Here is the puzzle at the heart of this topic: well-run companies, with smart leaders and loyal customers, are repeatedly beaten by products that are worse than their own. Not worse-but-cheaper as a quirk, worse on every measure their best customers care about. The companies see the threat coming. They have the resources to crush it. And they lose anyway. Understanding why is the difference between managing innovation and being managed by it.

The quick version

  • Sustaining innovation makes a good product better for your existing, profitable customers. Incumbents almost always win at this.
  • Disruptive innovation starts cheaper and simpler, serves the least-attractive customers (or non-customers) first, then improves until it's good enough for the mainstream. Incumbents almost always lose to this.
  • The trap is that ignoring disruption is the rational choice in the moment, the disruptor's market looks too small and too unprofitable to bother with.
  • Disruption is a process over years, not a single product launch. The move is to watch the trajectory, not the snapshot, and to fight it with a separate team, not your core business.

The idea in depth: two kinds of better

The vocabulary comes from Clayton Christensen, the Harvard Business School professor whose 1997 book The Innovator's Dilemma made "disruption" a boardroom word, and who spent the next two decades trying to stop people from misusing it. In a 2015 Harvard Business Review article with Michael Raynor and Rory McDonald, he drew the line cleanly. Sustaining innovations improve a product along the dimensions your mainstream customers already value. They can be small tweaks or major breakthroughs; what matters is that they help you sell better products to your most profitable customers. Disruptive innovations do something stranger. They take root at the bottom of a market, among customers your best product over-serves, or people who weren't customers at all, by being cheaper, simpler, or more convenient. They look unimpressive. Then they get better.

The critical word in the Christensen Institute's own definition is process: disruption is "a process by which a product or service takes root in simple applications at the bottom of the market … and then relentlessly moves upmarket, eventually displacing established competitors." It is not an event. The disruptor doesn't arrive fully formed and win on day one, it wins on day three thousand, by which point your escape routes have closed. So judge a competitor by its rate of improvement, not its current quality. A product that is bad today but getting better twice as fast as yours is a different animal from one that is bad and static.

flowchart LR
  A(["Performance over time"]) --> B(["Sustaining path:
incumbent improves
for best customers"])
  A --> C(["Disruptive path:
entrant starts below
'good enough', climbs fast"])
  B --> D(["Overshoot:
more than
customers can use"])
  C --> E(["Catches the mainstream
from below → incumbent
displaced"])
Two trajectories on one chart: the incumbent climbing past what customers need, the disruptor climbing up to meet them. Leaders Loop

Why good management makes it worse

The cruelty of the dilemma is that the incumbent's defeat is caused by doing everything right. Listen to your best customers; invest where the margins are highest; chase the bigger, more profitable market, this is textbook good management, and it is exactly the behaviour that makes you blind to disruption. The disruptor's early market is small, low-margin, and full of customers your sales team would rather not have. Every quarterly incentive points your resources away from it. By the time the disruptive product is good enough to threaten your core, the disruptor has years of cost advantage and learning you can't replicate from a standing start.

The canonical case from The Innovator's Dilemma is steel. In the 1960s, "minimills" using cheap electric-arc furnaces could only make low-grade rebar, the lowest-quality, lowest-margin steel there was. Integrated giants like Bethlehem Steel were happy to surrender that segment; it dragged down their averages. The minimills, led by Nucor, took the bottom, used the cash to improve, and marched upmarket into sheet steel. By 2014 Nucor was the largest steel producer in the United States and several integrated incumbents had gone bankrupt, each retreat upmarket a locally sensible decision that summed to collective defeat. Here's the practical tell: treat "we're happy to cede the low end" as an alarm, not a strategy. When you find yourself relieved to lose a segment, ask who is delighted to win it, and what they'll do with the cash flow it throws off.

"Disruptive innovations … originate in low-end or new-market footholds.", Christensen, Raynor & McDonald, HBR, 2015

Those two entry points matter for spotting a threat. A low-end foothold attacks the bottom of an existing market, the over-served customers paying for performance they don't use (minimills, budget airlines, basic legal-document software). A new-market foothold turns non-consumers into customers by making something so cheap or simple that people who previously went without can now buy, the transistor radio for teenagers who couldn't afford a tabletop set; the personal computer for people who'd never touch a mainframe. So scan both flanks: who is being over-served and would happily pay less for less, and who is shut out of your market entirely because your product is too expensive or too complicated?

flowchart TD
  Q(["A new competitor appears"]) --> A{"Does it serve
over-served customers
at lower cost?"}
  A -->|Yes| L(["Low-end foothold,
watch its upmarket climb"])
  A -->|No| B{"Does it turn
non-customers into
customers?"}
  B -->|Yes| N(["New-market foothold,
watch it pull your
mainstream across"])
  B -->|No| S(["Probably a sustaining
move, respond head-on
with your strengths"])
A quick triage. Not every threat is disruptive, and treating a sustaining attacker as a disruptor (or vice versa) leads to the wrong defence. Leaders Loop

Where the theory breaks down (be honest about this)

Disruption is a powerful lens, not a law of physics, and the most useful thing a leader can know is where it stops predicting well. In a 2015 MIT Sloan Management Review study, Andrew King and Baljir Baatartogtokh surveyed and interviewed experts on 77 cases Christensen and Raynor had cited as disruption. Only seven of those cases, about 9%, turned out to contain all four of the theory's own key elements (incumbents improving, then overshooting customer needs, the entrant starting at the low end and improving, and the incumbent then floundering). Their conclusion was measured: don't throw the theory out, but use its best parts alongside classical strategic analysis rather than as a one-size explanation. The theory's enduring value, they noted, is as a warning against managerial myopia, the standing temptation to dismiss a small, ugly competitor.

Christensen himself spent years correcting misuse. Calling every successful startup "disruptive" is the common error: in the 2015 HBR piece, he argued Uber is not a classic disruptive innovation, because it didn't start among over-served or non-consuming customers, it went straight for the mainstream taxi market with a better product. That might still be a brilliant business; it's just not "disruption" in the technical sense, and the distinction changes how incumbents should respond. The point of the label is diagnostic, not a compliment. The question is never "is this disruptive?" as a badge, it's "which playbook does this threat call for, sustaining or disruptive?", because the two demand opposite responses.

A worked example

Illustrative figures throughout, the company is a composite, the numbers are made up to show the mechanics.

Imagine you run a professional-grade video-editing suite. You sell to studios for £2,000 a seat, your software does everything, and your roadmap is a wall of features your power users have requested. That roadmap is pure sustaining innovation, better tools for your best, most profitable customers. It's working: revenue is up, churn is low.

Meanwhile a free app lets teenagers cut clips on a phone. It's laughably limited, no colour grading, no multi-track audio, a tiny screen. Your studio customers laugh at it, and so do you. There is no rational case to chase those users: they pay nothing, and serving them would distract from a roadmap that's making money. Every instinct says ignore it. That relief is the alarm.

Two years on, the phone app has added layers, transitions, and AI auto-editing. It still isn't studio-grade, but it's now "good enough" for the corporate marketing teams and small agencies who were paying you £2,000 and using a tenth of the product. They defect for something that costs £15 a month and is good enough. You've lost the bottom of your market, your per-seat average looks fine because only your demanding customers remain, and the app is still climbing. The move you wish you'd made: spin up a small, separate team, its own P&L, free to cannibalise you, to build a "good enough, dead simple" product for the customers you were over-serving, while the app's quality was still beneath you. Disruption is fought early and from the side, never late and head-on.

Frequently asked questions

Is disruptive innovation always better than sustaining innovation?

No, and this is the most common misreading. Most of the money most companies make, most years, comes from sustaining innovation: making a good product better for customers who already pay you. Incumbents reliably win at it. Disruption matters not because it's superior but because it's the threat your normal management instincts won't let you see. You need both playbooks; you just need to know which situation you're in.

How do I tell a disruptive threat from ordinary competition?

Ask three questions. Did it start by serving your least attractive customers, or people who weren't customers at all? Is it cheaper or simpler rather than better? And is its rate of improvement faster than yours? Three yeses suggest disruption, respond by setting up a separate unit, not by bolting features onto your core. A head-on, "better product, same customers" attacker is sustaining competition, and you fight that with your existing strengths.

Why can't a big company just copy the disruptor?

It usually can technically, but not organisationally. The disruptor's market is small and low-margin, so it loses every internal fight for resources against the profitable core. By the time the numbers justify acting, the disruptor has years of cost and learning advantage. This is why the standard prescription is an autonomous team with its own targets, insulated from the core's metrics, and allowed to undercut it.

Does the theory still hold up to scrutiny?

Partly. The King and Baatartogtokh study found that only a small share of the cases Christensen cited fit all four elements of the theory, far fewer than its popularity implies, and the term gets stretched well beyond its meaning. Treat it as a sharp lens for one specific failure pattern, incumbents dismissing a low-end or new-market entrant, not as a universal theory of why companies win or lose. Pair it with conventional strategy analysis.

Is "disruptor" just a marketing label now?

Often, yes, which is exactly why precision pays. Christensen argued that even Uber, frequently called disruptive, doesn't fit the technical definition because it targeted the mainstream market directly. Using the word as a badge tells you nothing useful. Using it as a diagnosis, which foothold, which response, tells you what to do.

Related in the Toolkit

Where to go next