Somewhere in a boardroom in 2019, a team of very well-paid people looked at a prototype with a plastic lens, a cheap sensor, and a screen that would not impress anyone, and they said: no. Not our customer. Not our margin. Not our problem.
That decision is the entire story of disruptive innovation. Not a technology problem. A decision problem.
I have watched this play out inside companies I worked with, and the pattern is so consistent it is almost boring. The disruption rarely wins because the incumbent is stupid. It wins because the incumbent is rational — rationally defending the wrong thing at exactly the wrong moment. And in 2026, with AI eating entire workflows and healthcare pricing collapsing from the bottom, the pattern is running faster than Christensen ever described it.
Key takeaways
- Disruption starts at the bottom of a market, not at the top — cheap, worse, and ignored by the leaders who own the profitable segment.
- The incumbent's refusal is usually correct in the short term. That is what makes it lethal.
- Modern examples live in healthcare, finance, mobility and energy — and most of them were dismissed as toys or charity projects.
- The theory has real limits. Some "disruptions" never arrive, and Christensen's own framework has been criticised for retrofitting history.
- You can spot a live disruption by tracking one number: the cost trajectory of the cheap entrant, not its current quality.
What "disruptive innovation" actually means (and what it doesn't)
Clayton Christensen, the Harvard professor who built the framework, drew a line most people blur. A sustaining innovation makes a good product better for existing customers: a faster chip, a sharper camera, a slightly cheaper insurance policy. A disruptive innovation starts worse on the metrics that matter to the incumbent's best customers, sells to people the incumbent doesn't want, and then improves until it takes the whole market.
The first automobiles were not disruptive. They were expensive luxury objects for wealthy enthusiasts. The Ford Model T became disruptive when it got cheap enough to serve people who had never owned a car at all.
The distinction matters because it explains why disruption is so hard to see from inside a successful company. Your best customers are telling you what they want. Your spreadsheets are telling you where the margin is. Both of them are pointing away from the thing that will eventually kill you.
The two types of disruption
Christensen's framework splits into two mechanisms, and I find this far more useful than the vague "Uber disrupted taxis" version you see everywhere.
- Low-end disruption. The entrant takes the least profitable customers — the ones your sales team actively avoids — and serves them adequately. Then it climbs.
- New-market disruption. The entrant creates customers who did not exist before, because the existing solution was too expensive, too complex, or too inaccessible.
- And a third case people forget: disruption that never happens at all, because the incumbent buys the entrant, or because the technology plateaus below the threshold customers actually need.
That last point is where most analyses stop being honest.
Disruptive innovation examples across modern industries
Here is where I want to be careful, because a list of company names is not an explanation. Anyone can list names. The useful question is: what was the entry point, who was excluded, and what did the incumbent do when it noticed?
Disruptive innovation examples in healthcare
Retail clinics are the cleanest new-market disruption in modern healthcare. Traditional primary care is built around appointments, buildings, and insurance billing cycles. The walk-in clinic inside a pharmacy served people who had no regular doctor, no time, and a minor problem — a strep test, a vaccine, an ear infection. It was objectively worse care for anything complicated. But it created demand that simply was not being served.
The same logic is now running through telehealth and remote monitoring. Continuous glucose monitors, for instance, started as an expensive tool for a narrow group of patients and are now cheap enough to be prescribed broadly — the classic bottom-up cost curve. I have a friend who manages a small clinic network, and she told me something that stuck: the thing that nearly broke them was not a competitor with a better service. It was a competitor with a worse service and a price that made "good enough" the reasonable choice.
Diagnostics is the next frontier. Cheap point-of-care testing does to lab work roughly what retail clinics did to primary care: it moves a service out of the specialised facility and into the hands of someone with less training, serving a market the specialists never wanted.
Finance and mobility: the version everyone knows, told properly
Mobile payment systems in emerging markets are the textbook new-market disruption. They did not take customers from established banks. They served people who had never had a bank account, using phone credit as currency. Once that base existed, the climb upward was inevitable.
Ride-hailing is messier than the story suggests, and I think it is a weaker example than most articles admit. It disrupted the dispatch layer, not the vehicle. The incumbent — the driver — is still there, still paid per trip, still bearing the cost of the asset. The disruption was real for taxi medallion owners. For the underlying economics of moving a person from A to B, less than the narrative implies.
Energy and software: where the cost curve does the talking
Solar and battery storage follow the low-end pattern almost perfectly. Early installations were expensive and unreliable, subsidised and niche. The grid treated them as a rounding error. Now they are the cheapest new capacity in most markets, and the incumbents who called them toys are buying them at a premium.
Software is where the pattern has accelerated past Christensen's original timeline. Cloud computing started as a way for small teams to avoid buying servers. It was slower, less secure, and a joke to enterprise IT departments. That joke now runs most of the world's infrastructure, and the enterprise vendors who dismissed it are renting capacity from the companies that made it.
How to tell a real disruption from a fad
This is the part I wish more people wrote about. Every year, a new technology gets labelled disruptive and most of them are not. Here is the test I actually use, built from watching both successes and a few very confident failures.
| Signal | Real disruption | Fad or sustaining innovation |
|---|---|---|
| Entry market | Serves people the incumbent ignores or rejects | Targets the incumbent's existing customers directly |
| Cost trajectory | Falling fast and predictably | Flat, or rising with quality |
| Initial quality | Worse on the metrics incumbents care about | Better or equal from day one |
| Incumbent reaction | Dismissal, then panic, then acquisition attempts | Faster feature response, or indifference |
| Who complains | Regulators and displaced workers | Nobody, usually |
The cost trajectory is the single most reliable indicator. If the entrant's cost per unit of service is falling on a predictable curve, the disruption is real even if the product looks terrible today. If the cost is flat and the product is just "nicer," you are looking at a marketing event.
Where the theory breaks (and why you should know)
Christensen's framework has been criticised, sometimes fairly. The main charge is that it explains disruption beautifully in hindsight and poorly in advance — pick any successful company, find its humble origin, call it disruptive. That is not a theory, that is a story.
There are also genuine cases where disruption did not happen. Incumbents sometimes do see it coming and respond. Regulated industries can slow the entry point for decades. And some technologies plateau below the threshold customers actually need — electric vehicles spent years in that zone before the cost curve finally broke through.
The honest position: the framework is a useful lens, not a law. Use it to ask better questions, not to predict the future with confidence. I have been wrong about this myself — I once bet on a "disruptive" logistics startup that died because the cost curve never bent the way the pitch deck promised.
So how do you actually spot a disruption early?
Watch the bottom of your own market. The customers you turn away, the segments your sales team finds annoying, the requests that are too small to be worth a meeting. That is where the entrant starts. Then track one number obsessively: the cost per unit of what they offer. If it is falling, take it seriously. If it is not, stop worrying.
The decision that matters is not technological
Every disruptive innovation I have watched win came down to the same moment: a group of competent people deciding that a small, unprofitable, slightly embarrassing opportunity was not worth their attention. They were right about the present. They were wrong about the slope.
The uncomfortable part is that you cannot fix this with better analysis, because the analysis is not broken. The incentives are. If your business is measured on margin and your disruption candidate has none, you will reject it — correctly, every single quarter, until the quarter when it is too late.
So the real question is not "which companies were disrupted." It is: which small, ugly, low-margin thing are you ignoring right now, and what would have to be true for it to matter in five years? Sit with that one for a while. The answer is usually more uncomfortable than the question.
