The Innovation Curve

Everyone has seen the smooth bell curve of adoption (innovators → early adopters → early majority → …). Moore's subversion is to see that the curve is actually not continuous between "early adopters" and "early majority" — a chasm cracks open there. The issue isn't product maturity; it's that these two groups are psychologically two completely different kinds of buyers.

Early adopters are visionaries: they buy a vision, and even take pride in "nobody else uses this yet." The early majority are pragmatists: they don't care whether the product is "good" — they care whether, if it breaks, they can explain it to their boss. So they demand peers already using it, a phone number they can call, an alternative vendor, retraining options — these are not product requirements, they are personal risk management. The visionary's enthusiastic endorsement reads to the pragmatist as exactly a warning sign: "only pioneers are using it" means "nobody dares to."

Hence the circular trap: pragmatists only trust someone like them having gone first, but if nobody is willing to be first, the product stalls in the chasm — early customers tapped out, mainstream customers immobile.

Moore's solution is counterintuitive because it requires you to deliberately shrink the market to something tiny. Not "sell to enterprises" — sell to the team at pharma companies that prepares FDA filings. Not "build a support SaaS" — build the one that handles e-commerce refund tickets, period. Conditions: pain sharp enough that they'll risk something, and a circle tight enough that one customer's news reaches the next. Pour everything in to dominate completely — and at that point what you offer is no longer software, it is what Moore calls the whole product: core features + training + third-party integrations + success stories + a support channel. Pragmatists have always bought the whole product; technology is one ingredient. Once a few word-of-mouth successes exist inside the small circle, you have manufactured with your own hands the social proof pragmatists demand, and you topple adjacent niches in turn (the next niche no longer sees "pioneers using it" but "the neighbors succeeded") — that's the bowling-pin play. Most companies die because they aimed at the "mainstream market" as their target from day one, skipping the step of manufacturing social proof.

Cases are mostly 1990s enterprise software (Documentum, Ariba); a single beachhead may not be enough in the era of SaaS / two-sided platforms / consumer AI. The theory is qualitatively strong but quantitatively weak — "where exactly is the chasm, and how many references does it take" is hard to measure in advance.

The usual explanation for big-company decline is "arrogance, rigidity, missed opportunities" — the unspoken subtext being "swap the CEO and it's fixed." Christensen's conclusion is colder: the very rational, good habits that made the company successful are what systematically drive it toward death. Try harder at "good management" and you die more reliably — that is why the book is called the Dilemma, not "Failure."

The mechanism lives in resource allocation, not in strategic vision. When a disruptive new technology appears, it has worse performance, a smaller market, lower margins, and is bought by the customers the company cares about least. Every rational process shoots it down: (1) Listen to your best customers — they don't want it, and for them it really is a downgrade; (2) Rank projects by NPV — the disruptive business's cash-flow model looks ugly for five years; (3) Protect margins — this thing drags gross margin from 50% to 20%, so finance opposes it first. Investing in the high-end upgrade for the mainstream is always "more worth it" than investing in the shabby low-end newcomer — and in the present moment that judgment is correct every single time. The fault isn't with individuals, it's with the process — and the process is functioning perfectly at its design goal (maximizing current profit).

What's truly fatal is the crossing of two trajectories. Technology usually improves faster than customer needs grow. The low-end disrupter climbs its own trajectory and sooner or later crosses the "good enough" line for mainstream customers — at which point its cheaper, simpler advantages all cash in, and the giant collapses at once. Meanwhile the giant is happily fleeing upmarket, because the high end has fatter margins and a prettier quarterly report — fleeing until it has handed over the entire low and middle. The disk-drive industry replayed the same script over and over: 14 → 8 → 5.25 → 3.5 → 2.5 inches; each new small platter was dismissed by the larger-platter incumbents as "customers don't want this" — their incumbent customers genuinely didn't, because the new platter was creating new customers. This is exactly why customer interviews can't save you — you interview your incumbent customers, who are precisely the ones pulling you upmarket and forbidding you to step down.

The remedy is therefore paradoxical: before disruption hits, you must disrupt yourself with an independent small organization. "Independent" doesn't mean "given freedom" — it means having its own cost structure, its own customers, its own P&L — because a disruptive business can never survive under the parent's "correct" financial standards. IBM put its PC division in Florida, deliberately not at HQ in New York, and that is the positive case. Kodak invented the digital sensor itself but stayed trapped in film margins — the negative case.

Jill Lepore's famous 2014 critique in The New Yorker noted the theory leans heavily on cherry-picked retrospective cases — many "successful disruptions" are reverse-engineered, while failed would-be disrupters are ignored. Strong explanatory power, weak predictive power: you can't reliably tell in advance which low-end product will actually succeed.

Ries's real insight isn't the everyone-says-it slogan "iterate fast." It's an unsettling diagnosis: in a startup, nearly everything that looks like "progress" is an illusion. You built the features on plan, hit the milestones, even gathered some users — but if all of it rests on an unvalidated core assumption ("people will want this"), then you have merely, efficiently, on schedule and on budget, built something nobody wants. Most failed startups never lacked execution; they poured execution into the wrong assumption.

So he redefines progress as validated learning — the only progress that counts is using real data to prove your judgment about "do users actually want this" has moved closer to the truth. The MVP is widely misread because of this: it is not "a crude little product," it is the cheapest possible experimental apparatus for testing your riskiest assumption. Dropbox's early "fake it" explainer video built no product at all — posted to Hacker News and watching the signups explode — what it validated wasn't "we can build a sync tool" but "enough people are genuinely in pain over multi-device sync," which was the riskier assumption. A landing page, a fake-it video, a service that secretly fakes automation by hand — anything that forces a real user reaction can be the MVP. Misreading the MVP as "a simple first version" is the trick: it turns it back into a product step rather than an experiment, and the original delusion survives.

Ries's most distinctive contribution is to confront a question others dodge: before you have any revenue, how do you know whether you're making progress? The answer is to distinguish two kinds of metrics. Vanity metrics (total signups, cumulative downloads, page views) only go up — a product can be getting worse and the cumulative curve still rises, because new traffic keeps being added. Actionable metrics are cohort retention and the conversion funnel: bucket the users who signed up in April, May, and June as separate cohorts, and look at what fraction of each remain on day 30. If a new version drops the latest cohort's retention, you know immediately that you've made things worse — regardless of what the total number is doing. This "innovation accounting" lets an early-stage company with no traditional financials use objective evidence to judge "is this worth keeping the investment in."

The whole method closes into the Build–Measure–Learn loop, and each loop must end with an honest verdict — pivot or persevere. That step is the real lever. A pivot isn't failure — it is "keep what you've learned, change one assumption." Instagram pivoted from Burbn (a complicated location-check-in + games + photo app) to "just photos and filters" — keeping what the team had learned ("users love the filtered photos") and cutting the rest. Dodge that honesty, and even the fastest loop is just accelerated waste.

The method assumes you can get real user feedback quickly and cheaply. Long-cycle, capital-heavy, highly regulated fields (aerospace, semiconductors, biopharma, medical devices), where a loop is years or a decade, simply can't be jammed into this frame — forcing it makes teams stall at "a barely-working MVP" and lose real technical ambition.

Economics textbooks treat "perfect competition" as the ideal. Thiel's whole book is built on a flat reversal of this: under perfect competition, profit is zero. All players are ground down by price wars to bare survival, with no slack to invest in the long term — any long-horizon R&D, brand, or talent pipeline is eaten by next quarter's marginal-cost war. So any company that truly creates value and is worth founding is, in essence, pursuing some form of monopoly. This is not a moral judgment but a structural fact: money and the room to innovate exist only outside of competition.

From this comes a checklist: monopoly is not black magic — it consists of at least one of four specific features. (1) Proprietary technology — at least 10× better than the next alternative (not 2×; 2× can be explained away as measurement error), e.g., Google vs. AltaVista. (2) Network effects — the more users, the more valuable, and the harder to leave the later you join. (3) Economies of scale — large fixed cost, tiny marginal cost; the first mover pushes the cost curve below where any follower can ever match. (4) Brand — what would otherwise be a commodity acquires a premium. Without at least one of these, however "unique" the story is, it's just packaging.

And from this comes a sharp observation: monopolists and competitors both lie, in opposite directions. Companies stuck in a bloodbath insist they are "unique" (defining the market extremely narrowly: "we are San Francisco's only organic British restaurant") to mask the truth that there is no profit; real monopolists do the reverse — they frantically downplay their monopoly (defining the market extremely broadly: "we're only a few percent of it") to dodge regulators and envy. Google publicly says "we are a small share of the advertising market" (measured against all global advertising), but in search advertising it owns nearly 90%.

Thiel's most offensive cut is aimed at incrementalism. He distinguishes "definite optimism" from "indefinite optimism": the former is holding a concrete, grand picture of the future in your mind and then building straight toward it; the latter is believing the future will be better without knowing what it is, so you dare only to keep options open and tweak in small steps. In his eyes, lean iteration is exactly the latter — it can polish an existing thing to a local optimum but can never find the higher mountain. SpaceX was never "test whether users want to go to Mars" — it was Musk deciding "we are going to Mars," and back-solving every engineering problem from there.

So the signature question — "What important truth do very few people agree with you on?" — exists to filter for people who dare hold a definite dissent. Everything that goes from zero to one begins with a secret that almost nobody at the time believed — and a secret, by definition, means the mainstream consensus is wrong. Airbnb's secret was "the vast majority of people are willing to stay in strangers' homes short-term," which struck nearly everyone at the time as absurd; Stripe's secret was "the real bottleneck in internet payments isn't the technology, it's the developer experience."

Intensely autobiographical — the theory rests almost entirely on PayPal, Palantir, and a handful of Founders Fund bets, a tiny sample with serious survivorship bias. The case for monopoly slides easily into impatience with antitrust. The prose carries a strong ideology, and many of its conclusions readers will not share.