarket. A breakthrough that arrives at the high end and serves the best customers first (a faster, pricier flagship) is a sustaining innovation, however dazzling. The distinction is the whole point; without it, "disruptive" just means "new and exciting," which predicts nothing.

If my company is the incumbent, what's the one question to ask?

"Is anything cheaper and worse than us improving faster than our customers' needs are rising?" If a low-end entrant is climbing the quality ladder faster than the mainstream's expectations climb, it will eventually be good enough for your customers, and your high margins are a reason you'll be slow to react, not a moat.

Does the theory actually hold up to evidence?

Partly. It's a genuinely useful lens, but it's been over-applied and the empirical support is thinner than its fame suggests, King and Baatartogtokh found only about 9% of Christensen's own cases fit all four elements of the theory. Treat it as a diagnostic that's sometimes right, not a law that's always right.

Doesn't this just excuse shipping a bad product?

It shouldn't. "Worse on mainstream metrics" is not the same as "bad." A disruptor is better on something its early customers care about, price, simplicity, convenience, even while it's worse on the incumbent's scorecard. If your product is simply worse with no compensating advantage for anyone, that's not disruption; it's a weak product.

Can a big company disrupt itself on purpose?

It can try, but rarely from inside the existing business, the core's metrics will starve any low-margin bet. The usual answer is a separate unit with its own customers, cost base and definition of success. Even then it's hard: most internal ventures die of politics or budget, not of bad strategy.

Related in the Toolkit

Where to go next