Startup Punditry's 25 Years of Failure
Government data proves the Lean Startup revolution failed: startup survival rates are unchanged since 1995, because once everyone follows the same playbook, no one has an advantage.
Read Original Summary used for search
TLDR
• U.S. startup survival rates are flat from 1995 to present despite widespread adoption of Lean Startup, Customer Development, and Business Model Canvas—the methods demonstrably don't work
• VC-backed startups are actually failing more: the percentage of seed-funded companies raising Series A has declined, not improved
• The theories are self-defeating—when everyone interviews customers and iterates MVPs, everyone converges on the same answers and competes head-to-head with no differentiation
• A real science of entrepreneurship must embrace the Red Queen hypothesis: you must constantly do something different from competitors, which means rejecting any fixed methodology
• The only valid rule is Feyerabend's "Against Method" principle: if everyone is doing something one way, you need to do it differently—including eventually rejecting this advice too
In Detail
The author presents damning evidence that 25 years of startup methodology—Lean Startup, Customer Development, Business Model Canvas, Design Thinking—has failed its own stated goal of making startups more successful. Government data shows U.S. startup survival rates at one, two, five, and ten years are essentially unchanged from 1995 to present. Even more striking, VC-backed startups show declining success: fewer seed-funded companies are raising Series A rounds now than 15 years ago. These methods claimed to be scientific, but they've produced zero systematic improvement in outcomes despite millions of copies sold and universal adoption in entrepreneurship courses.
The author explores three explanations for this failure. The theories might be flatly wrong, but that would show declining success rates (which we don't see overall). They might be obvious—founders naturally talk to customers and iterate anyway—but that just means we've formalized common sense without improving it. The most compelling explanation is that the theories are self-defeating: once everyone uses the same methods, no one gains a relative advantage. If every founder interviews customers, they all converge on the same answers. If everyone launches MVPs and iterates, they all iterate toward the same product. This is the Red Queen hypothesis from evolutionary biology—species must constantly innovate just to maintain their position relative to competitors. In a competitive market, doing what everyone else does drives profit to zero.
The path forward requires building a science at a higher level of abstraction—not prescribing specific tactics, but articulating theories about how to generate new methods. Drawing on philosopher Paul Feyerabend's "Against Method," the author argues that great scientists succeeded by violating the methodological rules of their time. A genuine science of entrepreneurship would embrace this: if everyone is running lean experiments, don't. If everyone is conducting customer interviews, find another way. The only axiom is that if you do what everyone else does, you get what everyone else gets. This creates a paradox—any paradigm, once institutionalized, becomes the next thing that needs burning. But that's the point: a true science of entrepreneurship must reject any attempt to permanently systematize it, including eventually this framework itself.