Voice interfaces were all over CES this year. Gadget makers added support for voice-based computing to just about everything they could, ranging from stuff you'd expect like, TVs, speakers, and smartphones, to ones which are little more out there, such as refrigerators, wireless routers, and even air purifiers.
The point is that voice is shaping up to be one of the next big computing interfaces. That's good news for Amazon, which as Ben Thompson lays out in an excellent piece of analysis from earlier this month, has a golden opportunity with its Alexa voice computing platform to build a successful operating system. They failed to accomplish this when they forked Android for their line of Fire tablets and phones (well, just one phone). Ben does a better job than I could of explaining the business models around operating systems and how and why they can be so fantastically profitable, so please go read that piece. Suffice it to say, Alexa gives Amazon a shot at owning a primary computing interface, not unlike how Apple owns iOS or Microsoft owns Windows.
Alexa has proven to be remarkably popular with consumers. While they have not released sales figures, Amazon's line of Alexa-powered Echo speakers are estimated to be present in around seven million US households and appear to have been a hit this past holiday shopping season, with Amazon's companion app for the Echo being the 5th most popular free iOS app the day after Christmas. (It probably didn't hurt that Amazon offered the smaller Echo Dot speaker for just $39, making it a nice, but not-too-expensive gift option.) Meanwhile, partners ranging from Ford, LG, Lenovo, Whirlpool, Huawei, and DISH have been integrating it into their products. With over 7,000 "Alexa skills" and counting made by third-party developers, for now they have a lead over their competitors, which include Google (which is a partner in betaworks' Voicecamp accelerator), Apple, Microsoft, and to a lesser extent, Samsung, which recently acquired Viv. However, for Amazon to maintain its lead -- or for Google or anyone else wanting to build a voice-based computing platform for that matter -- it's going to need to do what practically every successful operating system has done: be a great platform for developers. That means enabling others to build massively successful apps on top of it which in turn develop into massively successful businesses.
The phrase "killer app" gets tossed around casually these days, but Amazon doesn't necessarily need something that'll drive sales of the Echo all on its own, like how VisiCalc, a spreadsheet app, did for the Apple II back in the day (though of course it wouldn't hurt). You can see how the iPhone didn't truly take off until Apple opened iOS up to developers, and even then it took a little while until someone built Instagram, which turned out to be one of the first massively successful apps. Instagram made iOS more valuable and useful to both Apple and its users, and other developers, seeing that success, were driven to create even more apps for the platform, making iOS even more valuable for Apple and its users.
Alexa has over 7,000 skills, and while there are some great ones out there, so far there a killer third-party app for it hasn't emerged. What's tricky is that it's very difficult for any platform maker to engender this intentionally, since you can't just magically make it happen with aggressive dev evangelism. Part of the challenge here is that with any new computing interface it takes time to figure out how to create apps and services that feel "native" to the platform. Platform makers also need to offer a good SDK and the right APIs; often devs are too limited in what they can build in the early days of a new OS. Then there are issues around monetization (whether it's ads, in-app purchases, commerce, etc) and discovery (i.e. how users are going to find and share their apps); according to a report by VoiceLabs, 69% of Alexa skills have one or zero consumers, and retention numbers are poor. People will build plenty of stuff for fun when a platform is new and growing fast, but at some point they want to build businesses around those products. If app makers can't acquire users or make money, they're not going to invest their time into building for a new platform. Not surprisingly, there are still lots of open questions around how to accomplish either of those on voice platforms.
If you get this stuff right, then with enough experimentation (and time) others will eventually be able to build compelling, innovative products which couldn't exist anywhere else. The success of these kinds of products is almost impossible to predict beforehand, yet their emergence feels entirely obvious in retrospect, which is a major reason why we're doing our Voicecamp accelerator. We want to take a bunch of teams building promising new products for voice and create the right circumstances for them to explore what works and what doesn't when it comes to voice. It's possible that at the end of Voicecamp we’ll have a group of noble, but failed, experiments in trying to figure this out. However, we also strongly believe that taking that risk is worth it if there's a shot that one of them will build the first great voice app.