Amazon’s Echo – Intelligent Voice Control for Your Home and Life?

Amazon has been developing an increasingly sophisticated type of voice controller – called the Echo – that allows you, from anywhere the controller can hear you, to simply speak questions and express commands to it and receive spoken answers and responses back.  Want to order a pizza (or many products from Amazon, of course!)?  Wonder what the temperature is?  Wish to have some music playing?  With compatible devices, you can also adjust your room’s temperature, turn on and off lights and appliances, and so on.  If you have Siri on your iPhone (or Cortana in Windows), you’ll already be familiar with the concept.

Indeed, if you’ve ever watched the movie “2001 :  A Space Odyssey” (or countless less seminal movies and television shows over the years) you’ll have already been familiar with the concept of an omniscient all-powerful computer that one can communicate with via ordinary conversation and which speaks back to you in regular plain English.  Is this another part of science fiction lore that is becoming (or even ‘has become’!) science fact?

Amazon doesn’t disclose sales data for its new devices, the same as it has never revealed sales data for its Kindle eBook readers or its Prime membership.  But it does provide hints and clues, and the general perception is that its Echo range of products is proving to be sufficiently popular as to confirm their commercial viability, and Amazon is understood to be committed to increasing and improving its product range accordingly.  Apparently this year’s Black Friday/Cyber Monday, during which the Echo Dot was briefly on special, saw several million more units sold, causing the black version of the Echo Dot to go on back order for ten days (the white version remains available for immediate delivery).

With the recent release of their second generation hockey puck sized Echo Dot, the lowest priced of the units in the Echo family, it seemed time to try the system and service out and decide if it is truly a practical and useful thing, or if it is a gimmick best steered clear of.  Plus, with Siri on iPhones, unnamed general speech recognition for some functions on Android devices, and Cortana for Windows, what does Amazon’s device do that we can’t already do, with other devices, already?

Unfortunately, while the Echo Dot has many appealing capabilities, it doesn’t yet deliver its full potential and promise.  But future enhancements are likely to be in the software, not the hardware, so this need not be a barrier to buying one now if you want to experience the capabilities of the unit.

The Echo Product Range

Amazon’s Echo project and the Alexa branded intelligence that it offers has reportedly been in development since 2010, and the Alexa name has been used by Amazon for many years prior to that, but as an internet website ranking service.

The first unit to be released was the Amazon Echo, which was made available to Prime members in November 2014, and more broadly to everyone in June 2015.  The Echo can be thought of as being much the same as a Dot, but with a bigger/better built in speaker, and larger in size.  On the other hand, the Echo does not have an external speaker output, so it is a ‘six of one/half a dozen of the other’ tradeoff, although the Echo is very much more expensive – $180 vs $50.

The next model to come out was the Echo Tap, which used battery power rather than mains power, but the necessary sacrifice involved there is that the device is not always on and ‘listening’ for commands (if it was, its battery would quickly die – the battery only lasts nine hours).  You have to turn it on before you can use it.  The unit has a somewhat better price – $130.  Our guess is that after the first rush of early adopters were satisfied, Amazon has been finding it difficult to sell Echos for $180.  The same amount of money would buy three of their Fire 7 tablets or two of their Fire 8 tablets, making it an unusually expensive piece of gear and out of line with other ‘goodies’ they sell.

For most people, the essential element of the Echo concept is the ‘always on’ nature of the unit.  We suspect the Tap has not proven to be as successful as the other models in the Echo line because it lacks this feature, and note Amazon seems now to be selling it more as a Bluetooth speaker, with its Alexa/Echo functionality somewhat downplayed.  The thing about the ‘always on’ feature is that if you don’t have this, you may as well use the same app or a competing app on your phone.  The ‘no need to do anything, just speak’ feature is the major advantage of the Echo.

The third and newest model is the Echo Dot.  This first came out earlier this year and was (over)priced at $90.  A month ago it was re-issued in an updated form, slightly smaller in size and with slightly better microphone capabilities; best of all, the price dropped to $50.

This, to our mind, is by far the best of the three units.  It is very much smaller than the original Echo, and very much better priced – $50 instead of $180.  It also has a Bluetooth speaker connection   There’s no way that a slightly better speaker adds another $130 of value to the regular Echo; if you’re wanting to experiment with an Echo, get a Dot.  Indeed, get two or three, and you’re still paying less than for one regular Echo.

Amazon seek to encourage the notion of buying multiple devices – the idea is to have one within hearing distance of you wherever you are in your home, so you never need to think, you just know that any time you wish, you can simply speak a question or command and an Echo/Echo Dot will hear the command and respond.  They sell packs of Echo Dots on the basis of ‘six for the price of five’, and are currently also offering quantity discounts on orders of three.

What Echo Does

An Echo unit (it matters not whether it is an original Echo, an Echo Tap, or an Echo Dot – unless we need to specify which, we’ll just generically refer to them as ‘Echo’) is always listening for commands.  When it hears its ‘wake word’ (initially set to ‘Alexa’ but can be changed to either ‘Echo’ or ‘Amazon’ – because you’re less likely to be saying ‘Alexa’ in normal speech, generally it is best to leave it at that), it immediately ‘wakes up’, hopefully understands what you just said, and responds.

You can ask the unit questions.  You can ask it to play music.  You can ask it to do things.  You can ask it to control other devices.  You can ask it about the weather – either where you are, or somewhere else, and for news updates.

If asking it to play music, you can simply name a preferred radio station, and if that radio station streams via one of several different streaming services, it will play it.  If you have a Pandora or one of several other music streaming service accounts, you can ask it to play that service.  If you have music as part of your Amazon Prime services, you can have it play music from that.  Or you can just specific a type of music and let Alexa make the best choice for you, or even just say ‘Alexa play music’ and it will try to guess what you might like.

You can ask it to set an alarm, either for a certain period of time ‘Alexa set an alarm for three hours’ or to go off at a certain time ‘Alexa wake me at 7am’.  You can ask it to add items to a To Do list, or a shopping list.

If you have any of a growing number of intelligent home automation products that connect to the internet via Wi-Fi, the chances are Alexa can control them.  ‘Alexa, turn on the upstairs lights’.  ‘Alexa, raise the temperature two degrees.’  And so on.

You can summon an Uber or Lyft car, and you can order a pizza from Dominos, or flowers via 1-800-FLOWERS for a friend, and you can ask Alexa to convert units.  You can ask Alexa details about movies (including what might be currently playing in your area), phone numbers, restaurants and other local information.

And, yes, there’s more, much more.  Think of an Echo a bit like a smart phone – it comes with some built in apps, and then you can load as many more apps as you wish onto the phone.  In the case of an Echo, the apps are called ‘Skills’ and there are over 4000 such skills available, with more being added literally every day.  Here’s a list that completely erroneously bills itself as being a complete list of everything Echo can do, and here’s a more truly complete list.

The Need for a Screen

The essential concept of an Echo device is that you can just casually talk to it and have it talk back to you.  In theory, you don’t need to remember specific command words or phrases, but just have a dialog with the unit.

In real life, with the people around us, we can have conversations and seldom need to usual visual graphics as part of the conversation.  ‘Where shall we go for dinner tonight?’  ‘Oh, I don’t know, what do you feel like?’  ‘How about Thai?’ ‘Yes, that would work, but which Thai restaurant should we go to?’  ‘How about that new one on NE 8th Street?’.  And so on.

With an Echo, it is a bit more simplistic, of course.  I asked it for recommendations for good local restaurants, and was offered a Subway type store, and two private Microsoft campus cafeterias only open to Microsoft employees on the basis that the restaurants were popular.

When I did get some further specific restaurants, it directed me to its companion app for details.  And therein lies the Achilles Heel of the Echo devices.  No screen.  Anything that can’t be quickly and easily spoken is better conveyed with images or written details, and Alexa directs you to go to its app for such information.  Their app can work on iPhones/iPads and Android phones/tablets.

This is not a bad thing, per se.  But it means that the promise/ideal of the Echo unit – something that you can just talk with, conversationally, is compromised.  Sure, you can talk to it and it can reply, but sometimes, things are better done ‘in writing’.  With the restaurants, it then provided a page of restaurant information, jammed full with data – addresses, maps, customer reviews/ratings, and so on.  It would have taken many minutes to speak all that, and by the time it got to the third entry I’d have already forgotten the first couple of entries!

Amazon sort of concede that an Echo would be better with a screen, and are rumored to be developing a new device with a 7″ screen as part of it.  That is semi-good; cost-wise, it shouldn’t make a lot of difference, when you see them selling 7″ Fire tablets on special for less than $40, one would hope that adding a screen to an Echo would therefore add less than $40 to the device price.  But if you need to match your conversation with viewing text on a screen, you’re not going to be able to call out from the comfort of your arm chair to the device on the far side of the room for information and have it speak the reply back to you.  You’re going to have to get up and walk over to the small screen to view what it shows.  Or, of course, pull out your phone, turn it on, move to the Alexa app, wait for it to load, and view the information there.  The convenience factor just massively nose-dived, and by forcing you to a phone, the Echo service now has to compete with every other service you might have on your phone.

There is already a sort of a screen based Echo available now, for free.  On an Android device, you can load the Alexa app, and then communicate with it by pressing and holding the home button, exactly the same as you would with an Echo device.  It isn’t quite as convenient – you have to turn the device on and then hold the home button before you can start talking, but it is already available on devices with screens.

How Well Does the Voice Recognition Work?

The voice recognition is very good indeed.  Not perfect, of course, but very good.  While it works well ‘out of the box’ you can greatly improve it by going through one or more training sessions via the associated app.  If you move the device from one location to another, the acoustics will change and you should consider retraining.

Here’s a possibly counter-intuitive trick to working with voice recognition systems.  They often work better when you run your words together – ie, when you speak normally.  If instead you ‘speak like a robot’ and hesitatingly sound out each syllable and pause between words, that can actually detract from the way the voice recognition works.  The voice recognition already knows that when you say something like “this way’ you probably actually say “thi sway”, and it also gets contextual clues as to what each word is most likely to be by the words that are included around it.  ‘The car was stationary while the cyclist rode past’ is more likely to be understood as the car was stationary/stopped than the car was stationery/(office supplies) because it makes more sense in the first context.

So speak normally to your Echo.  But perhaps remember what your mother taught you, and don’t speak with your mouth full.

The other part of this issue is how ‘free form’ can you phrase your questions.  Do you need to remember any particular code words or phrases?  Can you speak your commands in any sequence, or do they have to go in a ‘subject verb object’ sort of rigid sequence?  The good news is that the Echo devices support something very close to fully free form commands.  For example, you can say ‘Turn off the upstairs lights’ or ‘Turn the upstairs lights off’ – either way works.  If you’re polite, you can include the word ‘please’ anywhere and as often as you like.  If you’re brusque, you can just say ‘Alexa upstairs lights off’.

But sometimes I find I get stuck, or just feel an inhibiting lack of confidence that something will work, which is silly – why not just experiment and try.  And sometimes, it tells me ‘Hmmm, I can’t find the answer to the question I just heard’ or ‘I wasn’t able to understand the question I just heard’.  But what does that mean?  Did it hear my question correctly, but can’t answer it – for example, my most recent question ‘when is the inauguration’?  Or did it mishear me and I should ask the question again?

Privacy Concerns

The device is actively on and listening all the time.  Amazon explains that it only activates and starts recording when it hears the ‘wake’ word (‘Alexa’), and when it does turn on, you know it is recording and transmitting what it hears because the lights illuminate around the rim of the device.  There’s a mute button on the device if you want to be absolutely sure that the device isn’t surreptitiously listening in on you at other times.

So you’ve nothing to worry about when it comes to privacy, right?

Well, if you believe that, I guess you’re also the proud owner of the Brooklyn Bridge.  There are already articles on the internet about how to hack the Echo devices – I don’t know if the hacking to date has extended to defeating the privacy issues, but it is trivially simple to do so via software if access can be obtained.  The thing is that it is only a software command that activates or deactivates the microphone.  We already know it is always listening – how else does it know when to wake up and respond.  And the further thing to appreciate is that the mute button is not a mechanical disconnect that breaks open the connection between the microphone and the rest of the unit.  It is just a software command that says ‘okay, the microphone is still there, but ignore it for now’ and turn on the red mute light’.  It is totally easy to change the command and instead have it tell the unit ‘when the mute button is pushed, record everything but don’t respond to any commands, and turn the red light on’.  We generally have to trust that it is doing what it says it is doing.  If trust doesn’t come easily to you, some type of network monitoring that can report on the traffic per connected device on your network would be needed to ensure that the Echo isn’t ‘talking’ to any off-site connections when it ostensibly should not be.

So the mute button doesn’t really mute at all, and the device is listening or potentially listening all the time.  We have to trust Amazon not to exploit that, and we have to trust all the world’s three letter agencies not to come up with exploits so they can covertly listen in, if/when they wish.  If the local police can readily get a warrant to tap your phone, and/or to compel Google to release your browsing history, how hard would it be for them to get a warrant to compel Amazon to activate the microphone covertly without any indication appearing on the device?

Isn’t this a bit like the modern day admonition that you should cover over the web cam on your computer when you’re not using it?  (And, for that matter, why does no-one ever say ‘and also block up the built in microphone’ – it is as easy or easier to access and remotely monitor your computer’s microphone than it is your computer’s webcam).

The only way to be sure the device is off is to unplug it.  Which sort of destroys its entire purpose, doesn’t it.  Bottom line – you are potentially giving up some privacy in return for convenience – a trade-off that we are being increasingly and uncomfortably encouraged to accept in more and more parts of our life.

As for the sound bites that the unit officially sends to Amazon, those are stored for an unspecified period of time.

How Does Amazon Profit from Echo Sales?

There’s an interesting issue to ponder within all of this.  Clearly, the units have a reasonable amount of profit in them – we guess that a Dot costs Amazon under $20, and the full size Echo probably no more than $30, and the units are sold for $50 and $180 respectively.

But that is just the tip of the iceberg.  Someone has to pay the costs of its ongoing operation.  The Echo is not a self-contained device.  Most of its ‘intelligence’ is located in the cloud, somewhere in an Amazon computer farm.  Every time you ask it a question, your device simply sends some kB of voice recording to an Amazon computer that has to translate that into a computer-understandable statement, then has to process the statement, and send back a response.  If you ask it to stream music, it is then streaming you music as well.

While the computer cost for each transaction is some teensy tiny fraction of a penny, all those fractional costs start to add up over the life of a device, and even more when there is an installed base of over 5 million (currently) and with the number growing by several more millions after the Thanksgiving shopping season.  What is in it for Amazon?  Why is it providing a lifetime of unlimited computer resource for free?

Perhaps Amazon  is getting a revenue share from companies that partner with Amazon to provide Echo ‘services’ – and lets understand that when you order, for example, a Domino’s pizza, while you might see that as a service, Domino’s sees it as a wonderful profit opportunity – another sale without having the bother and cost of providing a phone operator to take your order.  So we’d guess that Amazon picks up some ‘slice’ (ha ha) of the cost of your pizza, and perhaps the same if you order a Lyft car, and so on.

We don’t object to that at all.  It is a bit like a travel agent.  If the cost is the same and the end-user convenience is improved, why not add middle-men galore to any transaction.

We also expect to see Amazon more aggressively start allowing you to order its own products through the Echo units.  Sure, you can already order some things, but think about the last time you ordered anything from Amazon.  Don’t you want the whole ‘shopping experience’ – starting with seeing all the products that more or less match what you searched for, then reading about them, choosing the size/color/configuration you most want, maybe reading a review or two, seeing when it would be delivered and how much for faster delivery, and so on.  None of that is easy or convenient without a screen.

Support

I’ve had occasion to call Echo Customer Support way too often – four times so far (in just over 24 hours of using the device).  The calls have gone through quickly with little or no time on hold.  All times seemed to have Filipinos answer, with the first experience excellent in all respects.

The second experience was reminiscent of the sort of comedy skit about call center operators that you might see on Saturday Night Live, complete with robotic intonation, no knowledge of American conventions (how to give out a phone number, for example), and ending up with the brilliant advice that to get better support, what I should do is simply hang up and call back the very same number.  I pointed out ‘But, that might mean I end up speaking to you again and neither of us want that’ which caused for a lengthy pause before an alternate number was offered.

The third call was close to as bad as the second.  I asked ‘Why does the Alexa app show my Fire 7 and Fire 8 as Echo devices, too?  How can I use them?’  The answer was that I should simply ignore that, and it meant nothing.  This surprised me, and I shared as much – ‘So you are telling me that your app has a feature on it that means nothing and I should just ignore it?’.  After a pause, he offered to ‘troubleshoot my iPhone for me’ which involved a Keystone Kops type charade of remotely accessing my computer and viewing the directories on its C drive (yes, nothing to do with the iPhone at all).

I told him ‘I have no problem, I just don’t understand what this option means or does’ and also pointed out that my iPhone was not connected to my computer so nothing he was doing had anything to do with the iPhone.  This caused him to put me on hold indefinitely (I hung up after 28 minutes).  This is a classic ‘passive aggressive’ move of call center agents who choose to antagonize and ‘punish’ callers for asking questions, rather than agreeing to help as they are paid to do, and they feel safe in their anonymity at doing so.  But he forgot I knew who he was because he was identified in the remote control session.  So, Amazon, may I suggest some further training is required of your support agent, who, while using the American sounding name Steven Woodard, is almost certainly known under a totally different name by his friends and family.

The fourth time, if possible, was the worst of the lot.  The call was quickly answered, but all I heard was the background sounds of people talking, etc.  No-one answered me.

In the good old days, Amazon offered brilliant customer support with American support agents who were insanely knowledgeable and helpful.  Not so much, now.  Perhaps this is another type of job loss that our incoming president could look into and do something about!  🙂

On a similar note, Amazon extends an insulting short 90 day warranty on the devices.  It offers to sell you a full year, or two, or three of warranty as an extra cost item.  It is nothing other than corporate meanness to not include a one year warranty on all electronics.

Limitations of the Units

There are various ‘rough edges’ around the periphery of the products and their capabilities.  The companion/control app, at least on an iPhone, is very ugly and poorly laid out, and sometimes disconnects from the Echo device it is associated with, requiring a phone reboot to reconnect.  The same if you are as brave/foolhardy as to attempt connecting the Echo Dot to a Bluetooth speaker system; several times my Dot disconnected from the Bluetooth connection and I just gave up and ran a wire instead.

The biggest disappointment though is that while Amazon is encouraging us all to buy the units in lots of six and twelve at a time, they have not thought through the implications of multiple devices in one environment.  Say, for example, you have an Echo unit in the living room connected to your home stereo system, and another in the kitchen.  You’re in the kitchen and want to get some music playing on the stereo system.  That’s a foreseeable request.  But there’s no way you can ask the unit in the kitchen to talk to the unit in the living room and get it to send music to your stereo system.  Whichever unit responds to your utterance is the only one you can command/control.  That is ridiculously primitive and a totally unnecessary limitation.

Another example – the Echo would make a great ‘clock/radio’; it is so easy to set an alarm, unlike many of my other alarm clock/radio type units.  But, while the unit can stream radio channels, and can work as an alarm, you can’t get it, when sounding an alarm, to play your choice of radio station.  You have to accept its spacy alarm sound.  Why can’t this be changed?  There’s no reason, other than incomplete poorly thought through programming.

It has a ‘to do list’ feature, but you only have one list which everything would go on.  I’d like to have one list for home, one for work, one for a specific project, and so on.  Not possible.  How hard would it be to be able to create and manage multiple lists?  How about setting priorities for each item on the list?  Also not currently possible.

Another feature conspicuous by its absence.  Sure, it can stream audio, but how about some type of an interface to a video monitor so it can also stream video?  Or just an easy way to control one’s Amazon Video streaming experience?

To try and put a positive spin on these surprising limitations, the good news is that they are all software limitations, not hardware restrictions.  That means there is reason to hope that in the next, or the next after the next, new software release, some of these obvious features might start to appear.

But for a device that has been in development for six years and on the market already for two, these are astonishing omissions in their feature set.  For something that seems to be one of the key cornerstones of Amazon’s push into an even more omnipresent role in our lives, it is very surprising they’ve not allocated more resource to simple features, and to making their iOS app more polished and user friendly.

Summary

The Echo is an exercise in seeing your glass as half full, rather than half empty.  It is a wonderful futuristic concept that isn’t quite completely ready for prime time, just yet.  But in its present somewhat limited form, it already has appeal and value, particularly if you’re thinking about simply getting one of the inexpensive Echo Dot units ($50 each), and because the limitations may be resolved in the form of simple software updates in the future, there’s no reason not to get one now and then enjoy its ongoing enhancements.

Whether you use it as a glorified music player, an alarm clock, a source for weather forecasts and other informational requests, to remote control your smarthome gadgets, or for one of the other 4,000 potential uses that currently exist for it, you might well find more than $50 worth of value and fun in this $50 unit.

Recommended for early adopters and gadget lovers.  Ideal as a stocking stuffer for the people in your life who have everything (except an Echo) this coming Christmas.

Note that Amazon currently have several discounted offers on the Dot.  If you use the discount code DOT3PACK in your shopping cart, you’ll get $20 off the cost of three (ie $130 for three rather than $150) and if you order six, use the DOT6PACK code and get $50 off (ie pay for five and get six).