Believing in an idea when the room thinks it is crazy

Tim Westergren was a musician and film composer when he started the Music Genome Project in 2000, an attempt to describe songs by hundreds of musical traits so that a computer could recommend music by how it sounds. The dot-com crash arrived almost at once. By 2001 the company had run through its money, and for about two years its employees worked without regular pay while Westergren pitched investors. He has said he made more than three hundred pitches before one firm agreed to invest in 2004. The internet radio service built on that work, Pandora, launched in 2005 and has since grown to tens of millions of listeners.

From outside, the story looks like the line in action. Most people Westergren pitched thought the idea was unworkable or the business unfinanceable, and he kept going.

What the story also shows is why his belief was reasonable. He was not asking people to imagine a product. The genome existed, the recommendations worked, and the staff who stayed without pay had seen it work up close. The investors who said no were mostly doubting the business model and the market timing, not whether the technology did what he claimed. Belief backed by a working thing is different from belief in an idea alone.

The line is less useful when the people calling an idea crazy are the ones who would use it. Friends and investors can be wrong about a market they do not know. Target customers who try a product and walk away are telling the founder something, and treating their reaction as another obstacle to believe past is how founders lose years.

It also leaves out the cost of belief to others. The employees who deferred their salaries carried real risk for Westergren’s conviction. A founder is entitled to bet on an idea. Asking others to share that bet deserves plain talk about the odds.