The Real Reason Innovation Dies Inside Big Companies

Why "we need to take more risks" is the most reliably broken promise in business

Everyone says they want a culture of experimentation. They put it in the all-hands deck. They say the words out loud: “We need to take more risks.”

And then they go back to their offices and quietly punish the people who do.

These are well-meaning leaders, all genuinely convinced they want their organizations to experiment more — who then flinch the moment an experiment takes longer than planned or costs a dollar more than the safe route would have. They’re not hypocrites. They just don’t understand what experimentation actually requires.

Real tolerance for failure. Not the inspirational-poster kind.

The math nobody wants to talk about

Experimentation fails 99% of the time. That’s not pessimism — that’s the math. If you’re running genuine experiments, testing things you don’t already know the answer to, most of them aren’t going to work. That’s what makes them experiments rather than implementation.

The problem is that “99% failure rate” sounds terrible in a budget meeting. Time and money are easy to measure. The learning embedded in a failed test is not. So leaders who say they want innovation keep defaulting to the metrics they understand — and those metrics punish exactly the behavior they claim to want.

They’re not lying when they say they want more risk-taking. They just keep rewarding something else.

Culture is observational

You cannot install a culture of experimentation. You can only model it.

Culture isn’t what’s written in the employee handbook or posted on the lobby wall. Culture is what people observe — who gets celebrated, who gets promoted, who gets fired, and why. Everything else is decoration.

Geology works this way too. Ninety-nine percent of the change to a landscape happens in one percent of the time — earthquakes, floods, eruptions. The slow steady stuff barely moves anything. Culture is identical. Most of what you do day-to-day barely registers. But a handful of moments — specific, visible, impossible to misread — reshape the terrain completely.

Those moments are the three decisions every leader keeps making: who you reward, who you promote, and who you fire.

That’s the whole game.

The gap between saying and doing

If you say “fail fast” but spend forty minutes in a post-mortem grilling the team on why a test underperformed — you just taught everyone in that room that failure is dangerous. If you say “take risks” but give the bigger raise to the person who hit their number safely rather than the one who tried something ambitious and came up short — you’ve told your team exactly what you actually value.

You can’t talk your way to an experimentation culture. Ninety-five percent of the signal comes not from what you say, but from those three decisions. Reward. Promote. Fire.

At Netflix, we ran tests constantly — ten, twenty different things a day at our peak. The overwhelming majority failed. That wasn’t a bug in the system. It was the system. The subscription model that became Netflix’s core business came out of a year and a half of relentless testing. It wasn’t a brilliant idea someone had. It was what was left standing after everything else failed.

I had to genuinely stop caring whether any individual test succeeded. The only thing that mattered was whether we were learning.

Make sure your incentive system can tell the difference between the team that played it safe and hit their number, and the team that ran a fast cheap test and came up empty. Be loud about your own real failures — not the humble-brag kind. And when you make those three decisions, make sure people understand why. The story that circulates around them is the actual culture document of your organization. Not the one in the handbook.

Most leaders don’t actually want a culture of experimentation. They want the results of one — without tolerating the failure rate that makes those results possible.

You can’t have both. Pick one.