MOST OF THE inventing we do at Amazon goes like this: somebody has an idea, other people improve the idea, other people come up with objections for why it can never work, and then we solve those objections. It’s a very fun process. We were always wondering what a loyalty program could be, and then a junior software engineer came up with the idea that we could offer people a kind of all-you-can-eat buffet of fast, free shipping.

When the finance team modeled that idea, the results were horrifying. Shipping is expensive, but customers love free shipping.

You have to use heart and intuition. There has to be risk taking. You have to have instinct. All the good decisions have to be made that way. You do it with a group. You do it with great humility because, by the way, getting it wrong isn’t that bad. That’s the other thing. We’ve made mistakes, doozies like the Fire Phone and many other things that just didn’t work out. I won’t list all of our failed experiments, but the big winners pay for thousands of failed experiments.

So we tried Prime, and it was very expensive at the beginning. It cost us a lot of money because what happens when you offer a free all-you-can-eat buffet? Who shows up to the buffet first? The heavy eaters. It’s scary. It’s like, oh my god, did I really say as many prawns as you can eat? And so that is what happened, but we could see the trend lines. We could see all kinds of customers were coming, and they appreciated that service, so that’s what led to the success of Prime.