Two Possible Futures for Amazon’s Alexa

Amazon’s Alexa voice assistant was clearly the star of CES this year. No single consumer electronics device dominated coverage but lots of individual devices incorporated Alexa as their voice assistant of choice. The announcements ranged from Echo clones to home robots to cars and smartphones. It was clear Amazon had entirely captured the market for voice platforms. Only one or two integrations of the Google Assistant were announced and those are both future rather than present integrations.

It would be easy off the back of all this to say Amazon had won the voice assistant battle once and for all but I actually see two possible futures for Alexa, with very different outcomes for Amazon and its many partners.

Future 1: Amazon continues to dominate

The first possible future for Alexa is one where the current trends mostly continue and even accelerate. Amazon’s own Alexa-based products continue to sell well, with Dot probably taking a greater share of sales going forward relative to Echo (or Tap), selling into the tens of millions of installed units in the next couple of years. On top of that, the adoption by third parties that was so evident at CES continues, with even more devices offering integration. Importantly, Alexa starts to make an appearance in Android smartphones, making it as pervasive and ubiquitous as existing smartphone-based assistants, possibly even making an appearance in another round of Amazon smartphones.

What we end up with in this scenario is a massive ecosystem of devices which all offer users access to Alexa and its functionality. These devices perform their functions well, recognizing voice commands effectively, responding appropriately, and adding value to users’ lives. Because they’re all part of the same ecosystem, they work the same way — commands issued through one are reflected on the others. Amazon benefits from owning a massive new user interface and platform which can be used not just to push its e-commerce sales but to take an increasingly large share of media and content consumption across video, music, audiobooks, and more.

This scenario also assumes major competitors either don’t launch competing products or those products fail to take off. Google has, of course, already launched its Home device but, thus far, sales are far lower than Amazon’s and are handicapped by a lack of awareness and the lack of a major e-commerce channel. The Google Assistant, meanwhile, should be the default option for Android OEMs in all their devices but the way in which Google has held it back as it promotes its own hardware has also held it back, perhaps fatally, as a third party voice platform. If that doesn’t change, if Microsoft’s Windows-based Cortana strategy falls short, and Apple’s reticence to participate in this market continues, Amazon dominates with its devices and those third party products using its ecosystem.

Future 2: Cracks start to appear in the Alexa ecosystem

The secrets of Echo’s success

I want, though, to paint an alternative future for Alexa, one which is less rosy and more complex. Amazon’s genius in launching the Echo and Alexa was to pick a blank slate rather than an existing category for its experiments with voice. Instead of competing with another smartphone-based voice assistant, Amazon chose to compete in the home, with its relative quiet and better internet connectivity, and a device that was optimized for specific use cases: fantastic voice recognition and great audio output, even from across the room. That had two major advantages: first, it wasn’t going head to head with powerful entrenched competitors and second, it could deliver far better performance around voice recognition than smartphone-based systems.

The Echo performs fantastically well at what it does. Its voice recognition is indeed very good, inviting highly favorable comparisons to Siri and the like. It’s this success in providing great voice experiences that have propelled sales of its own devices and prompted other companies to build their own as well. The assumption on all sides is it’s Alexa that powers these phenomenal experiences and the Alexa Voice Service for device makers will power similar experiences on other devices.

Amazon’s limited control over Alexa devices

However, one look at Amazon’s guidelines for those wishing to incorporate Alexa Voice Service into their devices should prompt at least some skepticism. Echo and Echo Dot famously have a 7-mic array built into the top of the device, with beam forming, enhanced noise cancelation, and more helping to ensure the device does a phenomenal job of picking up your voice from up to 20 feet away. But look at the minimum specs for Alexa-powered devices and you quickly realize many of these devices won’t match up on hardware – the minimum standard for microphones is just one and additional technologies like noise reduction, acoustic echo cancellation, and beam forming are entirely optional.

Also optional is “wake word” support – in other words, the always-listening function that waits to hear Alexa (or another word of the user’s choice) and then springs into life. The Amazon Tap doesn’t offer this feature (and was hammered in reviews for it) because the “across the room” use case is a key part of Echo’s appeal. Even when a wake word is supported, Amazon only requires a minimum of one microphone for near-field recognition and just two for far-field (20 foot) recognition.

Where this second future scenario diverges from the first is the sheer range of Alexa-enabled hardware starts to put many devices into the market that don’t have nearly the appeal of Amazon’s own. Lenovo’s Echo clones appear to be using an 8-mic array and may very well perform at exactly the same level or better but the Huawei Mate 9 smartphone, which is due to incorporate Alexa later this year, has just 4 microphones and the device obviously wasn’t built with optimal voice recognition in mind. In a rush to get products to market, we’ll see many vendors putting out devices with the bare minimum specs and prominent Alexa-related branding.

All it will take at this point is a handful of terrible reviews for Alexa-powered speakers and other devices and the Alexa brand will quickly become tarnished. At that point, Amazon’s admirable openness with the Alexa tools may come to be seen as a huge mistake because it’s set so few limits on what can be done with the service and its brand. Even if third parties are committed to providing the best possible experience, voice recognition on a smartphone or other smaller devices likely is never going to match up to the Echo’s quality, which means the true Alexa experience will likely remain elusive outside of the home. Any perceived quality advantage will, therefore, fade as well, making Alexa a lot less appealing.

More compelling competitors

Meanwhile, competitors will move past their slow start in responding to the Echo and Alexa and will begin producing more compelling alternatives. I see no reason why competitors shouldn’t be able to build devices which perform at least as well, in terms of voice recognition, as Echo given the same parameters (home use, large devices, mic arrays designed for voice recognition). Indeed, Google’s Home has already demonstrated there’s no special magic there. In addition, players like Google and Apple have one huge advantage – they already own massive installed bases of hundreds of millions of devices running their operating systems and integrated voice assistants.

Google’s early misstep in limiting the Google Assistant to its own devices will be overcome in the next few months as it makes it available to Android OEMs more broadly and, at that point, its Home device will become a lot more compelling. Apple, too, has the potential to do really interesting things in the home speaker space should it choose to do so, given the increasing scope and availability of Siri and its AirPlay audio and video casting technology. Again, the appeal of using the same assistant everywhere, tightly integrated into devices, will be a big advantage over Amazon’s looser Alexa ecosystem.

Which future plays out?

On balance, I’m inclined to think the future will look rather more like the second scenario I’ve painted than the first. That is to say, I think Amazon’s advantages in the field of voice assistants are mostly temporary and, to some extent, illusory. Competitors will catch up fast in the home and exceed its capabilities outside it. That doesn’t mean Amazon can’t build a decent business with a more limited scope of opportunity around its first party devices and a handful of really compelling third party devices in an ecosystem but I suspect its future will be a lot less bright than its present in this space.

Published by

Jan Dawson

Jan Dawson is Founder and Chief Analyst at Jackdaw Research, a technology research and consulting firm focused on consumer technology. During his sixteen years as a technology analyst, Jan has covered everything from DSL to LTE, and from policy and regulation to smartphones and tablets. As such, he brings a unique perspective to the consumer technology space, pulling together insights on communications and content services, device hardware and software, and online services to provide big-picture market analysis and strategic advice to his clients. Jan has worked with many of the world’s largest operators, device and infrastructure vendors, online service providers and others to shape their strategies and help them understand the market. Prior to founding Jackdaw, Jan worked at Ovum for a number of years, most recently as Chief Telecoms Analyst, responsible for Ovum’s telecoms research agenda globally.

20 thoughts on “Two Possible Futures for Amazon’s Alexa”

  1. An analyst looking at Alexa so closely…
    I guess it’s going to be a “thing”. I’ll wait.
    Far from being from luddite circles, I don’t know anyone who actually has one, but that’s just me.

  2. I am not sure that putting an exactly the same agent on a smartphone as in a home is a perfectly good idea.

    A smartphone agent is more personal, a home agent is shared. A smartphone agent knows more about you, a home agent knows more about the world. A smartphone agent is more transactional, a home agent is more conversational. Unless the software can learn from the experience (can it?), I think a separation between the two such as Google Assistant – Google Home makes more sense. Of course the two would be able to talk between themselves…

    1. Very good writing, but let’s not confuse the business of tech, with tech itself.
      Not one equation or phenomenological description… Sheesh! 🙂

        1. Renting the stadium and announcing the lineup is not the same as actually playing the game. These business things can apply to widgets, not just tech.

          As I said the writing is good. There’s the business of tech and then there’s actual tech.

  3. I can’t bring to mind any scenario in which room-limited, home-speakers will be a better solution than a wearable. Once the wearable becomes good enough at hearing, etc, the party will be over for devices that aren’t “always-where-you-are”.

    Toward that end, Apple offers Watch and AirPod. Wrist and ear. Still untapped are the shoulder and the necklace area. With tech improvements, one of these four areas will please everyone and then today’s Alexa will have lost.

    1. It’s not about the device though, it’s about the service.
      Like most, I don’t have a smart watch, and no intention to get one. On the other hand, a smart speaker I don’t have to wear+charge, can use as a plain speaker too, have a few of already in the form of TVs and PCs… sounds less disruptive and more widely useful.
      Plus what happens to the soup in your spoon when you’re eating and want to ask iWatch-Siri something ? Or the book in your hands ?

  4. Your two scenarios approach is a brilliant way to balance perspectives and opinions. I like the way you end up being negative, but prove that you’ve also considered (and dismissed) the optimistic option.

    One thing that I miss with all the analyses on Echo that I’ve seen, is an evaluation of the strength of network effects for a Voice UI (or AI in general) for the home.

    1. The current crop of voice UI lends itself to only simple tasks. Therefore, 3rd party developers probably don’t have to write much code. This means supporting multiple platforms will be easy and that developer lock-in will be minimal.
    2. Regarding the control of smart devices, none of the major players in tech, with the exception of Samsung, have a significant presence in home appliances. Therefore, I believe ecosystem lock-in to also be minimal.
    3. Amazon could potentially use the largest online store to lock customers in to the Echo. Amazon may decline to offer integrations with Google Home/Assistant or Siri, in order to push the Echo. This does feel unlikely though.
    4. Assuming that AI really does need to understand everything about your life to be useful, then Google may be able to lock users to its AI platform. I personally don’t take this view, and I don’t think my smart home needs to know which websites I tend to visit, for example.

    Whether or not the Echo’s strong initial lead will end up winning the race for Amazon, depends on the strengths of the network effects for Voice UI/AI.

    1. 1. Sonos just leaked their intention to do just that.
      2. LG also does; Xiaomi is strongly pushing that way (they recently said they make more money with their non-smartphone lines), LeEco (bought Vizio) too.
      3- That’s more of an Apple move, see the AppStore forbidding any mention of competing ecosystems.
      4- Or does it ? If I tell it to “read me the news”, don’t I want them read off the website I go to for news ? If I request a recipe, shouldn’t it be the recipe I recently… googled then clicked on ?

      1. You bring up some very interesting points, and I think this is the discussion that is necessary to truly understand the Amazon Echo and copycats.

        1. It would be interesting to see if more companies also announce their intent to support multiple voice UI platforms. As I noted above, an analysis on how much code a developer actually has to write to support the Amazon Echo would also be helpful.

        2. LG certainly crossed my mind. I was unsure however, whether I should include it in my definition of “major players in tech”. The same applies to Xiaomi, LeEco, Sony and others. Of course, it might not be necessary to be a “major player in tech”, but instead it might be more important to be a “major player in home appliances” for the purpose of ecosystem lock-in.

        3. I agree. This is an Apple-ish move. Which is why I doubt that Amazon or Google would use this approach.

        4. People have different opinions on how much data is necessary for an AI assistant. I take the view that “Alexa skills” that read you news headlines should have an idea of your news preferences, but should not know what food you like. Conversely, recipe “Alexa skills” should know what food you’re interested in, but should not need to know what news you regularly read. Therefore I think that the advantage of a know-it-all AI like Google compared to domain specific AI like Apple News or Flipboard is actually slim or non-existent.

        1. Let’s dig into recipes, ‘coz I’m planning the week’s meals right now ^^
          Over the years, my recipes have ended up in 3 spots:
          – a text file, that evolved into a gdoc, of tried-and-true stuff mostly got from friends and family
          – another text file -> gdoc of rather random stuff I came across from various sources I’d like to try out some day, or that started that way and is now not quite a staple but no longer experimental.
          – an account on France’s leading cooking website, where I look for specific stuff that’s outside my family and friends’ culinary scope ,that is either a staple or on the to-try list.

          A cooking or shopping assistant is of very little value to me if it can’ t a) access those recipes b) make sense of them (they’re so summarized it’s almost shorthand).

          I could change my system, various cooking sites let you upload recipes, will generate shopping lists and prep instructions/checklists… except I haven’t done so up to now, probably never will, and am not considering doing the equivalent for every little thing that AIs want to meddle with (news, music, finances, health, fitness, packing for trips, planning for trips,…).

          I think that gives an advantage to more intrusive AIs. They kind of have to become able to help me in spite of myself. I’d be less bothered by Google spying on my recipes (et al.) than by having to import them into Alexa.

          1. You are truly amazing!

            Regarding recipes, I suspect the cooking websites will be the ones that come to Alexa, and friends. The one we have in Japan (CookPad) is very profitable, has an extremely good business model that integrates with both food manufacturers and supermarkets, is wildly profitable, has strong social features, has 2.6 million recipes and has a great website and mobile apps.

            I don’t see any way Google can compete with this.

Leave a Reply

Your email address will not be published. Required fields are marked *