Amazon’s services strategy is the key to Alexa’s success

Amazon and smart home businesses launched more than a dozen new devices at an event on Wednesday. The company also shared an eye-popping statistic: There are now 85,000 devices that incorporate Amazon’s digital assistant Alexa. You can buy twerking bears, Christmas trees, kitchen scales, light bulbs, and more that link back to Amazon’s digital assistant. A year ago, you could buy about 4,000 such devices, according to Daniel Rausch, VP of smart home at Amazon.

While a lot of us focus on Amazon’s smart speakers and the launch of voice when we think about Alexa, it’s more accurate to consider Alexa as a new communications and technology platform. The digital assistant is the future, and Alexa is Amazon’s attempt to make its digital assistant the chosen platform for as many people as possible. The key to expanding Alexa is to think of her as a service for consumers and to consider the rest of Amazon’s efforts as a way to spread Alexa to manufacturers and developers.

When viewed in this context, Amazon’s announcements of random devices and the creation of new infrastructures such as the Sidewalk wireless protocol or better neural networks for voice make sense. And the service-centric viewpoint that Amazon has had since day one has helped it by making the infrastructure for a consumer-oriented internet of things ubiquitous.

I saw a lot of people trying to compare the launch of Amazon’s devices and its lack of discipline around those devices to Apple, but the focus on hardware misses the point. Amazon is succeeding because it is willing to throw less-than-perfect hardware over the wall to see what sticks. Remember the original Echo Show? It was hideous, but the next-generation Show devices improved on the design and now the Show products are a huge seller for Amazon.

In the wake of this event last year, I wrote that Amazon had pulled an Apple on the smart home by making it easy to connect and manage products in a way that people traditionally associate with the iPhone maker. In fact, some of the criteria that Amazon has for its Certified for Humans program, which it announced for connected devices on Wednesday, are straight from Apple’s playbook.

But Amazon didn’t get to the point of designing smart home products for humans until almost six years after the first Echo was launched in December 2014. When it started giving Alexa the ability to control smart home products, the smart home universe was still a fragmented mess; at that point only tech lovers were excited about turning off their lights by voice. But as adoption has increased, so has the ease of use.

To date, Amazon has succeeded in the smart home by taking an anti-Apple stance; in other words, by opening up its services to anyone in order to see what they like. In fact, I would argue that Apple has lagged in the smart home because its hardware partners found HomeKit difficult to implement and Apple a challenge to work with. This matters because if we want to bring technology to the world around us (and Amazon certainly does) we have to let it grow organically. This isn’t a process that one company can control.

Does this mean Alexa will be embedded in stupid products (hello, twerking Alexa bear)? Yes. Does it mean mistakes will be made that Amazon will have to come back and fix later? Most certainly. In fact, the privacy rules that Amazon rolled out for Alexa are an example of how Amazon is trying to adapt to feedback. Because as a services company Amazon has to listen to its users and respond.

So even if Frames (Alexa in eyeglasses) or Loop (Alexa crammed into a ring) don’t work out, that’s fine by Amazon. The goal here is to put Alexa in as many places as possible so it becomes, as Rausch put it in our interview, the “connective tissue” between services. You can look at other news released at the show, such as the ability to spend 99 cents to get Samuel L. Jackson to replace the traditional voice of Alexa on your devices, as part of this strategy as well.

Jackson’s voice on the Echo is a way to test out and show off Amazon’s Neural Text-To-Speech service, where an AI takes existing recordings of a voice and mimics it with a particular style to make it sound more like a human (or more like Jackson in particular). It’s also a potential source of revenue for Amazon and app developers on the Alexa platform. In the not-too-distant-future, could I pay someone to record my voice and turn it into a digital assistant for my husband? I’m sure he’d LOVE that.

That’s a silly example, but it illustrates how Amazon thinks about taking the things it builds for internal use to make its own products better, then tries to make those tools available to others as a service. For example, last year it launched an easy sign-on service for bringing devices onto your home Wi-Fi or Zigbee network. This year it made that available for anyone to use.

A particularly interesting example of Amazon’s infrastructure building (which could one day become a service) is the Project Sidewalk wireless protocol it talked about on Wednesday. The protocol was designed for the Ring devices that people place in their front yards. At issue was how Wi-Fi didn’t always cover the area well and Bluetooth couldn’t make it at all. So Amazon tooled the existing unlicensed 900 Mhz spectrum and built a proprietary, and secure, mesh protocol on top.

The protocol supports low data rates, can extend over half a mile, and doesn’t require a lot of power. Rausch didn’t have a lot of details, but said that the radios are already in existing Ring products. He added that a current meshed network of 700 devices in the LA Basin has managed to create a network dense enough for Amazon to build a dog tracking device called Fetch to see how well one might track something over the mesh.

Rausch says Amazon may open up the protocol to others in a year, but was pretty vague on details about chips and the protocol itself. But make no mistake, the infrastructure Amazon has built so its ring hardware can communicate outside has the potential to become a long-range, low-power network for a variety of IoT devices. Just like we have a ton of devices that now incorporate Alexa, we may one day have a ton of devices that incorporate Sidewalk. And if Amazon can see a way to build a network like that, it can certainly see a way to build products and services that would run on top of it.

Is Sidewalk a threat to Sigfox, LoRa or NB-IoT today? No, but it certainly has the potential to be, and for everyone out there clinging desperately to the idea that Amazon is selling products and not services, or that don’t understand the importance of building an ecosystem where everyone wins, the IoT is going to be a rough road. Amazon gets it.