Count the Plays Before You Count the Catalogue

A track can be available on a streaming service without receiving a play, and a play can be excluded from payment. Catalogue reviews need separate counts for delivery, listening and eligible consumption.
An independent artist performs at a live music venue. An independent artist performs at a live music venue.

About 106,000 new recordings reached streaming services each day in 2025, yet 88% of tracks received 1,000 streams or fewer that year. For an independent label, the number of recordings delivered is a measure of supply. The number that attract listening is a different measure, and neither automatically tells the label what will be paid.

Delivery creates the denominator

Most new recordings now enter streaming through independent distribution: Luminate puts that share at 96% in 2025. That makes delivery a routine first step for a large population of releases. It also means that a growing catalogue can make an operation look larger while leaving its listening concentrated in a small part of the repertoire.

One distributor’s figures show the gap. Of approximately 300 million tracks it had released that remained available on digital service providers (DSPs) in the United States and Europe, only 40 million had received at least one stream. About 260 million, or 87%, had no play. Availability was real for those tracks; consumption was absent. A delivery report therefore answers whether a recording reached a service, not whether anyone used it.

The pattern appears in other measures with different scopes. Luminate indicated in 2023 that two-thirds of tracks on platforms had between zero and 100 plays. Deezer said in 2025 that 78% of new uploads had received no streams during the preceding 12 months. Those populations and periods differ, so their percentages should stay separate. Each measure still gives a catalogue manager the same practical prompt: identify which recordings moved from availability into actual listening.

A stream can leave the payment count

Even a recorded play is not necessarily a payable one. Deezer said fully AI-generated tracks made up between 1% and 3% of streams on its platform, then flagged 85% of those streams as fraudulent and excluded them from payments. The upload count, the stream count and the payment count are three different totals. Treating any one of them as a substitute for another will distort a catalogue review.

This distinction matters beyond AI-generated recordings. A team assessing a release needs to know which count it is reading before it assigns a commercial meaning to it. A track appearing in a distributor’s delivery inventory establishes availability. A service’s listening data establishes activity under that service’s reporting. A payment statement establishes which activity produced income under the applicable accounting and exclusions. The Deezer figures give a concrete reason to keep the last two columns apart: flagged streams did not enter payment.

Platform removals create another reason to date every count. Spotify said it had removed 75 million tracks it described as spammy and had removed 50 million AI-generated tracks from recommendations since the beginning of 2026. A snapshot of what was delivered cannot, by itself, describe what remains available or how a service presents it to listeners. When a catalogue total changes, the team should establish whether it is looking at new delivery, listening activity or a change in platform treatment.

Review the catalogue in stages

A workable catalogue report can follow the path a recording takes through the system. For each release period, keep these measures distinct:

  • Delivered recordings: titles accepted into the catalogue inventory, with the service and territory attached.
  • Recordings with a play: titles that have crossed from availability into observed consumption.
  • Depth of listening: the distribution of plays across titles, including how many remain at zero or in the low-play range.
  • Paid activity: the streams and amounts that appear in payment reporting after any service exclusions.

The 88% of tracks at 1,000 streams or fewer in 2025 makes that third measure useful. A single total for the catalogue could rise while most titles remain in the same low-play range. Sorting titles by listening depth shows whether growth reaches more recordings or stays with a small group. It also gives an independent team a basis for deciding which releases need audience work and which are already drawing repeat attention.

That decision should use the artist’s own catalogue, not a platform-wide average as a target. Spotify estimates that roughly 13 million people have uploaded at least one song to its platform, while about 250,000, fewer than 2%, pursue music professionally. The uploader population includes many different intentions. An independent label can set a more useful review point from its own release costs and goals, then compare each title’s listening and paid activity against that point.

Make the next decision from the right count

The immediate action is to stop using delivered titles as a proxy for reach. At each reporting date, ask how many catalogue recordings received a play, how listening is distributed among them, and which activity reached payment. If a title has no plays, the next decision concerns audience discovery. If it has plays but no corresponding paid activity, the next task is to reconcile the service and payment reports. The distinction starts with a simple number: 40 million of one distributor’s approximately 300 million available tracks had received even one stream. Count that transition before assigning value to the size of the catalogue.

Previous Post
Litmus Music announced new capital commitments from Carlyle and KKR-managed funds.

Separate the Capital, the Administration and the Income

Next Post
Ignacio Meyer, president of Universal Music Latin America and the Iberian Peninsula.

An Executive Title Needs an Operating Metric