The three systems, not one
People say "the Spotify algorithm" like it's one thing. It's three overlapping systems, each with different inputs and time windows. Confusing them is why most artist advice is wrong.
- Editorial — humans curating Spotify's flagship playlists (RapCaviar, mint, etc.)
- Algorithmic — Discover Weekly, Release Radar, Radio, Daily Mix
- Personalized — your On Repeat, Repeat Rewind, Your Top Songs
You optimize for each one differently. Confuse the systems and you waste effort.
1 · Editorial playlists
Editorial is humans. About 100 people in NYC, London, Stockholm, São Paulo, and a handful of other regional hubs decide what goes into the flagship playlists. You can't game it; you can only pitch correctly.
How to pitch correctly
- Pitch through Spotify for Artists at least 7 days before release. Track must be confirmed in your distributor delivery
- One pitch per release. Your strongest single, not all five tracks
- Fill every field: genre, mood, instrumentation, language, region. The metadata routes your pitch to the right editor
- The "story" field is what the editor reads. Lead with what's interesting (collab, sample, regional context, sync, milestone)
Submitting a pitch through Spotify for Artists unlocks Release Radar exposure to your existing followers. That alone is worth doing — it's the algorithmic boost most independent artists ignore.
2 · Algorithmic playlists
This is where the actual machine learning lives. Three relevant ones:
Release Radar
Personalized weekly playlist of new music from artists each user follows or has streamed recently. Updates Fridays.
How to enter: Have followers, have recent listeners, and submit your release through Spotify for Artists. That's it. Release Radar is the most underrated algorithmic playlist for independents — guaranteed exposure to your warm audience.
Discover Weekly
Personalized weekly playlist of music users haven't heard but should like. Updates Mondays.
How to enter: Get listeners with overlapping taste profiles to save and repeat-play your track. The algorithm uses collaborative filtering — if listeners who like X save your track, your track gets recommended to other listeners who like X. Saves and repeat plays matter more than raw streams.
Radio (artist & track radio)
Continuously updated stream that builds on a seed track or artist. Most underrated source of streams for niche genres.
How to enter: Audio similarity. Your track needs Spotify's audio feature analysis (BPM, key, energy, danceability, valence) to align with established tracks. Genre tagging in metadata helps the seed-matching.
The signals that actually matter
Spotify hasn't published exact weights, but credible analysis from former Spotify staff and consistent observation by curators gives this rough ranking:
| Signal | Weight | Why |
|---|---|---|
| Save rate | Very high | Strongest signal that a listener wants to come back. Rate matters more than absolute count. |
| Full-listen rate | Very high | Streams under 30 seconds don't count for royalties OR algorithm weight. The 30-second filter is hard. |
| Repeat plays | High | Same listener returning to the track within days signals genuine taste fit, not curiosity. |
| Playlist adds (user playlists) | High | User-created playlist adds are stronger than editorial — it's "voluntary" listening. |
| Skip rate | High (negative) | High skip rate before 30 seconds tanks algorithmic distribution. Worse than no plays. |
| Source diversity | Medium | Plays from many countries, devices, and contexts beats concentration. Bot streams concentrate. |
| Follower count | Low (direct) | Mostly affects Release Radar reach, not Discover Weekly or Radio. |
| Total stream count | Low | Counter-intuitive but true. A track with 10K streams and 8% save rate beats 100K with 0.3% save rate. |
This is why bot streams kill careers. They concentrate plays, lower save rate, lower full-listen rate, and signal "fake fan base" to the algorithm. Spotify's anti-fraud detection got noticeably more aggressive in 2024–2025.
The 28-day window
Algorithmic playlists use rolling 28-day windows for most engagement signals. The first 28 days post-release are the highest leverage you'll ever have on a track.
- Day 0–7: Release Radar pushes to followers. Editorial decision (yes/no)
- Day 7–14: Algorithm starts evaluating engagement metrics — if save rate is good, Radio inclusion begins
- Day 14–28: Discover Weekly evaluation kicks in. If signals are strong, you enter the recommendation loop
- Day 28+: Without sustained signal, algorithmic weight decays. Tracks rarely "come back" without a new push (sync, viral moment, big playlist)
This is why concentrated promotion in the first month beats spread-thin promotion over six months.
Where independent curators fit in
Curator playlists like the ones Rapture runs aren't algorithmic. They're human-curated user playlists. But adds from genuine niche curators feed the algorithm in three ways:
- Save rate boost — users who follow niche curators tend to save tracks at 5–10× the average rate
- Repeat-play boost — niche playlists get repeat-listened by genuine fans, not background-noise listeners
- Source diversity — distributed plays across the curator's audience signal organic discovery
This is why a 5,000-follower niche playlist often delivers more algorithmic lift than a 500,000-follower vanity playlist with low engagement. Save rate and repeat plays beat raw audience size.
What doesn't matter (much)
- Bot followers — Spotify detects them and discounts the signal
- Cross-platform "engagement" — TikTok virality only matters if it drives Spotify saves/full-plays. Views ≠ algorithm
- Submission to "playlist farms" — bot-driven playlists actively damage your algorithmic profile
- Paying for Spotify ads — useful for marketing but barely moves algorithmic playlists. Save your budget for distributor + curator submissions
- Track length — longer doesn't mean less play. As long as the hook lands fast, length is neutral
The 72-hour launch checklist
If your release is 72 hours away:
- Confirm Spotify for Artists pitch is submitted (must be 7+ days before, but check it's actually queued)
- Pre-save link active and in every social bio
- Cover artwork loaded into Spotify (not just distributor)
- Artist Pick set on your Spotify for Artists profile to point at the new track
- Email list teed up with the pre-save link, sending day-of-release
- 2–3 niche curator submissions queued (genre fit, not volume)
- Day-1 social posts ready: short clip with hook, save link, story