Podcast Episode 34 – A good mix! How streaming service Deezer bets on human-curated playlists as well as algorithms15. April 2020
With more than 14 million active users in 180 countries, finding out what a local market wants is a very special challenge. In the newest episode of our podcast, Benoit Terpereau, VP Product at music streaming service Deezer explains why product managers need to be regional experts and how curators put together human-curated playlists to supplement the algorithm.
Digitale Leute Insights is the podcast for passionate product people. We interview product developers from around the world and take a closer look at their tools and tactics.
Host: Tilo Bonow
- In this episode you will learn:
- How the product teams are structured at Deezer,
- why Deezer takes every effort to provide the users with a good mix of human-curated playlists and machine learning algorithms,
- that design plays an important part when facing heavy competition,
- why bringing your passion to work is a good start for young product managers.
“As a trained software engineer, I’m not used to talking to people,” Benoit says jokingly about himself. He refers to the first lean startup event he attended. His idea: A travel app, which turned out not to be what his potential customers wanted. They wanted a book rather than an app. Fast forward ten years, as VP Product of Deezer, Benoit is now consolidating feedback from customers from several communities, from Twitter, app reviews, Reddit, customer support and surveys to find out what to develop next.
Benoit uses a threefold rule for the decision-making process itself: Check the data, listen to user insights and use your feelings and expertise. That way they released the often requested dark mode end of 2019.
A joyful UI and human-curated playlists
In a market where Apple, Spotify, and Amazon are the top dogs, creating a USP is a challenge. One way Deezer wants to stand out from the competition is by creating a joyful experience within the app with more colors than all competitors in the UI. For example, Deezer uses changing backgrounds in a playlist depending on the title playing.
But the look of an app is not worth much if you don’t like what you are listening to. This is why Deezer decided to not only build a machine learning generated algorithm. With local curators putting together dedicated playlists for each country, they add a local flavor to the algorithm.
I think we are the best streaming service creating a balance between algorithm and curated playlists.
LinkedIn: Benoit Terpereau
About the Host
Tilo Bonow is the founder and CEO of PIABO, Germany’s leading PR partner for the digital economy headquartered in Berlin. As a long-time entrepreneur, communications expert and investor, he actively supports his clients in achieving their local and global growth targets, including tech giants such as Stripe, GitHub, Merck, WeWork, and Silicon Valley Bank.