The voice interface has a problem: discoverability

By Adam Tinworth

10/02/2017 | The voice interface has huge potential to be one of the next big ways with which we communicate with computers. But there's always a problem with new interfaces in the early stages - people try to map the paradigm of the past onto the new. And it doesn't always work - look at the attention to put an iPhone-like app screen on the Apple Watch. It really doesn't work.

The same seems to be happening with apps on Amazon's Alexa platform - known as "skills". Recode summarised the findings of a report by VoiceLabs, which doesn't paint a pretty picture of the existing voice app ecosystem:

For starters, 69 percent of the 7,000-plus Alexa “Skills” — voice apps, if you will — have zero or one customer review, signalling low usage.

OK - but given that there's less of an App Store interface, could that not be as bad as it seems?

What’s more, when developers for Alexa and its competitor, Google Assistant, do get someone to enable a voice app, there’s only a 3 percent chance, on average, that the person will be an active user by week 2, according to the report

Now that's concerning. The comparable figures are 13% and 11% for Android and iOS, respectively.

From my own experience with the Amazon Echo, I use the functions of the device that don't require me to specifically invoke a particular Skill, but I rarely remember which Skills I've enabled, so really can't invoke them by name. I suspect the future will see the voice platforms opening up hooks for the sorts of tasks that people will want to do - and let the user decide which skill fulfils that task. A request for a taxi could pass to Uber or Lyft, and so on.

New interface. New interaction models. They'll come.