Exclusive: Amazon says 100 million Alexa devices have been sold — what’s next?

More than 100 million devices with Alexa on board have been sold. That’s the all-too-rare actual number that Amazon’s SVP of devices and services, Dave Limp, revealed to me earlier this week. That’s not to say Amazon has finally decided to be completely transparent about device sales, however. While the company claims it outstripped its most optimistic expectations for the Echo Dot during the holiday season, Limp wouldn’t give a number for that. Instead, Limp says, Amazon is sold out of Dots through January, despite “pushing pallets of Echo Dots onto 747s and getting them from Hong Kong to here as quickly as we possibly could.”

But back to that 100 million number. Depending on how you count, it’s either seriously impressive or a serious problem for Amazon. On the one hand, 100 million pales in comparison to the number of phones that have either Siri or Google Assistant pre-installed. On the other hand, the word “pre-installed” is the key thing to pay attention to. With Alexa devices, you could argue, consumers are making an active choice to purchase an assistant instead of just getting a default.

It’s been a while since we’ve had a good, old-fashioned, knock-down, drag-out, winner-take-all platform war. But one might be brewing right now in the world of intelligent assistants. As in any platform war, the numbers come out front and center, and Amazon has the lead on many of those numbers: more than 150 products with Alexa built in, more than 28,000 smart home devices that work with Alexa made by more than 4,500 different manufacturers, and over 70,000 Alexa skills developers. The numbers for Google Assistant were lower across the board the last time we heard them, but it’s likely Google will use CES to check in with new ones.

How Amazon got to those numbers is a lot more interesting than the numbers themselves. Amazon’s strategy for Alexa reveals a fundamental philosophy: speak softly and empower everyone to just ship.

Is this really a platform war? Limp, a veteran of Amazon who started out building the Kindle, doesn’t seem to think so — or at least, he believes that it will play out differently from platform wars of the past. “I don’t think it falls into a sporting event,” Limp says, “where there’s going to be one winner.” He adds, “There will be multiple players for the foreseeable future. I don’t think it’s only going to be only two, either. I think there will be more than that.”

It’s easy to see that attitude as standard executive demurral. Why start a fight if you don’t have to? But having talked to Limp since the early days of the Kindle, I see it as part of his character. He is a laid-back executive who is remarkably candid, pragmatic, and almost chill about issues that other executives would dance around.

As just one example, Limp doesn’t seem too worried that third-party Alexa devices will hurt the Alexa brand. There are more than 150 different products right now with Alexa built in, more than 100 of which shipped in 2018 and aren’t made by Amazon at all. Some of them are pretty great — the Sonos Beam soundbar and Bose QC35 II headphones are strong examples — but inevitably, some of them are going to be pretty junky.

But Limp’s not worried. “There are hits and misses,” he says. And though Amazon works to ensure that everything with Alexa follows Amazon’s privacy policy and doesn’t misuse data, Limp thinks consumers won’t pin blame on Alexa if they happen to buy a dud. He just wants to make sure there’s at least one good example of each different type of Alexa-enabled gadget. “As long as there are existence proofs in each of the categories,” Limp argues, consumers will figure it out.

Another thing you’d expect in a platform war: walls, exclusivity agreements, and stuff that doesn’t work together. Limp’s not interested in those either. He’s all for allowing device makers to create gadgets that support multiple assistants. He pointed to the Facebook Portal and the partnership with Microsoft to pair up Cortana and Alexa as examples.

When I ask him if Amazon is doing anything in contracts with companies to limit what they do with other assistants, the question was so outside of Limp’s worldview that he is taken aback by it. “Uh, no,” is the simple reply. Limp is even open to working with competitors to make different assistants work better together — for example, by letting them all use the same standard for identifying rooms so you don’t have to set up everything multiple times.

“We’re still a big believer in multiple assistants,” Limps says. “We think they will interact in lots of different ways.” In other words, if you’re wondering what’s keeping Sonos from shipping Google Assistant on its devices, don’t blame Amazon. Of course, it’s easy for Limp to be all kumbaya about competition when Alexa is in a pole position and also when other divisions of Amazon are playing hardball with competitors like Google and Apple.

Still, it’s hard not to see Alexa, Google, Apple, and, to a lesser extent, Microsoft fighting for dominance. And CES is the perfect place to watch that fight play out. It may not be the venue where The Next Big Thing gets revealed, but CES is the place to see where the trends are heading.

This year at CES, you can expect Google to reprise its shock-and-awe strategy of literally wallpapering Las Vegas with “Hey Google” ads as it erects a massive booth in the parking lot outside of the convention center. Google started a little behind Alexa when it launched its Home products, and it’s well behind Alexa when it comes to third-party support, so going all-in on the event where all of those third parties are partying makes perfect sense for Google.

But not so much for Amazon. “Customers do not care about an ad campaign on the Las Vegas Strip,” Limp says. “They just don’t. It’s playing to the industry. It’s not playing to who really matters. Will we have a slide with a diving board and a pool? No, that’s not the plan. But I think we’re going to have lots of partners with really good products, which I’m excited about.”

Amazon won’t have a gigantic booth, nor will it stage an over-the-top keynote, but it will have a presence. That presence will be the ubiquity of “Works with Alexa” labels on gadgets throughout the show floor alongside new devices that have Alexa in them. “CES is about making sure that the partners we have [are supported],” Limp says. “It’s less about Amazon itself.”

When it comes to Alexa products, that’s just how Amazon rolls. There’s no clearer example of how Amazon does things than how it announced the last wave of Echo devices. This past September, Amazon held an event in Seattle where it announced literally dozens of Echo and Alexa products. It showed off new Echos and wall clocks, new screens and wall plugs, and even a microwave.

There was no preamble, no splashy magazine feature, just Amazon saying something like, “Here is everything we’re releasing. Here’s what they do. Have at it.” That strategy can backfire. It certainly doesn’t build hype, and it makes it all too easy for products to get lost in the shuffle. “What we want to do is not speak through an ad campaign or a keynote,” Limp says, “but instead speak through the results.”

The result for Amazon is a big lead in the raw number of devices that either support Alexa directly or can be controlled by Alexa through skills. It has a major lead in the home, too. Limp says, “At least as it relates to home assistants — we can certainly talk about mobile assistants, [too] — I think we are by far the leader. Our data would suggest we’re way out ahead.”

But let’s talk about mobile assistants. Google and Apple can include their own assistants on board as the default on their phones, leading to install bases that could number in the billions instead of the 100 million Alexa devices Amazon has. That really could be a serious problem for Amazon: if it turns out that consumers prefer to just use one assistant, in the long term, that edge Alexa out of people’s lives.

Limp’s response is that he’s not actually interested in competing directly with “assistants” as a category. He focused on a bigger goal. The platform war isn’t for a voice interface to set timers and turn on your lights. It’s for, as Walt Mossberg has called it, “ambient computing.”

“We have a very singular focus on this idea of an ambient user interface and where that makes the most sense,” Limp says. That’s why Amazon started with the home and why it’s headed to cars next. In September, it announced a prototype-looking product that lets you use Alexa in your car, and Limp says that nearly 1 million people have signed up to get an invite to purchase one.

As for phones, Limp argues that a “phone is not an ambient device. It’s meant as a personal device.” Just at a pragmatic level, your phone spends a ton of time sitting in your pocket. It’s not ambient; it’s something you directly interact with.

Of course, Amazon has to say this: it’s locked out of being allowed to be the default assistant on the iPhone, and convincing Android users to switch their default assistant to something other than Google is incredibly difficult. (Just ask Samsung how Bixby is doing.) Instead, Amazon is focused on the places where it could make sense to speak out loud into a room: your home, your car, and, increasingly, the workplace. Alexa can work with Polycom speakerphones to set up conference calls, for example.

Amazon’s strategy for Alexa in 2019 is a continuation of what it’s been doing all along: build the foundation for an ambient computing platform. That sounds heady, but the pieces required to build that foundation are remarkably practical and — as with any software foundation — nerdy.

Limp wants to get as many devices to use Alexa as possible, and he wants those devices to do as much as possible. “I think someday, third-party Alexa-enabled endpoints may be vastly larger than what we produce ourselves,” he says. So his team’s focus is on building out tools and APIs for third-party developers alongside making Echo products.

Limp thinks that he can use the same playbook that Amazon used to turn Amazon Web Services (AWS) into the default web services platform that a million software startups use. “You need breadth and depth. … The way to get the breadth is to build tools that are self-service, that are free, that are easy to use, and that are widely available for developers.”

Those tools include things like the Alexa Connect Kit, a hardware module that can make adding Alexa to a gadget nearly plug-and-play. They include APIs and SDKs that make it easy to integrate software products into Alexa, which led to the surprise support for Apple Music on Echo.

“You can end up with something like AWS,” Limp says. “You didn’t have to call anybody to build a storage system based on S3. You just basically put your credit card in, signed up for an account, and next thing you knew, you could do puts and gets. … You should be able to come to our website, sign up for an account, and be able to get the breadth of all the things that happen.”

Building up that foundation also means moving beyond the basics of keywords and skills. “We started with an app store-like metaphor for skills, not because we think that was right, but because that’s what we could do quickly,” Limp says. Alexa needs to become more conversational, make skills more discoverable, and generally just become less awkward to use. That won’t happen quickly (and Google has some advantages there), but it’ll be a big focus going forward.

Amazon intends to focus heavily on enabling those partners to just make more Alexa stuff in 2019, and also to push hard on international expansion to drive those adoption numbers even higher.

Whether this is a platform war or not, one thing is clear: bigger numbers are better.