Big Mood Machine

Liz Pelly

Music and emotion

  • Our listening habits often signal something deeply personal and private.
    • Think about why you choose certain songs at certain times.
  • Music streaming platforms are in a unique position within the greater platform economy.

Spotify’s Emotional Goldmine

  • Streaming platforms, especially Spotify, possess troves of data related to our emotional states, moods, and feelings.
  • But it’s not just sitting there…
  • Spotify has been selling access to that listening data to multinational corporations.

Making Moods Explicit

  • Other platforms might need complex algorithms to infer emotional states.
  • Spotify streamlines the process: Users literally click on boxes indicating their moods.
    • Happy Hits
    • Mood Booster
    • Rage Beats
    • Life Sucks
    • Wake Up / Chill / Focus playlists

Playlist Generation/Refinement

  • How do they choose the right songs for “Sad Indie” vs “Life Sucks”?
    • AI analyzes audio features (tempo, energy, valence) and textual data (lyrics, genre tags, user descriptions) to categorize songs by mood.
    • Recommendation algorithms (like collaborative filtering) suggest these mood playlists to relevant users based on their listening history and inferred context.

Discussion

  • How do you typically choose music on Spotify (or other streaming services)?
  • Do you often use mood-based playlists?
  • Do you select music to match your current mood?
  • Or do you use music to try and change or induce a mood you desire?
  • Or do you mostly just stick to artists/albums you know?

Spotify’s Narrative: Curation & Discovery

  • Over the years, Spotify aggressively marketed mood/activity playlists.
  • The story: This helps users navigate the “infinite choice” (40+ million songs).
  • Positioned as a benefit for both listeners (finding music) and artists (getting heard).

Discussion

  • Do you think mood playlists are necessary to find music you want to listen to in the streaming era? Why or why not?
  • How is picking a “Chill Hits” playlist fundamentally different from the past, when people would turn on a radio station known for playing, say, smooth jazz or soft rock (genres often associated with certain moods)?

Advertising Revenue

  • A closer look shows the push for moods was strategic, particularly leading up to Spotify’s IPO (Initial Public Offering).
  • The goal: Grow Spotify’s advertising business.
  • Spotify’s enormous access to mood-based data is a pillar of its value to brands and advertisers.

Targeting by Emotion

  • This mood data allows advertisers to target ads on Spotify based on listeners’ likely moods and emotions.
  • Imagine ads for comfort food targeting “Life Sucks” listeners, or energy drinks targeting “Workout Beats” listeners.

AI-Powered Ad Targeting

  • Real-time Bidding (RTB) and programmatic advertising platforms heavily rely on AI.
    • AI algorithms instantly analyze a user’s profile (current listening mood, history, demographics) the moment an ad slot becomes available.
    • They match this profile against bids from advertisers targeting specific emotional states or contexts.
    • AI predicts the likelihood a user will engage with an ad, optimizing placement for effectiveness and revenue.

Selling Data

  • It’s not just used by Spotify for its own ad platform.
  • Since at least 2016, Spotify has shared this mood data directly with the world’s biggest marketing and advertising firms (e.g., WPP, Omnicom, Publicis).

AI for Data Segmentation & Packaging

  • AI algorithms perform sophisticated user segmentation, grouping users based on combinations of listening habits, inferred moods, explicit playlist choices, demographics, and other factors.
  • Machine learning helps identify high-value audience segments (e.g., “likely stressed commuters,” “weekend party planners”) that advertisers want to reach.

A Shift in Identity?

  • Liz Pelly offers a counterpoint to the “music discovery” narrative:
  • Spotify’s pivot to moods was less about music discovery and more about gathering granular data on listening habits and emotional states.
  • This transforms Spotify into more of a marketing and data firm that uses music as its medium.

Conclusion

  • Music is deeply emotional, and streaming platforms capture data reflecting this.
  • Spotify explicitly categorizes users by mood via playlists
  • While presented as a discovery tool, this strategy fuels Spotify’s advertising business
  • This mood data is processed and segmented using AI before being shared directly with major advertising conglomerates.
  • This raises questions about privacy and whether we see Spotify primarily as a music service or a data broker