Amazon Wants Alexa to Leave Home and Learn to Forget
What Amazon’s Echo Updates Tell Us about the State of Voice-Enabled Tech
The last time we wrote about voice interfaces, Amazon had just put Alexa in a microwave and smart speakers were forecast to be in half of U.S. households by the end of 2019. Fast forward to now, Amazon has unveiled an updated lineup of Echo speakers, introduced new features to Alexa, and put it into a smart oven and a trio of wearable devices. And yes, smart speaker adoption has hit half of U.S. households as predicted, thus cementing the idea of voice-enabled devices as a viable channel to reach consumers at home. Overall, voice technology seems to be keeping its growth momentum throughout 2019 so far, staying on track to become an important way that we interact with personal devices and digital services.
In reality, however, the actual applications of voice interfaces have yet to land a use case that has truly moved the needle. Voice shopping, long heralded as a killer use case for voice-enabled devices, has yet to catch on despite early hype and support by retail heavyweights like Walmart and Target. According to a recent survey by market research firm GfK, although 74% of consumers are aware of voice shopping, only 8% have actually tried it and will continue to do so.
As adoption of smart speakers, voice assistants, and other voice-enabled connected devices continues to ramp up, the service layer that they are supposed to enable has seemingly been stuck in a trough, thus preventing voice interfaces from achieving their full potential. Despite there being over 100,000 Alexa Skills available (and, for comparison, about 5,000 Google Assistant Actions), discoverability of third-party voice experiences remains a major hurdle for users, leaving them to use voice assistants primarily for basic requests, such as playing music, checking the weather, and setting a timer, and barely scratching the surface of the unique experiences and value that voice interfaces can deliver.
To their credit, Amazon is attempting to address this issue by introducing a slew of new Alexa features and initiatives while staying true to its “Alexa Everywhere” strategy. What the company announced this week offers a good reflection of where the voice technology is in late 2019, the key challenges it faces, and some indicators of how the industry is aiming to solve them and push things forward. Let’s unpack the key trends gleaned from Amazon’s announcements and what they tell us about the state of voice technology.
Low Price Spurs Adoption
Altogether, Amazon announced 15 new hardware products at its press event on Wednesday, including updates to its existing Echo lineup. Despite their distinctly different functions and positioning, almost all Echo devices come with a rather affordable price tag that undercuts competitors. The new Echo Dot, which, along with Google Home Mini, was heavily discounted during holiday seasons and responsible for over 50% of smart speakers sold, gets a rudimentary LED display to show time and temperature and will retail for $59.99. The flagship Echo speakers got a nice polish, a new blue color option, and will retain its $99 price tag. Even the most high-end Echo product, named Echo Studio, which boasts Dolby Atmos and 3D audio, will be available for just $199. In comparison, both Apple’s HomePod and Google Home Max are currently priced at $299. In addition to Echo speakers, a new Ring indoor camera will be available for just $59.99, making it the cheapest Ring product so far.
It is clear that Amazon is less concerned with making a profit on selling these Alexa-powered hardware products and more so with using affordable pricing to lower the entry barrier and help it conquer the market share. This strategy, coupled with Amazon’s first-mover advantage, has worked well for the Seattle company so far. As of last month, Echo speakers maintained 70% market share of the U.S. smart speaker market, according to the latest data from CIRP, with Google a distant second with 25% market share and Apple’s HomePod at a paltry 5%. To a lesser degree, the relatively large variety of Alexa products also contributes to Amazon’s dominance of the U.S. smart speaker market.
The reason why Amazon can afford to sell Alexa devices at cost is that the true value of gaining an Alexa user remains elsewhere in Amazon’s ecosystem. Part of that is using Alexa as a data-collection channel to learn more about its customers and their behaviors at home, and using those insights to fuel Amazon’s many other ventures. Part of that is hoping to turn Alexa into the software layer of smart home to control consumer access in the long run. Even though voice shopping has yet to take off as Amazon has hoped, it certainly doesn’t hurt to control this crucial choke point for the future of IoT commerce. Therefore, low profit margins on Echo devices is the price Amazon is more than happy to pay in exchange for market share.
This cost-driven growth strategy also points to a deeper issue with the current wave of smart speaker adoption. Most people bought one (or two) not necessarily because they are sold on their usefulness, but rather because they are novel and cheap. In other words, a sizeable portion of the smart speaker users today are primarily treating their smart speakers as a low-cost tech toy to handle basic tasks rather than a voice interface to explore and experiment with. In this regard, low consumer interest, coupled with the aforementioned discoverability issue and the failure of platform owners to help developers properly monetize their voice applications, are stalling the growth of voice platforms.
Pushing Alexa Out of Home
Beyond the standard Echo device update, Amazon also added Alexa to a rather experimental set of wearable devices, including wireless earbuds, an eyeglasses frame, and a clunky ring. Perhaps it is better to categorize them as “hearables” since they don’t come with a screen and rely entirely on Alexa to function. The Echo Buds — which incorporate Bose’s noise reduction technology and is available for preorder for $129 — are Amazon’s answer to Apple’s AirPods, and are clearly a more viable product compared to the latter two. The Echo Frames are eyeglass frames fitted with directional microphones and speakers to solely bring hands-free Alexa to the sides of your head, while the Echo Loop is a bulgy black ring with a haptic engine that lets you wear Alexa around your finger. Both are billed as what Amazon calls “Day One Edition” products: early devices that aren’t quite ready for a wide release and will be sold via an invite-only program.
One key weakness of Alexa is its lack of presence outside the home due to its lack of mobile integration. Although Alexa is available through its own apps on both iOS and Android, Siri and Google Assistant, respectively, remain the default, integrated choice for smartphone users. Alexa may have conquered the voice-at-home space, but it is losing valuable data and access points in other scenarios. With this in mind, it makes sense for Amazon to eagerly rush out this trio of Alexa-enabled wearable devices in an effort to increase Alexa’s accessibility outside home environments, despite two of them feeling a little half-baked. While Amazon is commendable for their willingness to experiment, this does call to mind some previous Echo products that were rushed out of the gate and received poorly, such as the Echo Auto and Echo Look.
It is also interesting to consider Amazon’s decision to skirt around smartwatches, the most popular wearable product category today, and opt for screenless, voice-only wearable devices instead. Sure, some Android watch manufacturers have been taking advantage of Alexa’s open APIs and integrating Alexa into their products, but the fact that Amazon has yet to do that themselves shows that Amazon is smartly choosing to focus its R&D resources on innovative hearable devices rather than attacking Apple Watch’s dominance in the smartwatch market head-on. In order to get Alexa to leave home, Amazon will have to help it find its own path, and hearables, along with the Echo Auto that came out last year, are now part of its grand plan to expand the reach of its voice platform.
If Amazon wants to convince consumers that Alexa isn’t just for the living room and it can play a role in all aspects of our daily lives, then it has to get the mobile OS on board, since few wearable devices today can run independently without a smartphone. Curiously, the Echo Frame won’t work with iPhones to begin with, hinting at a possible wall that Amazon ran into with Apple. Even the more open platforms of Android and Wear OS don’t always offer the level of customization that device makers want.
Alexa Everywhere in a Walled City
Experimenting with hearable devices is just one of the many things Amazon is trying in order to expand Alexa’s reach as it holds steadfast in its “Alexa Everywhere” strategy. As an update to its Alexa-powered Microwave announced last year, Amazon unveiled an Alexa-enabled smart oven that also doubles as a microwave and an air fryer. The acquisitions of video doorbell company Ring and mesh-network maker Eero bore fruits in the form of a lineup of Alexa-equipped connected security cameras, video doorbells, and Wi-Fi mesh routers. All these non-speaker home products further build out a smart home ecosystem enabled and connected by Alexa, further solidifying Amazon’s stranglehold on the home platform and all the data users generate.
Not content with just inside the house, Amazon is also looking to expand the coverage of its smart home platform to include the entire premise of your house. With a new Amazon Sidewalk program, which establishes a low-bandwidth, long-range network over the free 900MHz band of the radio spectrum, users may be able to connect to compatible devices from “up to a mile away.” While this is initially designed for users to take control over their smart gardens, outdoor lights, and mailbox sensors, Amazon also created a connected dog tag that will leverage the Sidewalk network to alert owners if their dogs ever leave the perimeter.
Moreover, Amazon announced a partnership with GM that will bring Alexa to millions of its vehicles’ OEM infotainment system, including 2018 and newer Buick, Cadillac, Chevrolet and GMC models, which further expands Alexa’s native integration with connected cars. GM research showed many customers prefer to use the same voice service in their vehicles that they use in their homes. Leveraging the Alexa Auto Software Development Kit, GM offers customers access to standard Alexa skills and capabilities that they already use at home, but is also able to create custom skills relevant to a customer’s vehicle.
Interestingly, the most ambitious bid to make Alexa ubiquitous that Amazon made this week was the Voice Interoperability Initiative it announced on Tuesday. This program, which unites Alexa with a number of third-party voice assistants into one big alliance with the notable exception of Apple, Google, and Samsung, aims to make it possible to talk to a variety of voice assistants such as Microsoft’s Cortana, Salesforce’s Einstein, Spotify Voice, and many more to get specific tasks taken care of without having to own multiple devices. Simply saying the appropriate wake word would trigger the proper assistant to respond. It is also worth noting that Baidu and Tencent, whose voice assistants Xiaodu and Xiaobing, respectively, lead the Chinese market along with Alibaba’s Tmall Genie, are also on board with this initiative, hinting at a crossover play for Amazon to serve the Chinese-speaking market without having to develop mandarin capabilities for Alexa themselves.
Of course, Amazon is not opening up its huge base of Alexa-compatible hardware devices to all these smaller voice assistants out of pure altruism, or some grand vision of kumbaya unity amongst voice assistants of the world. Rather, by forging an alliance with all these 30+ other voice assistant with limited capabilities, Amazon is leading a pack that positions Alexa as the “general-purpose” ringleader, a pack that will further enrich its own voice platform by drawing from the various specific skill sets of each digital assistants, so that it can better compete against the market leaders Siri and Google Assistant. In a clever move, Alexa is looking to close out its main competitors by opening up its voice platform to band together all the voice assistants that won’t pose any meaningful threat to its market dominance.
Compared to Apple’s walled garden approach, Amazon wants to build a walled city where Alexa reigns. But that doesn’t mean that the smaller voice assistants won’t be able to work with other voice assistants, just priority of access. For example, it announced on Thursday that Microsoft will add Google Assistant support in English for the Xbox One in beta, with the ability to launch games, take screenshots, and more. By all accounts, this is not a real solution to the interopratity issues that users of multiple major voice assistants will inevitably run into. Still, it is a compromise that makes sense for the market landscape today, and a necessary step to make the future of device-agnostic ambient computing a remote possibility.
Alexa Gets Smarter and Learns to Forget
At the end of the day, the quality of voice experiences that Amazon can offer users depends more on Alexa itself than the hardware. In that regard, Amazon also added a bunch of new features to make Alexa more understanding, more versatile, and more trust-worthy.
For example, Alexa added support for multilingual mode that will allow multicultural households to interact with Alexa devices in two languages simultaneously, as well as a Communications for Kids mode, which will let children talk to each other through Alexa devices under parent supervision. Extended “Alexa Hunches” will preemptively help users troubleshoot their IoT devices while crowdsourced answers via Alexa Answers will hopefully help it get better at answering random but general questions without resorting to a dry list of web search results. On the developer tool side, a new skill personalization tool in the Alexa Skill Kit will allow voice developers to deliver customized information based on who is speaking with Alexa Voice Profiles, while the addition of an Education Skill API will allow developers to easily create voice interfaces for education applications.
All these new features aim to make Alexa a more helpful home assistant that also happens to appear more “human” and fun. In that vein, Amazon will also start selling “celebrity voices” as alternative add-ons for Alexa, starting with Samuel L. Jackson and with more celebrities coming next year, to make Alexa more customizable and engaging. Similarly, a new Alexa Frustration Detection feature can now detect when you’re getting frustrated with Alexa getting your requests wrong, after which it’ll try to apologize.
Earlier this year, Amazon came under fire for hiring people to occasionally listen in on Alexa conversations. Though implemented as part of a legitimate effort to help improve Alexa’s understanding of human speech, it was not properly disclosed to Alexa users and caused major privacy concerns. As a result, Amazon added several new commands to tamp down privacy concerns, such as opting in ongoing, rolling deletion for recordings in 3 and 18-month intervals or for troubleshooting weird Alexa behavior like unprompted activations. Both Google and Apple faced similar outrages over privacy voice recording as well, and they will likely follow Amazon’s lead here to come up with privacy-minded features to appease the increasingly alert user base. Data privacy is quickly becoming a fundamental test of consumer trust, and all voice tech players are realizing how important it is to the long-term future of voice assistant usage.
Voice Computing in Late 2019
With this Echo event, Amazon tried to address a lot of the issues that are bubbling up in the voice technology space, such as voice’s lack of out-of-home usage, the interoperability issue, and privacy concerns. But these efforts are largely buried by the frenzy of announcing 15 new products at once, and the reason for such a “throw-everything-at-the-wall” approach hints at Amazon’s eagerness (some would call it desperation) to figure out the right use case for voice interface both at home and beyond. Ultimately this stems from Amazon’s lack of control over mobile and wanting to establish Alexa as a platform in preparation for a post-mobile era. Ironically though, Amazon’s lack of attention towards AR may leave them dependent on integrating with other platforms again in the post-mobile future. After all, a voice interface alone won’t be enough for all use cases, especially if AR becomes the next computing paradigm.
As of late 2019, the voice computing market is shaping up to be a one-on-one battle of Amazon versus Google
As of late 2019, the voice computing market is shaping up to be a one-on-one battle of Amazon versus Google as both companies appear to be going after the same thing: pushing data capturing services in the home. So far, Amazon is leading the market because they accepted the low-margin costs to gain market share, but it is not a sustainable ecosystem long-term if they don’t figure out the right use cases for voice tech beyond the basic things they can do today. About 70% of people will gradually use voice assistants to replace visits to a store or bank in the next three years, per a recent survey by the Capgemini Research Institute, but that will not come to be if stores and banks have little incentive to develop useful voice experiences to engage with customers. Plus, our social norm will also need to evolve if we wish to keep expanding the use case of voice assistants and enter the age of hearables and voice-enabled ambient computing. Voice tech is gradually slipping down the trough of disillusionment, but the major players are already looking for the slope of enlightenment to lift us up.