Loading...
In 2019, three Stanford MBA graduates launched Verdant, a same-day organic grocery delivery platform targeting health-conscious millennials in the Bay Area. Before writing a single line of code, they'd devoured every retrospective on DoorDash, Instacart, and Gopuff — companies that turned logistics chaos into billion-dollar valuations. "We've reverse-engineered the playbook," co-founder Ava told a room of angel investors at Demo Day, projecting a slide showing Instacart's growth curve with Verdant's logo pasted over it. Nobody in the room asked about the hundreds of delivery startups that had quietly shut down that same year. The failures didn't write blog posts. Ava's team — two software engineers and a former brand strategist from Nike — had never managed a warehouse, negotiated with ...
Popular framing: Founders fail because they're not smart enough or don't want it hard enough; the good ones make it.
Structural analysis: The visible startup canon is a survivorship-biased sample — the thousands of failed companies don't write retrospectives — so founders pattern-match on a distorted distribution and underweight base rates. Planning-fallacy timelines, overconfidence about adjacent expertise, and sunk-cost commitment keep teams executing a plan whose premises stopped holding. Most failures are predictable from the priors, not the people.
The gap matters because it determines the interventions we design. If failure is a founder-cognition problem, the solution is better education and self-awareness. If failure is a systemic problem, the solution requires redesigning information flows (mandatory failure post-mortems), fundraising incentives (LPs demanding operational due diligence), and base-rate visibility (public failure registries). Misidentifying the level of the problem produces interventions that are individually rational but collectively ineffective — more founders reading Kahneman while the ecosystem that generates their priors remains unchanged.
