Amazon may have started out simply, selling books online, but these days it peddles everything from pickled carrots to two-carat diamond rings. The Internet's biggest retailer says it logged the best holiday season in its history despite the rough economy. CEO Jeff Bezos credits the $4 billion the company spent on R&D over the past decade.
Last summer, Amazon  added a new movie-download service to accompany its MP3 and electronic-book services; the Kindle, its electronic reader, sold more units in its first year than the iPod did in its debut, according to Citigroup analyst Mark Mahaney, generating about $136 million in sales; and the company's cloud-computing and data-storage business now counts more than 440,000 customers and 29 billion objects stored. "Amazon Web Services already uses way more bandwidth than our retail business does," says Bezos. "I'm astonished by that."
Is it harder to take risks now that Amazon is so big [a projected $19 billion in annual revenue]? On the contrary, insists Bezos: "What's dangerous is not to evolve."
Here's what Bezos sees ahead.
FC: You've gone from selling electronics to making one, the Kindle. Why?
There are two ways to extend a business. Take inventory of what you're good at and extend out from your skills. Or determine what your customers need and work backward, even if it requires learning new skills. Kindle is an example of working backward.
We saw maturing technologies we could meld together: electronic-ink display and wireless connectivity. You don't need to find a hot spot. You can download a book in less than 60 seconds. It's the same advanced network that 3G cell phones use. There's no monthly bill. We bundle the delivery into the cost of the book. When we launched, we sold out in five-and-a-half hours.
The Kindle's sales (about $136 million in 2008, estimates Mahaney) are a drop in the bucket for Amazon ($19 billion in projected annual revenue). How do you justify the investment?
We want to plant seeds that grow into big trees, and that may take five to seven years. You also have to be willing to repeatedly fail -- and to be misunderstood for long periods of time.
Why rent out your computers and servers to other companies, even competitors?
On the Web, the price of admission is the network engineering and infrastructure. It's an extraordinarily sophisticated, hard thing to do well. When we were designing the APIs [application-program interfaces] on top of that infrastructure, we realized that with a little extra work, we could build them so that everybody can use them. And we could turn it into a new business. Amazon Web Services already uses way more bandwidth than our retail business does. I'm astonished by that.
Is it harder to take risks now that Amazon is so big?
It's actually easier. You're not betting the whole company. You're doing new things, and if they don't work, you can change direction. What's dangerous is not to evolve.
What will Amazon look like 10 years from now?
The details, technologies, and competitors will change. But the big things won't. It's impossible to imagine customers 10 years from now saying, "I love Amazon, but I just wish their prices were higher or they delivered a little more slowly." We'll continue to pick things that we know really matter to customers and invent around them.