I've often complained that the business model of most companies publishing Web 2.0 APIs — even Google — is generally, 'Let's put it out there, we'll figure out how to make money from it later on.' The exception is Amazon Web Services. All its API services are designed from the outset to make money. So as a follow-up to its recent launch of Amazon SQS (see Amazon and Microsoft in mashup pact) I spoke to Amazon last week about its business model to find out what others could learn from its example.
There's no doubt that AWS is meant to become a substantial business that contributes profit to Amazon by delivering services to developers. I interviewed Adam Selipsky, VP product management and developer relations, and it was clear from the outset that AWS takes a business-like approach that's a-typical of the Web 2.0 crowd:
"It's about developers having distinct needs they've told us about and us charging for the service for the value being provided," he told me. "That's a very different approach from a lot of companies that are providing web services."
Every Amazon web service is designed to generate real, measurable revenue for Amazon. Some are monetized indirectly rather than directly — the Amazon E-Commerce Service (ECS), for example, is free of charge to use — but there's never any doubt about the commercial objective: ECS is designed to help developers sell more Amazon products from their websites. Amazon is careful about its beta trials too; SQS was available on free beta for about a year-and-a-half, but the cost of operation was virtually nil. The more costly to run S3 storage service, in contrast, had a private beta with a handful of clients before it launched as a chargeable service.
The business model is simple: listen to customer needs and identify where Amazon has economies of scale or some innovative engineering skills that can be translated into a marketable offering. The S3 storage service for example:
"S3 was built in large part because of customers coming to us and telling us the storage problems they had," Selipsky told me. "It all springs from customer need and customer demand."
Amazon was well aware that it had a very cost-effective storage infrastructure compared to what was affordable for smaller companies. So this was a clear case of being able to translate economies of scale in the infrastructure into a service that could be priced attractively to the market. I was particularly interested in how Amazon sets its pricing, which for services such as S3 and SQS is totally based on consumption — there's no setup fee or monthly subscription. You just pay for what you consume.
Is Amazon pricing exceptionally low in order to buy market share, I wondered? Or on the other hand is it creaming off a lucrative margin because customers are prepared to pay a lot more than its actual costs? No on both counts, Selipsky told me:
"We aim to pass our cost efficiencies on to our customers," he said. "We know exactly what our costs are so we know how aggressive we can be in passing those costs on."
This is an interesting strategy that puts Amazon in direct competition with infrastructure vendors that are big in online storage and middleware, for example. S3 seems to have had some early successes. Does that mean vendors such as Sun, EMC or Network Appliance are going to start losing sales to Amazon S3? That probably depends on the extent to which AWS can appeal to the wider enterprise market, which I'll come on to in part 2 of this interview.