Q4 may not be quite as intense a period for many people in the music industry as it used to be in the physical era, given the fact that an album released early in the year (or in Ed Sheeran’s case, early in the previous year) can still be streaming up a storm in December. But when it comes to smart speakers like Amazon’s Echo and Google Home, Q4 absolutely is the bumper sales quarter, as people rush to get their first (or second, third, fourth…) talking speaker for Christmas.
Focus on the ‘first’ there – millions of people are going to be interacting with a smart speaker for the first time this Christmas, so improvements to the capabilities of those speakers’ voice assistants (especially for music) are notable. Which brings us to Amazon Alexa’s latest new features, which add to its music-discovery capabilities in an interesting way.
It’s partly about a proper conversation: rather than making a single request (‘Alexa, play Drake’ or ‘Alexa, play Christmas songs’ etc) people will now be able to start a request with ‘Alexa, help me find…’. For example, ‘help me find dinner music’ or ‘help me find a holiday playlist’. Alexa will then ask some more questions to fine-tune whatever it eventually plays in response.
Amazon also says that there’ll soon be an ‘Alexa, recommend some new music’ command as well as ‘Alexa, what should I play?’ – in both cases, Alexa will also follow up with more questions, using the answers and its existing data on your listening habits to refine its recommendations. Meanwhile, new commands including ‘Alexa, I like this song’ and ‘Alexa, I don’t like this’ will also help people to tune Alexa’s understanding of their tastes – complete with a follow-up ‘Alexa, play music I like’ command to focus playback on songs that have this seal of approval.
This is all fun for owners of Echos and other Alexa-powered devices, as is the nudge towards a generic ‘Alexa, play music’ command that will use the new conversational features, as well as that archive of listening data. It’s a concept that’s been talked about and explored a lot in recent years, from Spotify’s Paul Lamere dreaming of “a music player where you hit the play button” back in 2014 to Deezer’s early launch of its ‘Flow’ feature the same year.
In the smart-speaker context, the thought of people’s commands getting simpler and more general, requiring the recommendation algorithms to be even smarter, is an important trend. Not least if you’re one of the growing number of label staff tasked with trying to better understand (if not quite reverse-engineer) these algorithms, to understand how to give your music the best chance of being chosen by Alexa (or Google Assistant, or Siri…) as they get more conversational with their users – and smarter about interpreting their listening habits.