Artificial Intelligence

Why voice-first tech isn't resonating with developers

The voice-first revolution keeps promising to change everything...if only developers were paying attention.

Sometimes a pipe is just a pipe, as Freud once quipped to a student. And sometimes a voice-activated assistant is just a voice-activated assistant, and not the keys to a grand, voice-driven shopping universe. As a deep-dive report by The Information uncovers, for the 50 million Amazon Echos (Alexa) sold, a mere 2% have been used to actually buy something (90% of which never bothered to do so again). Indeed, the vast majority (77%) use Alexa-powered devices to listen to music—something that Apple's Home Pod is arguably better at.

This leaves Alexa in the role of being a revenue generator for Amazon, but not necessarily a strategic pillar for its going-forward strategy. Is that a bad thing?

The old Alexa curiosity shop

Despite the much ballyhooed potential of our voice-first revolution to change the world, voice is still a distant last place to, well, everything else when it comes to interacting with our machines. Yes, my family routinely badgers Alexa to start or stop cooking timers, and I did once ask Alexa to play an excerpt from A Prayer for Owen Meany so that my wife could hear the protagonist's voice, but the reality of voice-first is that it's still voice-last.

SEE: IT leader's guide to the future of artificial intelligence (Tech Pro Research)

When we do turn to voice-activated speakers like the Amazon Echo (Alexa), the interaction is light on engagement and heavy on playback, as a Reuters survey surfaced by Benedict Evans has highlighted:

voice.jpg
Image: Reuters

While consumers seem disinclined to use the Echo for what Amazon must surely wish they would (shopping), a scant 2% do that, as mentioned. In a statement to The Information, an Amazon spokesperson declared: "Millions of customers use Alexa to shop because it is the most convenient way to capture needs in the moment," but this seems...untrue. Far more (20%) have used Alexa to ask about orders they made on another device, sources told The Information.

The problem with shopping with just voice is that shopping is primarily a visual thing. You want to see (if not the physical product at least a picture of it, as well as reviews/etc. around it) before you buy.

Not that this has stopped analysts from projecting the huge revenue potential of voice. As noted in The Information's report, OC&C Strategy Consultants sees voice-driven shopping to jump from today's $2 billion to $40 billion by 2022. RBC Capital Markets, meanwhile, says that Alexa could be good for as much as $6 billion by 2020.

Possibly.

About those developers...

If you believe that developers are the new kingmakers (and I do), then the next question is "what do developers think about voice first?" This isn't an easy question to answer.

For example, if we look at the total number of Alexa skills created by developers, the number is 45,000 and climbing. Just one year ago that number was 15,000. However, based on both the Reuters data and earlier studies by VoiceLabs and others, those Alexa skills have a 97% fall-off rate. That is, only 3% of those who use a skill will go back to it a week later.

SEE: How we learned to talk to computers, and how they learned to answer back (cover story PDF) (TechRepublic)

More worrying still, the vast majority of those 45,000 skills will never be discovered in the first place. Aside from the stock skills (like "Play music") that come with a voice-activated assistant, most are like the tree falling in the forest.

No one but the originating developer knows that they're there.

As such, I suspect most of those skills being developed for Alexa (and far fewer for other platforms) are tire-kicking exercises by developers. So much hype has been blathered about the potential of voice that of course developers will give it a try. But when the return on that investment falls short—and it is falling short, across the board—we'll see developer investments in voice taper off.

In sum, developers have kicked the tires on Alexa but basically none are making money on the platform. Users, meanwhile, basically treat Alexa like a newfangled radio. In short, quoting Evans, "It is at least possible that Alexa will have no more strategic impact on Amazon than the Kindle Fire." (Hint: this means "none at all.")

Will Amazon still print lots of cash on Echo sales? Yep. Will it be a platform for selling gobs of goods? Probably not. This isn't a bad thing, it's just a reality check on the voice-first hype.

Also see

echo.jpg
Image: Ian Knighton/CNET

About Matt Asay

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox