How to measure playlist results: saves, skips and cost per save

Most artists measure playlist promotion by streams, which is the one number that cannot tell you whether it worked. A playlist add is worth what its listeners go on to do: save the song, add it to their own playlists, replay it, follow you. Four numbers separate a placement that built something from a placement that rented a number for a week: save rate, skip rate, cost per save and algorithmic share. All four are visible in Spotify for Artists and a spreadsheet.

Public benchmarks give the bands to aim for. MusicPulse treats a save rate above 3% as strong listener intent, with 2 to 4% typical for editorial playlists (5 March 2026). Chartlex's published campaign tiers work out to roughly $0.005 to $0.04 per stream, and Ones To Watch reports 2,000 to 8,000 streams from a $150 to $300 campaign (Chartlex; 22 April 2026). Your own numbers decide whether a specific placement beat those bands.

The four numbers that matter

Save rate

The formula is simple: saves divided by streams for the same period. Spotify counts a save when a listener taps the plus button, and saves from music videos count too (Spotify for Artists).

To attribute saves to a playlist, compare three windows: seven days before the add, the seven days after, and the playlist's share of plays in the same window. If saves rose with the playlist's plays and stayed after the spike, the placement brought savers. If saves stayed flat while a playlist poured in plays, it brought traffic, not listeners. MusicPulse's retention analysis makes the useful distinction: a placement that produces plays without saves, follows or repeats is reach without growth (24 March 2026).

Skip rate

Skip rate is the fastest signal that a placement found the wrong audience. Spotify reports it per song, and the pattern to watch is the direction, not the absolute number. A skip rate that climbs while a particular playlist is active says the list's listeners do not want this song, no matter how large the list is. That is exactly the feedback loop you want: a wrong-fit list tells you to drop it, and a right-fit list shows up as falling skips and rising saves.

Two cautions. First, skip rate reacts to the song's opening seconds more than to anything else; a long intro produces skips everywhere. Second, short playlists and discovery surfaces have naturally higher skip rates than one's own fanbase. Compare a playlist's effect on skips with your baseline month, not with a single ideal number.

Cost per save

This is the number that makes paid and free routes comparable. Cost per save = total spend on promotion ÷ saves attributed to that promotion over 28 days.

Worked example: a $60 credit run produces 300 attributed plays and 18 attributed saves. Cost per stream is $0.20; cost per save is $3.33. Compare that with the same $60 spent on a video clip, a mixing session or ads to your profile. If the campaign's cost per save is higher than the cost of earning the same saves organically, the campaign did not pay for itself in audience terms, even if the stream number looked good.

The calculation also exposes bad services quickly. A $129 stream-target package that produces 5,000 low-engagement plays might look like $0.026 per stream, right in the published band, but if those plays produce no saves, the cost per save is infinite. Streams you cannot convert are not the product you needed.

Algorithmic share

Open Segmentation > Source of Streams in Spotify for Artists. Spotify splits your plays into active sources, such as the artist profile, the listener's library and queue, and programmed sources, such as editorial playlists, algorithmic mixes and other listeners' playlists (Spotify for Artists). The trend you want after a placement is:

• Programmed plays rise while the placement is active. That is the placement doing its job.
• Active plays rise afterwards. That means listeners came back on their own, which is the compounding part.
• Unclear sources stay near zero. A large unexplained share is one of the signs Spotify's artificial-streaming page flags (Spotify for Artists).

Third-party platforms such as Soundcharts can track playlist and chart movement across stores if you want a second view, but Spotify's own source split is the primary record.

The 28-day tracking table

One sheet, one row per playlist. Fill it at day 7 and day 28.

Artist Advisory's pitch guidance recommends reading saves, skips, adds and listeners across days 1 to 28 rather than day one, because late adds and algorithmic pickups are real (15 April 2026). The day 28 pass is where the verdict gets written.

Keep, ask again or drop

1. Keep: saves per 1,000 plays beat your release average and the skip rate did not rise. Pitch the next release first.
2. Ask again: plays were strong but saves were average, and the curator was responsive. A different song from the next release may fit better.
3. Drop: plays had no saves, skips rose, or the cost per save exceeded your alternative. Remove it from the list and, if it looks automated, report it via our bot playlist guide.
4. Log the reason. One line per verdict. The reasons are what improve the next list.

Compare your numbers with the public benchmarks in do playlists still work, and check the per-playlist view our stats after placement guide covers before you renew any spend.

Five measurement mistakes

1. Measuring streams only. Streams without saves measure the playlist's size, not its value to you.
2. Judging on day one. Placement effects and late adds build over weeks. Check at 7 and 28 days.
3. Never writing down the baseline. Without a pre-placement average, you cannot attribute the change.
4. Mixing paid and free placements in one number. Track cost per save separately for each, or you cannot tell what the money bought.
5. Renewing a subscription you never measured. If the last three months produced no attributable saves, cancel and put the budget into the next release's best-performing lane.