The Lean Startup

Eric Ries spent his twenties watching well-funded startups ship beautifully-built products into total customer indifference, including his own. After his first venture, IMVU, launched in 2004 and made $300 its first month, he reverse-engineered the failure pattern and arrived at a simple claim: a startup is not a smaller version of a company but an experimental organization whose only real output is validated learning. The Lean Startup, published in 2011, imports Toyota's lean manufacturing playbook — Taiichi Ohno, kanban, the Five Whys — and adapts it to the radical uncertainty of new products.

In 2004, Eric Ries and his cofounders launched IMVU after months in stealth. They had built a 3D-avatar instant messenger they were certain would ride on top of AOL Instant Messenger and MSN — users would import their existing contacts, the network would compound, and growth would explode. The first month, the company made $300. When the team finally pulled teenage users into focus groups, the verdict was unanimous: nobody wanted to add another layer to AIM, and nobody wanted their existing friends to see them as a cartoon avatar.

Drew Houston had a problem in 2007. He believed people needed seamless file sync across machines, but building Dropbox required months of distributed-systems work, and explaining it to investors made eyes glaze. So he recorded a three-minute screencast showing the product as if it already worked, narrating over a mocked-up desktop. He salted the video with inside jokes for the Hacker News crowd — a Tay Zonday cameo, references calibrated for the audience — and posted it on the site.

Wealthfront didn't begin as Wealthfront. It launched in 2008 as kaChing, a fantasy stock-picking platform where users built portfolios and competed for status. Engagement was strong, but the business model wasn't. The team noticed that some amateur pickers had genuinely good records, and other users wanted to copy them — so they pivoted into a marketplace where talented amateurs could manage real money for strangers.

IMVU's monthly signups looked like a textbook hockey stick. Total registered users, cumulative downloads, gross revenue — every chart in the board deck pointed up and to the right. Investors were thrilled. But Ries was suspicious. When he broke the numbers down by cohort — what each month's new users actually did over their first few weeks — the lines were flat. New users in October behaved no differently than new users in March. The product wasn't getting better; the marketing was just getting bigger.

The Five Whys was developed by Taiichi Ohno on the Toyota factory floor in the postwar years. When a machine stopped, an operator wasn't supposed to just restart it — they were supposed to ask why? five times in succession, drilling through symptom, mechanism, process, and policy until they hit something a person had decided. The answer to the fifth why was usually a human or organizational choice that nobody had ever revisited.

Ries's gift is reframing the startup from a smaller version of a company into an experimental organization whose only real output is learning. Once you accept that the plan is mostly guesswork, the question stops being 'how do we execute?' and becomes 'what's the cheapest test that could prove us wrong?' You leave thinking less about polish and more about how quickly the next experiment can run.