In the summer of 2003 Amazon was a fast-growing online bookstore. Nine years old, it had pioneered many innovative components of e-commerce that we take for granted today: customer reviews, a personalized shopping experience, relevant product recommendations, and frictionless buying.
Jeff Bezos had founded the company in 1994 and began selling books online. His advantage, he thought, was that he could sell a bigger selection of books than any traditional book store (or chain). And right from the start he committed to outstanding customer service. Amazon’s mission is “to be Earth’s most customer-centric company, where customers can find and discover anything they might want to buy online, and endeavors to offer its customers the lowest possible prices.”
Which is why it wasn’t an easy decision in the summer of 2003 to hire a team of 50 or so engineers, and set them upon the task of build “the operating system of the internet,” as Andy Jassy, Sr Vice President of Amazon Web Services, likes to call it.
The Leadership team took an assessment of its assets:
“We were doing an assessment ourselves, as a senior leadership team, about what Amazon was really good at. And initially you can imagine what we were talking about: we were good at retail, we were good at detail pages on our retail site website – all the things that were the primary business at the time. But as we dug deeper we realized we were really good at running these infrastructure services deep in the stack, and then we were also good at running these reliable, scalable, cost-effective data-centers. So the first realization we had was that we had a real core competence running infrastructure.”
They had committed to loose-coupling and autonomy beforehand:
“We really got religion in the company about building in a services oriented architectures fashion – very loosely coupled. It really permeated all the teams such that we had goals for the teams that they had to have well-documented APIs that were hardened so that each of the internal teams inside Amazon could use those respective services to build applications – that was a real cultural shift.”
They recognized that other developers and companies likely shared their challenges:
“When I would talk to our product development leaders they’d say ‘Look I know you guys think these projects should take 2 or 3 months end-to-end, but I’m spending 2-3 months just on the storage solution, or just on the database solution, or just on the compute solution.’ And what they were building didn’t scale beyond their own project, and they knew that multiple people on other projects were building the same thing. And that was a very interesting realization for us that there was all the reinvention of the wheel inside Amazon – there was this real thirst for infrastructure services. I think if we had never built AWS and opened it externally, we would have build all these services just for Amazon the retailer to move more quickly.”
They experienced and recognized the power of APIs:
“We have this associates business which allows third-party websites to merchandize Amazon retail products […] and that team was trying to figure out how to be more ubiquitous across the million associate websites. They tried several tests, and one of the tests they tried was they took all the product data – title, pricing, availability, etc – they put all of that in an API. They though if they could decouple the data from the presentation that associates would do more with the data than we had time to or could think of. And that led to much better conversion for those that used it. But what really surprised us was: with no promotion, thousand and thousand of developers flocked to these APIs. And they used them for things we didn’t necessarily anticipate them using them for. And they asked us to open up all kinds of parts of our platform that we hadn’t really contemplated.”
Putting this all together, the Amazon leadership team recognized that “if developers at businesses will build application from scratch on top of these Web Services (which people called the cloud now) then the operating system becomes the Internet – which was a really different model then before.” Fortunately for Amazon, when they looked to see if any of the key components of this Internet operating system were already built, they didn’t find them. They felt like they were sitting on a secret:
“We realized we could contribute many of the pieces of the internet operating system. We decided to pursue this much-broader mission which was to enable developers and businesses to be able to use these web services to build any sophisticated, scalable application they wanted.”
Jassy asked Bezos for a team of 57 people to start building AWS. They said yes. The team formed in South Africa and started building the cloud: EC2 for computing, S3 for storage. The rest is history.