How well does Spotify really know you?

Carol Liu
December 15, 2021
5 min read

TLDR: Spotify knows a lot about you, based on your music-listening activity, but there’s one essential thing they don’t know - how music affects your mood.

If you’ve been tuned into the Internet at all, you’ve probably seen the deluge of Spotify Wrapped posts showing everyone from avid Drake listeners, to too-cool-for-you bedroom indie hipsters. "My Audio Aura" was a fun new feature in this year's Spotify Wrapped which gave you a coloured gradient representing the mood of your music.

One thing’s for sure - Spotify has a keen sense for what genre we prefer to listen to. And they’re really good at recommending us more of the same.

What most people don’t know, is that Spotify also scores the happiness, energy, danceability and acousticness of the songs you listen to. You can check out your own scores on the Obscurify website. Here's how Obscurify rated my Spotify library:

Image: Obscurify


Image: Obscurify


According to Spotify, my music taste leans more towards a low happiness score, and it’s been getting even sadder/angrier. The first part didn’t surprise me, since I’ve always known that I personally prefer melancholic or emotional music, but I was surprised to learn that my current taste scores even lower on the happiness scale than my average all time taste!

In the past couple of months, I’ve actually made a conscious effort to listen to more ‘happy’ music in an attempt to boost my mood during the dreary seasonal-depression, gets-dark-at-four-PM period. This statistic gave me a hard slap in the face, and led me to wonder how my music-listening habits are contributing to my mental health.

The question is - if Spotify really knows us that well, are they just giving us more of what we want? Or are they able to serve us the music that we need?


The question is - if Spotify really knows us that well, are they just giving us more of what we want? Or are they able to serve us the music that we need?

In the past, we often categorized music by genre, language, or geography. Mood and purpose are now more important to listeners than genre when it comes to music classification. This has led to the rise of functional music, where more users are choosing playlists that are context-based instead of content-based. This can be seen in Spotify’s top 20 playlists worldwide - titled “Get Turnt”, “Peaceful Piano”, or “Mood Booster”. On YouTube, the channel Lofi Girl who livestreams “lofi hip hop radio - beats to relax/study to” has over 9.6 million subscribers. Listeners are familiar with using music for a functional purpose, to elicit a specific emotion. But how well do these playlists work?

When it comes to generic functional playlists, there’s a lack of rigor and objective process behind the curation process. The songs in the “Mood Booster” playlist are not necessarily composed with the intent of helping someone who’s in a funk, get out of it. So who decides that this particular song is a mood booster? And will it provide a mood boost for anyone, regardless of their personal music preferences? Of course, there’s also context - how the listener experiences different times and moods. For somebody in a melancholic mood, ‘Dancing Queen’ by ABBA may not cheer them up the same way as it may if they’re in a mellowed state. Often, these generic playlists resemble a one-size-fits-all solution that doesn’t take into consideration the variability in personal taste and context.

Spotify can’t curate according to your mood because it’s missing one key piece of information - it doesn’t understand your emotional response to music. Generic playlists don’t know if you’re actually feeling better, worse, happier, or angrier, or generally if you’re anywhere close to their promised emotional outcome after listening.

Spotify can’t curate according to your mood because it’s missing one key piece of information - it doesn’t understand your emotional response to music. Generic playlists don’t know if you’re actually feeling better, worse, happier, or angrier, or generally if you’re anywhere close to their promised emotional outcome after listening.

On the other hand, an affective music AI can personalize to your music preferences as well as your current mental, emotional and physiological state. Using intelligent, data-driven tools throughout the composition and curation process can provide an evidence-based approach to functional music. This approach would create more personalized functional music as opposed to the static one-size-fits-all playlists.

So how would an affective music AI compare to popular generic playlists? We decided to run a study to find out.

In this study, participants listened to either Spotify’s “Calm Vibes” playlist - the top relaxation playlist with over 750 million followers - or a personalized music experience curated by LUCID’s affective AI.

LUCID’s affective AI curates digital music therapy with integrated ABS (auditory beats stimulation), with the goal of decreasing anxiety and increasing valence. It is an interactive music experience that collects the user’s current mood state and anxiety levels through an emotional assessment grid, before and after their experience.


The results - LUCID’s music intervention reduced stress by 20.5%, while Spotify's playlist reduced stress by 12.3%. Valence, or positivity,  was boosted 11.6% by LUCID compared to 3.6% by Spotify.

The results - LUCID’s music intervention reduced stress by 20.5%, while Spotify's playlist reduced stress by 12.3%. Valence, or positivity,  was boosted 11.6% by LUCID compared to 3.6% by Spotify. Activation (energy levels) was also measured, wherein a lower activation score indicating a more relaxed state. LUCID decreased activation by 12.6% while Spotify increased activation by 2.4%



According to these data points, listeners of LUCID’s music intervention were more relaxed, less stressed and in a more positive mood as opposed to listeners of the “Calm Vibes” playlist. These findings are consistent with the results of our Adult Anxiety Study, in which participants with moderate trait anxiety experienced a significant reduction in anxiety after a single session.

The future of music will gravitate towards data-driven curation and hyper-personalization for functional outcomes. Music’s role in health & wellness can be unlocked through these methods, generating positive outcomes in a variety of environments from digital health to clinical settings. We are committed to conducting further research and development to turn music into medicine.

Read the full study and comparison data here.

Image credits: Spotify & Obscurify