Last week's release of Amazon CloudDrive (a digital music locker that provides storage in the cloud for consumers) should give consumer web services which store, stream, or otherwise provide simple webapps on top of S3 pause. It should also give a bit of heartburn to investors chasing the pot of gold at the end of music streaming, photo sharing, video hosting, backup and other thin front ends to public storage utilities.
Until now Amazon has been very good about sticking to its "sell the picks and shovels" strategy with its web services business unit. Having experienced phenomenal growth— the chart here is for the number of objects (files) in S3 since inception but I bet there is a similar chart for MIPS rented through EC2— mainly on the back of startups— this platform provider is now looking to get a bit closer to the consumer. And thanks to that side business known as "Earth's Biggest Bookstore," these guys have got the chops to do so: they take credit cards, have one of the most sophisticated data-driven merchandising infrastructures out there (constantly testing and iterating every offer), and a brand that caused a friend recently to ask whether he couldn't just direct deposit a part of his paycheck to Amazon directly.
Let's look at the classic platform-eats-its-young economics: Dropbox, a fantastic consumer file syncing product whose storage is provided by Amazon's S3, sells 50GB of storage for $99/year. As of last week, Amazon intends to do the same for $50/year or 50% of that. And to boot, they'll offer a 20GB plan for $20/year an option which I suspect Dropbox can not match due to their paid user acquisition (in a very competitive category, $20 to acquire a user, a rough guideline being to spend first year's margin doing so, doesn't go very far). Naysayers might quickly point out that CloudDrive is for music files, has no sync or fine-grained access control, and lacks features X, Y, & Z. All of this is indeed true but it is early still, and the important thing to realize is that as a beachhead service, music streaming has much broader appeal with consumers than file sync (as Jim from the EchoNest [disclosure: an investment] likes to say, it is the "gateway drug to every other type of media").
I would bet that Amazon is about to begin that delicate dance across the consumer and SMB application ecosystem it hosts (at 300B objects, I'm going to guess at least 300,000 individual applications) all the while cherry picking those most likely to have broad appeal, and in cases like this one, well aligned with the strategic direction of its e-commerce business (the 20GB option is free if you buy one MP3 album for them per year). And much like Microsoft in the "Office Wars," they will end up eviscerating quite a few of the existing services which can not structurally achieve independent economics. In fact, it is likely to be worse because unlike packaged software, rented storage/MIPS carries a marginal cost with which Amazon will always have the advantage.
What to do about it?
Run up the stack as fast as humanly possible. Broad horizontal services that depend on storing consumer/SMB data for one part of the value proposition are played (arguably they were done before this, but now it seems certain).
On the SMB side this may mean providing rich workflows around the data which which go well beyond simple backup or sync. And on the consumer side, companies should be thinking hard about data-based value added services that Amazon is unlikely (or unwilling) to be able to provide any time soon. For sure this is more than just a pastel UI on top of a file explorer or a SendBigFiles service with encryption on both ends. It may mean deeper ACLs based on social graph data, or more likely, applications which make sense of vast data stores backed by S3. I'd also be moving aggressively to support other cloud storage backends (though given the relative strength of Amazon, this may be a bit of a pipe ream).
In the end, you can't blame the platform player for eating its most promising youth— it is in its nature after all as Apple is also proving when it comes to iOS and its own rich new ecosystem. It's also why platform plays that win are so valuable (look out all of you thousands of companies building telephony assisted apps on top of Twilio/Tropo).
Given that most of the technology value creation that takes place by non platform players (because let's face it, while most of us try and all of us like to call ourselves "platforms") happens because of temporal arbitrage across features not yet baked into the core platform offerings and the resulting M&A premiums paid, running like hell to high ground seems to me to be the best option.