Amazon – Show and tell.

Reply to this post

RFM AvatarSmall

 

 

 

 

 

Screens help to alleviate digital assistants’ stupidity.

  • I think that the Echo Show is more about addressing the shortcomings of voice interaction with machines than it is about launching a series of new and exciting Digital Life services.
  • Amazon has launched an ugly looking device called Echo Show that is effectively Alexa with a 7 inch screen attached to the front.
  • The form factor is disappointing as even Baidu with no hardware experience managed to come up with a far more appealing looking product (see here).
  • Amazon has also upgraded the speakers to give a louder and richer sound profile but I see this being about giving Alexa another medium with which to communicate with the user given the limitations of voice.
  • The problem is simply that Alexa (and all other others) are far too stupid to be able to hold a meaningful conversation with a user.
  • Google Assistant is currently the best but remains woefully short of what one would consider to be a useful assistant.
  • Digital assistants were designed to replace the human variety but because their intelligence is so limited, they are unable to hold a coherent conversation with the user.
  • Human assistants do not need to use screens to understand requests, relay information and carry out tasks meaning that the perfect digital assistant should not either.
  • Hence, I think that the Echo Show has been created to make up for the huge shortfall in Alexa’s cognitive ability
  • This type of interaction is what RFM refers to as one-way voice where the user asks a question and the results are displayed on a screen.
  • RFM research has found (see here) that the vast majority of all man to machine interactions are one-way voice and with this device, Amazon makes these interactions easier.
  • Furthermore, for those that depend on advertising having a screen also helps to maintain the business model of lacing a Digital Life service such as Search or Social Networking with advertising.
  • Consequently, I think that Google is likely to follow up with a similar product which will take advantage of the fact that the necessary communication apps that the device will use are already installed and ready to use on all new GMS Android compliant devices.
  • In Alexa’s case, it looks like the user will have install another app on his phone in order to communicate with the Echo Show.
  • The Echo Show will come with all of 12,000 Alexa’s skills but these skills have been designed for a device with no screen and so I do not see the screen improving the already very poor user experience that these skills currently offer.
  • At $230 or two for $350, the Echo Show is priced to sell but I think that volumes will be small given that the vast majority of Echo’s shipments are made up by the cheapest member of the family, the $50 Echo Dot.
  • Hence, I do not see a sudden rush by developers to upgrade their existing skills or develop new ones to make use of the screen.
  • This is where Google Assistant has a huge advantage as it has already been designed to run with a screen (smartphones) meaning that adapting to having a screen on the Google Home product should be much easier and much better.
  • I still think that Google Home has the advantage here as it has a much better assistant than Alexa, but its lack of developer support for the smart home is starting to be a real problem.
  • Google really needs to pull its finger out and show developers love, especially as Microsoft looks set launch something similar to Echo Show but using Cortana.
  • I continue to struggle with Amazon’s share price whose valuation I think demands that investors pay for profits that never seem to materialise.

RICHARD WINDSOR

Richard is founder, owner of research company, Radio Free Mobile. He has 16 years of experience working in sell side equity research. During his 11 year tenure at Nomura Securities, he focused on the equity coverage of the Global Technology sector.

Blog Comments

Baidu are making a 2003 iMac with motor noise that will interfere with the mic and speakers for voice assistant or video calling. Unfortunate.

Google released their assistant to github, which was the move that made Amazon gain such quick developer adoption. To boost this, Google also supported MagPi, a raspberry pi enthusiast publication by distributing hardware (a hardware on top board with mic and speaker) so that subscribers can assemble their own Google Home out of a raspberry pi.

Google is late, but they’re certainly showing developers love as you call it.

My research shows that people use these assistants for kitchen timers, reminders, playing music, and at a distant third, controlling smart homes. Google is ahead because their device also controls the chromecast media (strangely, not when invoked from a phone.)

Adding the screen isn’t about supporting a legacy ad revenue business. It’s about making voice-first shopping easier, something amazon has pushed with its echo-only deals, but users are uncomfortable purchasing sight unseen even when they know amazon has a stellar returns policy.

The recipes skill now can show images as you step through cooking, which is a benefit for hands feee kitchen use, a place we already know the echo is commonly placed in the home.

Video chat is present as are song lyrics because, what else can you do when you have a screen?

wrong sort of developers… hardware will not drive the platforms adoption and see off amazon… better UX and functionality will…they need to tap into the android developer pool.

Agree. Smart Home is 15% of usage. The biggest usage by far is dumb bluetooth speaker which is around 70% of usage. This is where show falls over… there is no cast function…

The skills are awful.. almost universally and in most cases its easier and quicker to perform the function the old way by getting of your arse.

That the opportunity to fix I think and Google is not doing much there as far as I can see which is a shame….

Disagree. Ease of prototyping goes a long way – If you’re Google or Amazon, you want as many people using your assistant as possible. You do that through making useful skills easy to develop, and through making the hardware that runs the assistant ubiquitous. $129 USD Google Home isn’t going to get them there. The third party $49 USD speaker, or the fictional $149 thermostat with it integrated into it, does. You only get that kind of integration by letting all-comers run it on ubiquitous hardware, to make it easy to test the skills against, and make it easy to run on hardware that can be shipped (pi is too costly to integrate in a shipping product, but it’s great for prototyping.)

More skills, more third party hardware, and pretty soon they’ll have matched Amazon.

Dumb BT speaker is not the largest usage. The largest usage is streaming music initiated by voice – no phone paired – whether that’s Amazon Music, Spotify, Pandora or YouTube on Home. Dumb BT speaker is very low on the list, because very few users realize it works that way in addition to the other things it does (exception: Amazon Tap, which screams to people that it does BT.) A big part of the reason for this is Amazon’s onboarding – they get the Echo or Dot on Wi-Fi, but don’t do anything to guide the user to set up Bluetooth. You say, play $SONG from Spotify, and it informs you to set up Spotify in the app. This is how the vast majority of people use it.

Google Home casts to Chromecast, putting it ahead slightly. Amazon has yet to make any of the Echo products aware that FireTV exists, even though FireTV also has Alexa in it. That’s one of their oversights that makes Google better for media consumption. (For some definitions of ‘better.’ You can kick off a show from Netflix, but not specify which episode. “Play Black Mirror from Netflix on my TV” (where my Chromecast has been renamed ‘TV’) works. But that’s as much control as you have, and it has to be in that form. Also, someday Google should realize that no one wants to have a device named Chromecast7402 in their home, and should suggest ‘living room’, ‘lounge’, etc. as a part of onboarding a Chromecast device. I expect them to understand this late; They still don’t think they need to anthropomorphize their assistant. They do.)

I’m curious to know which skills and devices you’ve tried – they’re a mixed bag, not universally terrible, because each one comes from a different developer with different ideas about syntax and desirable use case. The ones that are the worst are the ones with convoluted syntax. The ‘getting off arse and doing it the old way’ is the challenge: any smart home device has to be better or faster than doing it the old way.

We’re at the dawn of this. The next phases are where they get better at recognizing intention without requiring specific syntax. This is coming.