How Spotify Uses Personalization (to Connect Artists and Listeners to Music)

A few days ago I was at a Spotify event in Boston, titled “Music and Data”. Featuring three folks from the Spotify offices in Boston, the session outlined how Spotify uses data to  customize and personalize the artist and listener experience in Spotify. There were three separate presentations, and all three were informative both from the artist side and listener angle. The highlight of the session was learning that Spotify has a superfan program, which I will get into detail, so keep reading!


Presentation 1: How Spotify Connects Users to Music

Back in the day, when record stores ran the music consumption experience, we did not have the personalization features at the level we have today. We mainly had categories of genres and artist names in alphabetical order.

So what kind of data does Spotify look at? There are three types of data they look for: Explicit signals, implicit signals and contextual signals. An example of explicit signals are hearts or song likes. Implicit signals are song skip rates, adding to playlists etc. Contextual signals are things like what time of day is it, are you on a speaker / computer or mobile, what day is it, etc.?

The presentation went on to describe that Spotify is very focused on new users. For them, user acquisition is super important, but the challenge is that Spotify doesn’t know anything about them initially. So they have a system called “taste overboarding”. This is when you open the app the first time, the app asks you to select 3 or more artists you like. This helps users springboard their content on Spotify.

Spotify regularly uses machine learning as well.  If you type in “fr”to the search bar for instance, for one person Frank Ocean might show up, but for another Frank Sinatra. If you type in j, you might get “Juice World” based on your listening history, while someone else could get “Jessie J” or “J Balvin”  based on what they listen.

The presentation then went on to explain other ways Spotify uses personification. Overall, Spotify combines aggregate data with personal data. The ‘browse genre’ menu changes from person to person, as well based on what genre they listen. Spotify likes to personalize everything based on the user likes. They also have cultural experts who help Spotify select the kind of songs different kind of playlists: These are playlist editors. Lastly, they also have concert personalization where Spotify will email and tell you who’s playing in your city. They also will show this info to you in the app.


Presentation 2: Voice Systems and Omakase

The second presentation focused on voice systems such as Alexa, and the concept of Omakase.

Omakase is the term when patrons of restaurants in Japan will leave the food selection to the restaurant rather than ordering from menu. Spotify is like this that way, especially in terms of the voice systems. For instance, when you say “Play David Bowie”, or “Play Something Funky”, Spotify will do it for you. This sounds pretty straightforward, but Spotify wants to personalize and localize this experience as well. As a basic example, “Funk” in United States and “Funk” in Brazil are two completely different genres, so Spotify wants to be able to make this distinction.

They also want to personalize when users are less specific like “Play me some music” or “Play some Guacamole making music”. These are the challenges for Spotify, because apparently about half of Spotify listeners make an omakase, or vague searches like these,  every month. The reason Spotify thinks getting the omakase right is because ultimately this is better experience for both users and artists. Better omakase creates additional revenue for artists and better match for users.


Presentation 3: Fans First Project and Superfans

Fans First Project is Spotify’s project, where they help Spotify artists identify their SuperFans. With Fans First emails, Spotify reaches out to artists’ biggest Spotify fans with unique or exclusive offers. Some treats they offer include access to presale tickets for upcoming concerts, access to exclusive merch items not available anywhere else and Invites to special artist events (such as Spotify Concerts).

Of course the biggest question then is who gets these emails and how do you get one?

Spotify defines a superfan as a measure of how much a user is the fan of an artist. So for instance, how do they figure out who are the Top 50 fans of a particular artist? Apparently they determine a score for every user. This score is a combination of user listening data (skips, plays etc), song like/dislike data, save/playlist add data, track data and follower data. The score is combination of these five. Based on the top 50 users scores, they get an email!

All three presentations were great for understand the challenge for personalization for the music experience of Spotify. For me the third presentation was most interesting, as the idea of determining the ‘Top 50 Fans’ of an artist seems to open up new opportunities for all. One of the top questions is whether Spotify shares this “Top 50 Fans” data with the artists for instance (which I’m guessing is partially no, because Spotify Artists platforms shares song play, save and playlist add data, but not likes/dislikes or skips of a particular song or artist).

Adva Mobile is a marketing services and technology company for creative artists. Using Adva’s mobile services, you can let your fans learn about your latest creative work, run contests, take surveys, reach out to your superfans and engage with them.

Alper Tuzcu is a Berklee College of Music and Denison University alumni, and a Boston based guitarist, songwriter and producer. His new album ‘Aurora’ was released on October 19, 2018 and his debut eclectic album ‘Between 12 Waters’ featuring 8 different vocalists is available on Spotify. In addition to being a musician, he regularly teaches workshops and masterclasses internationally. You can follow him on Instagram or Twitter, and for more information you can visit his website



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s