Alex Underwood, VP and head of global strategic partnerships & verticals at Spotify, addresses this topic in the November issue of Admap, which focuses on personalisation and features nine papers from thought leaders in the UK, US and Asia.
He stresses that data should serve your audience, not yourself. In one sense it could hardly be otherwise at Spotify, as it deploys machine learning to drive effective personalisation in order to help its audiences navigate more than 40 million songs.
And the evidence supports the thesis: its streaming data shows that users of Discover Weekly – which delivers two hours of custom made music recommendations powered by algorithms that combine personal listening data with data from users that have similar listening taste profiles – stream more than twice as long (2.1x) as non-Discovery Weekly users.
“Personalisation should be a trigger to inspire positive emotional resonance and reaction,” Underwood says, and any personalised content or communication needs to be “authentic to your brand and true to the category in which your brand exists”.
From a brand perspective, that means getting permission in order to get personal – “otherwise, you risk leaving your audience feeling manipulated by your brand”.
Timing is another factor: Spotify updates Discover Weekly on Monday mornings, a time when people are likely less inspired or motivated to select their own music; marketers need to identify the most appropriate contexts for their brands to get personal.
But simply demonstrating that a brand knows and understands an individual is not enough, says Underwood. “You also need to create a relationship with your audience to become a part of their lives.”
And that relationship needs to evolve, something Spotify can do on a number of levels – compiling a list of nostalgic tracks from peoples’ teens and early twenties, for example.
While algorithms are a crucial part of Spotify’s approach to personalisation, it still has a human touch as machine-driven choices are handpicked and filtered by a team of human curators to produce what it terms “algotorial” recommendations. Click below to hear more about this in an exclusive Q&A with the author.
These owned and operated playlists, such as RapCaviar, Hot Country and Viva Latino, have evolved into multimedia experiences and become “epicenters of culture in their own right”, Underwood reports. His takeaway? Machine and human are a winning combination.
WARC subscribers can access the nine articles as well as a deck summarising the key thinking and advice from all authors.
Sourced from Admap