How Anthropic, Costco, and Patagonia all build incorruptible companies | Eric Ries

From Lean Startup to "Protect What You've Built"

Ries opens by connecting this book to The Lean Startup: the first was about building a successful company, this one is about not watching that company get hollowed out later. He describes a force "that no one controls but everyone obeys" — the drift toward mediocrity, bureaucracy, and eventual betrayal that founders usually notice only after it's too late.

The Corruption You Can Literally Taste

His most vivid example is a dinner where a friend took one bite, pulled out his phone, and said, "I could tell this restaurant got taken over by private equity. I could taste it." That's Ries's whole point: companies often aren't destroyed by competition but by success itself, because the more valuable the goose, the stronger the temptation to butcher it.

Why Founders Should Stop Assuming They're the Exception

Ries pushes back hard on the usual founder confidence here. He cites Harvard Law School data that only 20% of venture-backed founders remain CEO three years after IPO, then tells the story of a hot company whose founder ignored his governance warnings, had a successful IPO, and was ousted just five months later after one market shock.

"Too Early" Turns Into "Too Late"

One of his sharpest arguments is about timing: mission protections always sound premature until the exact moment they're impossible. He walks through the full sequence — lawyer says wait for product-market fit, VCs say wait for scale, bankers say bundle it with the IPO — until the founder finally asks about mission provisions and the CFO says, "Oh, you were serious about that? Now it's too late."

Novo Nordisk and the 100-Year Proof That Structure Matters

To show this isn't idealistic fluff, Ries tells the origin story of Novo Nordisk. In 1920s Denmark, Marie Krogh's diabetes diagnosis led her and Nobel laureate August Krogh to commercialize insulin using a two-tier structure: a for-profit company governed by a nonprofit foundation, built specifically to prevent future exploitation. His point is that these "industrial foundation" models aren't experimental — Zeiss used one in 1885, and research shows such firms are six times more likely to survive to year 50.

The Brutal Vectura Example: Fiduciary Duty Can Force the Wrong Answer

Ries then drops the ugliest story in the episode: Vectura, a UK inhaler company, received acquisition offers from Philip Morris and private equity, plus the option to stay independent. The board accepted Philip Morris's higher bid — 165 pence per share versus 155 — because they believed they had a fiduciary duty to take the highest price, and within three years Philip Morris had written down roughly $900 million and broken the company apart.

Harder Is Easier: Cloudflare, Groupon, and the Culture Bank

The practical leadership principle is "harder is easier": do the principled thing even when it costs you. He contrasts Groupon, which let short-term ROI logic push it from one email a day to eight, with Cloudflare, where a junior engineer challenged Matthew Prince on charging for SSL; Prince's answer — "let's figure it out" — led the team to make encryption free, sacrifice near-term conversion, and massively expand trust and top-of-funnel.

Anthropic, OpenAI, and the Need for a Mission Guardian

On AI, Ries says the stakes are so high that none of the major labs use ordinary startup governance. He recounts advising Anthropic early, before ChatGPT and before it was a hot company, and says the team wrote mission protections into its charter from day one, later creating a long-term benefit trust whose outside trustees are AI safety experts with no equity. His broader takeaway: whether through founder control, a nonprofit, a perpetual purpose trust, or another mechanism, companies need a real "mission guardian" with legal power, because mission alone won't survive pressure.