The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail

Miles: Early in nineteen eighty five, engineers at a leading maker of hard drives have something that works. A smaller drive, three and a half inches across, built and running.

Miles: The marketing people do what a well run company is supposed to do. They take it to their best customers, the firms building desktop computers, and they ask. Those customers want more capacity. This little drive holds less, and every megabyte on it costs more.

Miles: So the forecast is cut, the programme is cancelled, and the money goes where the customers pointed. Every step of that is defensible. Three years later the company ships the small drive anyway, and almost none of them end up in the portable machines the small drive turned out to be for.

Miles: Today: The Innovator's Dilemma, by Clayton Christensen. In two sentences: great companies don't fall because they're badly managed, they fall because they're managed well, listening closely to their best customers right up to the edge. And the threat shows up looking like a worse product, sold to people the leader has no reason to want.

Miles: Sixty Books. Presented by Stoar Consulting and Coaching.

Miles: Thirty seven, and the reading room's open again. You've got Miles, on Sixty Books. The arrangement holds: a single management book each episode, sixty chosen from a century of them, every one asked the same two things. What it offers somebody leading, or working for, a mind built along unusual lines. And what it still has to say once AI has pulled up a chair at the desk. Nora carries the research, and this one arrives with a quarrel attached.

Nora: Last time Rumelt wanted a choice made out loud, with a diagnosis underneath it. This book takes a firm that does exactly that, quarter after quarter, and shows you it can still lose.

Miles: So a good choice isn't enough.

Nora: Christensen's managers diagnose, set a policy and act on it coherently. The trouble is where the diagnosis comes from. It's built out of what the best customers say, and that's the instrument that fails them.

Miles: So Rumelt hands you the decision, and this one explains why a building full of capable people can't take it.

Miles: Start where the book started, which I gather wasn't as a book.

Nora: As a doctorate. He went back to Harvard Business School in nineteen eighty nine, in his late thirties, wanting to know why strong companies lose. And he records a piece of advice he got. Geneticists don't study people, because a human generation takes decades. They study fruit flies, which in the telling he passes on are conceived, grow up and die inside a single day.

Miles: So he went looking for the fruit fly of business.

Nora: And found it in disk drives. Products turned over in a couple of years, companies came and went within a decade, and an annual trade report had been tracking the products year by year.

Miles: And what did the flies show him?

Nora: The book's own count from that data. Seventeen firms made up the industry in nineteen seventy six, most of them large, diversified companies. By nineteen ninety five every one of them except IBM's drive operation had failed or been bought.

Miles: And the easy explanation is that they got lazy.

Nora: Which he rules out in so many words. Not passive, not arrogant, not afraid of risk, and not outpaced by the engineering. On most new technology the established firms led. Where they lost was on a small number of changes that were technically simple, and the cold open is one of them.

Miles: Fill it in.

Nora: Seagate, early nineteen eighty five, with working prototypes of the smaller size. Its desktop customers wanted forty to sixty megabytes. The new drive offered twenty, at a higher cost per megabyte. Forecasts were cut and the programme was shelved. Seagate started shipping that size in nineteen eighty eight, and by nineteen ninety one almost none had gone into portables. They went into desktops, mounted in frames built for the older, larger drive.

Miles: So they built the new thing and sold it to the old customer.

Nora: That's the book in one company. Keep the name in mind, because it's at the centre of the argument later.

Miles: The author.

Nora: Clayton Christensen, born in Salt Lake City in nineteen fifty two. Economics at Brigham Young, a Rhodes Scholarship to Oxford and a master's there in applied econometrics, then a Harvard MBA in seventy nine. After that, consulting, a year as a White House Fellow working for two transportation secretaries, and then a materials company he co founded and ran.

Miles: So he'd sat in the seat the book is about before he studied it.

Nora: He had. The doctorate came in ninety two. His history of the drive industry ran in the Business History Review in ninety three and won its prize for the best paper that year. He joined the faculty, and the book came out in May of ninety seven from the school's own press.

Miles: And it was a hit straight away?

Nora: No. A trade magazine reported in early ninety nine that it had sold more copies in the first six weeks of that year than in all of ninety eight. It had already taken a Global Business Book Award as the best business book of ninety seven, and a business magazine had just put him on its January cover beside Andy Grove of Intel. Later, The Economist named it one of the six most important books about business ever written.

Miles: And after it?

Nora: Sequels, starting with The Innovator's Solution in two thousand three, and years more teaching. He had a stroke in two thousand ten that affected his speech, and he died in January twenty twenty, at sixty seven, of complications of leukemia.

Miles: Take me through what he argues, chapter by chapter.

Nora: Two kinds of technology to begin. Sustaining technology makes a product better on the measures its best customers already use: more capacity, more speed. Incremental or radical, the leaders win nearly all of it. Disruptive technology is worse on those measures when it arrives. What it offers is something else entirely: cheaper, simpler, smaller, easier to use. It sells first to customers the leader doesn't want, at margins the leader can't live on.

Miles: So worse, on purpose.

Nora: Worse on the scorecard the big customer keeps. Then the second idea, which is speed. Technology tends to improve faster than customers can use the improvement. So a product that starts too weak for the mainstream can climb into it, while the leader keeps climbing and ends up offering more than most buyers ever needed.

Miles: Right.

Miles: Then why can't the leader see it coming?

Nora: Chapter two's answer is the value network. A firm sits inside a set of customers, suppliers and costs that decides what counts as a good idea. Mainframe buyers rank capacity first. Portable makers rank size and weight. The same drive is excellent in one network and useless in the other, and the firm's whole machinery for judging ideas has been tuned to its own.

Miles: Then he leaves drives behind.

Nora: For mechanical excavators. Cable operated shovels give way to hydraulics over about twenty five years, and of thirty makers of the cable kind, four came through. Then steel. Minimills start at the bottom, making reinforcing bar, the lowest margin product in the trade, and the integrated mills are glad to let it go. The minimills climb: bars, rods, angle iron, structural beams, eventually sheet. By the time of the book they hold forty percent of the North American market, and as of ninety five no integrated company anywhere had built a minimill of its own.

Miles: So the big mills just kept walking upstairs.

Nora: And every step up was the more profitable one, which is the title of chapter four in a sentence. What goes up can't go down.

Miles: And the second half of the book.

Nora: Five principles. Companies depend on customers and investors for resources, so in practice the customers allocate the budget. Small markets don't solve a large company's need to grow. Markets that don't exist can't be analysed. An organization's capabilities define its disabilities. And what technology supplies isn't necessarily what the market demands.

Miles: Take the third one.

Nora: It's the one a planning department should find uncomfortable. For a market that doesn't exist yet, the forecast is the one number in the building you can be certain is wrong, and the process won't fund anything without a forecast.

Miles: So the spreadsheet votes no.

Nora: Every time. Which is why his remedy is never a braver executive. Chapter five says give the new technology to an organization whose customers need it. A separate unit, small enough that a small win excites it, with costs low enough to live on thin margins. And near the end there's a long worked case, the electric car as it looked in the mid nineties, run through every one of the principles.

Miles: Now the quarrel.

Nora: I'll give it as an exchange, because that's what it was. June twenty fourteen, The New Yorker, the historian Jill Lepore, in a piece called The Disruption Machine. Her general charge is about method. A handpicked case study, she writes, is a notoriously weak foundation for a theory.

Miles: And the specific charges.

Nora: Two industries. On drives, she says Seagate wasn't felled by disruption at all. Its sales doubled between eighty nine and ninety to two point four billion dollars, which an industry report put at more than all its American competitors combined. In ninety seven, the year of the book, it was the largest company in the industry, with revenues of nine billion. She adds that most of the entrants the book celebrates no longer exist, their success brief in some cases and illusory in others.

Miles: And steel.

Nora: That the account leaves out labour. Twenty two thousand U.S. Steel workers off the job in a dispute across eighty six and eighty seven, a workforce unionised for generations, set against newer minimills that were mostly non union. And her line on where it ended up: in twenty fourteen the largest steel producer in the country was still U.S. Steel.

Miles: So what did Christensen say?

Nora: That week, in an interview with Bloomberg Businessweek, he said she'd judged the theory on the nineteen ninety seven book alone, and left out the later work that he said had taken up the very problems she raised. He also agreed with her on something. The word disruption, he said, was being used to justify whatever anybody wanted to do.

Miles: And the longer answer?

Nora: Harvard Business Review, December twenty fifteen, written with Michael Raynor and Rory McDonald. Three of their points meet her directly. Disruption is a process and not an event, and it can be slow: they write that it took Nucor more than forty years to match the revenue of the largest integrated steelmakers. Success isn't built into the definition, so not every disrupter wins. And an incumbent shouldn't tear down a business that's still making money. They also wrote that the theory gets criticized for shortcomings that had already been addressed.

Miles: And what does the book itself say about Seagate?

Nora: It tells the story we opened with. A firm that built the smaller drive, shelved it on its customers' advice, and sold it late into the market it already had. It doesn't claim Seagate died. So part of this exchange is about what the theory predicts in the first place: who leads the new market, or who's still standing.

Miles: Has anybody tested it outside the argument?

Nora: Twice in ways that matter here, and they point different ways. In twenty fifteen, in the MIT Sloan Management Review, Andrew King and Baljir Baatartogtokh took seventy seven cases from two of Christensen's own books and put each one to people who knew that industry, seventy nine experts in all, through a structured survey. Four elements: incumbents improving along a trajectory, that improvement overshooting what customers needed, incumbents able to respond and not doing it, and incumbents floundering as a result. Seven of the seventy seven fit all four. About nine percent.

Miles: That's a small number.

Nora: And it was disputed in the same journal. Juan Pablo Vázquez Sampere replied that they'd missed more than forty articles testing the theory, that it has more than four elements, and that seventy nine experts is thin coverage for seventy seven industries. So it's an expert survey with its critics on the record, and not a ruling.

Miles: And the second test.

Nora: An economist going back to the fruit flies. Mitsuru Igami, Journal of Political Economy, twenty seventeen. A dynamic model of the industry, fitted to records of drive makers from nineteen eighty one to ninety eight, asking why incumbents brought out the new size later than entrants did. His answer: reluctance to eat into their own sales explains at least fifty seven percent of that gap, even though incumbents had strong reasons to move first and lower costs than the newcomers.

Miles: So the mechanism survives.

Nora: In that industry, in that model, the core of it does, with one change of emphasis we'll need in a minute. It comes out as a rational calculation, not a blind spot. And earlier, King and Christopher Tucci used the same industry to show that experience in previous markets made a firm more likely to enter the next one, so incumbents weren't uniformly frozen either.

Miles: So where does that leave it on the shelf?

Nora: This one grew out of a doctorate and a peer reviewed history, and it's still practitioner literature: a book of selected cases written for managers, not a study you could repeat from the page. What's held up best is the narrow mechanism, that a firm listening hard to its best customers reaches a new market late, and it's held up in economics he didn't write. What's held up worst is the reach. Read as a law about who survives, the cases are disputed in both directions, and an expert survey found the whole pattern in a minority of the book's own examples. None of that needs a motive from anybody. The cases are public, and you can read them yourself.

Miles: Our half.

Nora: And the book hands it over before I have to. He says outright that these firms didn't fail through arrogance or passivity. A small market's signal came in, the firm's own machinery for deciding what matters scored it low, and the money followed the big customers. That's a description of how information gets processed. There's no character anywhere in it.

Miles: Is there language for that at the level of a company?

Nora: A whole school of it. William Ocasio, Strategic Management Journal, nineteen ninety seven, the attention based view of the firm. His argument is that what a firm does follows from how it channels and distributes the attention of the people making decisions, and that its rules, resources and relationships route each issue toward a particular table or away from it. That's theory, not an experiment, and Christensen's cases read like illustrations of it.

Miles: And at the level of a person?

Nora: Here I'll go carefully. Remington, Swettenham and Lavie, twenty twelve, Journal of Abnormal Psychology. Fourteen autistic adults and fourteen typical adults, matched on age and nonverbal reasoning, searching for a letter in the middle of a screen while also watching for a faint grey squiggle off to one side. As the search got harder, the typical group's detection of the squiggle fell. The autistic group's didn't, and at the heaviest load they picked it up better.

Miles: So the small thing at the edge got through.

Nora: It did. And the authors' reading is the part I'd keep. The same trait that shows up elsewhere as being easily distracted, they argue, isn't a failure to filter. It's spare capacity spilling over. A follow up in children, twenty four autistic and thirty nine typically developing, found the same pattern for a surprise extra item nobody had been told to expect. Small samples, a narrow laboratory task, intellectually able participants. These are group results. They don't test for anything, and they don't describe any single person.

Miles: And for somebody running a team?

Nora: Two halves, and the second is the argument. The first is a habit. When a team misses a signal, or when one person keeps raising the same odd one, ask what your process did with it before you ask what sort of person they are. That's Christensen's own move. His remedy is structural, a separate unit with different customers and a different cost base. He never once tells anybody to try harder to see.

Miles: Then the back half.

Nora: That a team with differently allocated attention in it may pick up a weak signal sooner. I think that's plausible, and nobody has tested it. I found no study that puts neurodivergent staff and a firm's response to a disruptive threat in the same design. The bridge between those two literatures is mine, and I'd rather label it an argument than let it borrow the laboratory's authority.

Miles: Cautions.

Nora: One, and it comes straight from the book. Noticing isn't acting. Christensen's firms had the prototypes in hand. If the person who catches the edge of the screen reports into a process that scores everything by the big customer's measure, you've wasted what they saw, and you've taught them to stop mentioning it. Fix the routing before you celebrate the noticing.

Miles: Okay.

Miles: One minute, in the frame we keep. The Innovator's Dilemma says well run companies lose to worse, cheaper products because everything that makes them well run, listening to the best customers, chasing the best margins, funding the biggest markets, tells them to ignore what's starting at the bottom. For anybody leading a mind built along unusual lines, or working inside one, here's Monday. When something got missed, look first at the route it travelled and not at the people it passed on the way. Give the odd signal a path of its own, a person who owns it, and a budget that doesn't have to beat the core business to exist. When one person keeps pointing at the small, strange customer, treat that as information before you treat it as temperament. And keep noticing and acting apart in your head, because the one who sees the edge needs a process that will carry what they saw.

Miles: Now the tools.

Nora: The fixed caution about this whole kind of tool sits past the closing music, in words that never change, so I'll leave it there and use the time on what only this book supplies.

Miles: Go on.

Nora: Three uses, and all three point the same way, toward the bottom of the market. First, take the deals you lost and the customers who left, and have a tool sort them by where they went. The ones who went to something simpler and cheaper are the book's signal, and they're usually the pile nobody reads.

Miles: Second.

Nora: Write the memo your own process would reject. Ask for the strongest case for the small market, formatted for the meeting that decides funding, with the numbers it can have and the numbers it can't. Then watch what your budget process does with it. That's chapter five run as an experiment.

Miles: And third.

Nora: Draw his two lines. What your most demanding customers can make use of, and what you're shipping. A tool can read through release notes and support tickets and show you where you're overshooting. That gap is where somebody cheaper gets in.

Miles: Is there a catch?

Nora: One with a study behind it, and it's this book's catch exactly. Boussioux, Lane and colleagues, Organization Science, twenty twenty four. A crowdsourcing challenge for circular economy business ideas: a hundred and twenty five solvers from around the world sent in proposals, the researchers generated more with a language model under human direction, and three hundred outside evaluators each rated a random thirteen of the two hundred and thirty four. The assisted ideas scored higher on viability, on financial and environmental value, and on overall quality. The human crowd's scored higher on novelty, on average and at the most novel end.

Miles: So the machine wins on the scorecard and loses at the edge.

Nora: In one contest, on one kind of problem, rated by strangers, so keep the direction and leave the size. And look at what that profile is. Better on the measures already in use, weaker at the far end of new: that's the incumbent this whole book describes. Use a tool to go looking at the edge. Don't let it be the judge of whether the edge matters.

Miles: Finish it for us.

Nora: His, turned on the tools themselves. When something new looks worse than what you have, the question this book teaches isn't whether it's good enough today. It's how fast it's getting better, and for whom.

Miles: That's The Innovator's Dilemma.

Miles: Next time: Thinking in Bets, by Annie Duke. Christensen's managers made the sound call by every rule they had and lost the market anyway, and Duke builds a whole book on the idea that you can separate the quality of a decision from the luck of how it turned out.

Nora: After this hour, that sounds like a bigger assumption than it did before.

Miles: Good company on a hard one, Nora. Thanks.

Nora: Good one to argue about.

Miles: Next book, then.

Miles: Stoar Consulting and Coaching produces this show. Logan Williamson is a Licensed Professional Counselor in Texas and Louisiana. What you heard was one book read closely, for those who run teams and those who sit on them. It was neither coaching nor counseling, and none of it was a judgement about you.

Nora: The source for every claim is on the episode page, alongside the transcript. Book thirty eight follows in two days.

Miles: A few facts about the voices. Both are synthetic. Neither Miles nor Nora is a person, neither holds a licence in anything, and nothing in this episode was said about whoever's listening. AI helps draft every script. Logan Williamson, a Licensed Professional Counselor in Texas and Louisiana, edits each one and signs off on it for Stoar Consulting and Coaching.

Nora: And one note about the AI segment. The evidence that these tools help neurodivergent people at work is early. There are controlled trials now, but they measure things like conversation skill rather than how a job actually goes, and for work itself most of what exists is individual providers reporting success. So take it as promising and not proven. The tools also get rebuilt every few months, so whatever you set up today is built on something that moves.