The Spotify PlaylistCleanup, in our data

Music-promotion channels attract manipulation once enough money flows through them. Radio had payola; playlists developed bots and pay-for-play. Spotify intensified enforcement in 2025. But what happened in 2026?

Spotify's cleanup is visible in the data.

From April 2025, Spotify has been clearing out streams it counts as artificial: by its own reporting, 75M+ spam tracks in a year, with plays pulled back even from the biggest acts. We wanted to see whether it was working, so we watched the follower histories of the most exposed playlists in Chartmetric's daily data. The logic was simple: if the cleanup is real, the inflated giants come down, and real discovery has to be happening somewhere else, on the playlists where actual fans find new artists. That is what the rest of this report goes looking for.

The most powerful playlist in hip-hop is just where the turn reads cleanly.

• Eight years, one direction. RapCaviar, ~16 million followers, did exactly one thing for its entire tracked life: go up. Every quarter from 2017 to early 2025 was flat or higher.
• Then, for the first time, it turned. It peaked at 16.16M on 19 April 2025 (Chartmetric) and has posted thirteen straight months of decline since, to 15.74M today. The magnitude is modest (−2.6%); the direction change is the regime break.
• The plays fell harder than the followers. Label staff reported RapCaviar stream volume down 30-50%[1]; this separate estimate points in the same direction as the follower data.
• One instance of a wider turn. The same reversal is visible across the giants; RapCaviar is only the cleanest place to read it.

The figures above are measured. The explanations below are hypotheses; Spotify has not confirmed how much each factor contributed.

• The decline is gradual. Rolling removal of bot and dormant accounts could produce this pattern, although the follower series alone cannot establish the cause.
• It surfaces on the giants first. A fake account has to look human, so it follows the obvious lists: Today's Top Hits, Top 50, RapCaviar. Scrub the profiles and followers drain off the defaults, while the independents real fans chose keep climbing.
• Several factors overlap. Three shifts converged in 2024-25: the purge, the demotion of curated playlists for personalised surfaces, and hip-hop's cooling share. All three predict a monotonic riser finally turning down, so the cleanup is real but not the whole cause.
• The practical question is what remains useful. Follower totals have become less reliable as a proxy for reach. Demonstrated artist discovery is a better basis for evaluating playlists.

Large playlists are flattening while independents continue to grow.

The cleanup section above measured what Spotify removed; this is the market it left behind, in three movements. Strip the manufactured audience out and what stands is the part that was always real: the small hand-curated third-party list. Five measurements below, the follower curves, the deletion census, the discovery placements, the genre cut, and how reachable those curators are. Independent playlists are gaining relative importance as large Spotify-owned lists flatten.

How the sample was built: from our own cache we ranked every playlist by followers and took the top 200 Spotify-owned plus the top 200 third-party. The table below lists all 400; two are personalized "For You" lists (Release Radar, Best of the Decade For You) with no single follower series, shown without a trend. The follower-trend charts above are built from the lists that carry continuous history. The selection is purely mechanical and was fixed before any outcome was known, ranked by size because fakes cluster on the giants, so both classes are compared at maximum scale.

Indexed to January 2023, the independents Spotify doesn't run have compounded +70% while the owned editorial and algorithmic giants are up +19% and flat for the last twelve months. The gap widens, month after month.

• Worse than flat at the very top. The four most-defaulted playlists, Today's Top Hits, Top 50 Global, RapCaviar, Viva Latino, all peaked around 2025 and have lost followers every month since (−0.6% to −2.6% over the year). These are exactly the lists a fake account follows to pass as real, so a profile purge surfaces on them first; RapCaviar's −2.6% is the same pattern, not something specific to hip-hop.
• Chosen lists compound. Catalog titles like All Out 2010s (+101%) and phonk (+210%), and the independents generally: +70% median over three years vs editorial's +19%, and +10% in the last year vs +1%.
• What personalisation changes. Giants are defaults, the lists Spotify assigns you; independents are choices, the lists a listener seeks out. Strip the fake followers and all the growth sits on the chosen side. Personalised editorial keeps the placement and loses the loyalty.

Every playlist in the follower pull, sortable and searchable. 12-mo is the change in followers over the trailing year; from peak is how far the list sits below its own all-time high. Click a name to open it on Spotify.

Two percent of active playlists die in a matter of months, and the deletions concentrate on the ones selling placement.

• 825,000 re-pinged, ~1.9% gone. Every one of them was confirmed alive months earlier, so this is a live kill rate, not an archive rotting.
• The dead ones are the ones selling. A deleted playlist is 3.5× more likely to publish a contact email and 2.2× more likely to list an Instagram promo handle. Pay-for-play is being removed at more than double the base rate.
• And it takes the operator with it. Of the 216 curators behind those deletions, 42 (19.4%) lost the entire account, not one list.

• These were operations, not stray playlists. We watched individual submission networks disappear in full: every playlist in the network gone, and the profile running them terminated along with it. One carried 125 playlists; others were built on a single flagship with tens of thousands of followers.
• They all ran the same shape. A throwaway account, a catalogue of genre-titled playlists, and a "submit your track to [email]" line sitting in the bio.

Where Spotify credits artist discovery.

This section uses Spotify's “Discovered On” data to compare the playlist types associated with artist discovery.

• Independent lists carry 80 to 92% of discovery for every artist under a million monthly listeners.
• Editorial barely registers until the top. 1% at the bottom of the market, rising to about 27% for the largest artists.
• Independent playlists dominate discovery credits for smaller artists. By the time editorial shows up, they are already big.

Median follower count of the playlists that actually discover an artist, by artist size, independent against Spotify-owned. The lists doing the discovering are not the famous ones, and they are not close in size. Cut by genre so no average hides the cases that matter:

• 30 of 30 genres: the independent lists doing the discovering are smaller than the Spotify-owned ones in every single size tier.
• The gap closes as the artist grows, a median of 36× at the bottom of the market against 11× at the top. 28 of 30 narrow that way; Country and Indian Film/Desi are the two that do not.
• Runs on every artist we can place, all 64,687 with a genre and a Spotify-credited discovery list, six times the chart sample above. Same shape.

• The size is exactly what the purge is marking down. Fakes collect on the big defaults, so editorial's counts are the inflated ones being corrected: +70% independent vs +19% editorial over three years, on real listening.
• The discovering lists carry it on far fewer followers: a tenth of editorial's at the top of the market, a thirtieth at the bottom, where discovery decides a career.
• Relative to reported followers, independent playlists show 10 to 30× more discovery activity than Spotify-owned playlists, and the difference is largest for smaller artists. This is an observed ratio, not a causal estimate of placement value.

• Editorial is a narrow gate. All of it runs through about 12,000 distinct playlists, no more than ~2,400 at any single artist size: a small set of gates a handful of people control.
• The independent market is wide open. Discovery is spread across 257,000 lists: large, competitive, and the part of discovery a tool or a campaign can actually move.

The fair challenge first: are these small independent playlists the pay-for-play operations the census caught Spotify deleting? Mostly no. The clearest marker is an email funnel, a "submit your track to [email]" line, and we checked the playlists that drive discovery against a baseline of 687,474 ordinary playlists in our cache.

• A funnel is the exception. 93% of the playlists that drive discovery carry no submission pitch at all, at any artist size, and the solicitation marker the census tied to deleted playlists tops out at 3.5% in the most solicited band, under the 3.8% baseline. Below it at every tier.
• Quiet is not the same as closed. Not advertising for submissions says nothing about whether a curator can be reached. Most simply run their lists without a pitch inbox, and where a route does exist it is more often a social profile than an address.
• About one in eleven has a contact route. It reads like scarcity only if you are checking playlists one at a time. Across the pooled chart alone it is 4,382 reachable curators, every one of them credited by Spotify with actually discovering artists; that is a working list rather than a shortage, and a bigger one than any team could get through.

Finding these curators requires aggregating artist-level discovery records. A contact route is only visible once you have pulled the discovery graph for the whole market and re-keyed it by playlist, artist by artist. We did, and that is the rest of this report.

The Playlist Discovery Index.

The findings above turned into an instrument. We read the discovery graph of every artist we can reach and re-key it by playlist; what comes out the other side is a ranking of 107,240 independent playlists by demonstrated discovery, not follower count. Artists in, playlists out.

Playlist discovery performance by genre.

For every genre, the Playlist Charts list the independent playlists Spotify's own “Discovered On” data credits with artist discovery: ordered by observed discovery, screened for genre fit and artist size, and flagged where the curator is reachable. Below is a preview of the top ten per genre from the pooled all-sizes chart.

Access the full database with a PlaylistSupply Viral Annual or Record Label Annual plan.

Unlock the full charts →