Spotify has made its biggest move to date into the booming voice market by rolling out the “Hey Spotify” feature to a test base of users. Initially limited to its Android and iOS apps in the US, the streaming service says it will slowly expand out from there.

The company is, in a blog post about it all, at pains to stress this is all opt-in for users and that it will only spring into life when the words, “Hey Spotify…” are uttered by the user. It will also only be operational for the duration of their command.

Voice search is a naturally swifter way for most users than typing to get what they are looking for and that notion of delivering convenience to our hectic lives has been the key selling point for the multitude of speech-based services on the market.

Amazon launched Alexa, its voice-controlled virtual assistant, just over eight years ago. It was partly in reaction to Apple’s launch of Siri 18 months earlier to build voice integration into the iPhone 4. What was different about Alexa was that it pushed a whole new hardware category into the mainstream – that of the smart speaker. Amazon dominated in the opening years but has subsequently faced fierce competition from Google GOOG (with Nest) as well as from Chinese companies Baidu BIDU and Alibaba BABA not to mention, albeit to a lesser extent, Apple (with HomePod). Over 150 million smart speakers were sold last year globally according to Strategy Analytics research.


There is, however, a lot more than just amenity happening here. If this is a hardware arms (voice?) race, Spotify cannot afford to be a silent witness.

The logic underpinning voice assistants is that the more you remove barriers of all stripes the more users will use your service. This can be read within the context of Spotify’s move away from being a pure “music service” towards becoming a broader audio company – notably by acquiring podcast companies like Anchor and Gimlet as well as signing up huge names to make podcasts for it exclusively. Spotify is less concerned with what you are listening to than it is with how much you are listening to it on its platform.

Speaking to different record labels last year, they all said the same thing – that voice was having a huge impact on listening behaviour due to many people now having to work from home. Context was changing consumption.

Many reported that streaming of lullabies and children’s songs in particular were sharply increasing as young (pre-literacy) children were able to ask Alexa (or whatever device was in the family home) to play their favourite songs. Again and again.

“Hey Spotify,” could just becomes another verbal means for children to assert more autonomy over their listening at home. It all links to its Spotify Kids offering and how it is recruiting new users from the kindergarten upwards. The convenience of the verbal benefits Spotify’s growth as much as it benefits the user’s listening experience.

In announcing “Hey Spotify”, the company added a separate post covering off the privacy issues here, looking to assuage user reservations by insisting that it is only gathering voice data when issued with direct commands. However, it will (if switched on in the app) be running constantly in the background, keenly awaiting its next request before springing into life. “When listening for the wake-word, Spotify listens in short snippets of a few seconds which are deleted if the wake-word is not detected,” the company says.

This is the knife-edge that all voice-command technology is asking the user to walk along, pushing the convenience angle to try and deflect from any worries relating to privacy and device surveillance.

Spotify is no stranger to any of this and has run into these issues in its early days as well as its recent past.

It had to row back in 2011 when it partnered with Facebook and automatically shared what everyone was playing to their Facebook feeds.

Then at the start of this year it was revealed that Spotify had been granted a patent – one it had applied for in 2018 – that would allow it to analyse a user’s voice and from that serve up musical suggestions based on what it understood to be your “emotional state, gender, age, or accent”. It was a move that some felt was just a little too intrusive, a little too… dystopian.

“Hey Spotify” is arguably the beatific face of this wider push to normalise voice commands without drawing too much attention to what exactly is purring under the hood.

Setting the arguments about speed and convenience in music search to one side, the question users now have to resolve for themselves is this: when I listen to Spotify, just how much is Spotify listening to me?