The Biggest Catalogues Are Leaking Royalties
There's a comfortable story the catalogue market tells itself about unclaimed royalties. It goes like this. The big copyrights, the valuable ones, the songs everyone actually knows, are all administered perfectly. The money that goes missing is the long tail. Bedroom producers, dormant catalogues, songs nobody streams. The implication is that if you own something genuinely valuable, you don't have a problem.
The long-tail framing has been around for years. It stuck. It's the kind of thing that sounds sensible, gets repeated, and quietly becomes the thing everyone assumes is true.
The numbers the MLC published about the MLC
When the Mechanical Licensing Collective launched in 2021, it received roughly $424.4 million of historical unmatched royalties. Money that had accrued, that was owed to somebody, and that nobody had been able to match to the right rights holder. By the end of that first year, the accumulated pool had grown to around $561 million.
The MLC has done real work since then. By its 2024 Annual Report, the historical unmatched pool was down to roughly $195.8 million. Genuine progress, and worth saying out loud.
But here's the part that complicates the long-tail story. In comments to the Copyright Office's 2024 re-designation review, Jeff Price of Word Collections put the total sitting in limbo closer to $1 billion. That's money where, in a lot of cases, somebody knows roughly who it belongs to and it still isn't moving.
A billion dollars is not a rounding error on the bedroom-producer long tail. Some of that is sitting against songs you'd recognise.
What we found when we looked at the sharp end
We've been building cover version models for a while. The idea is fairly simple to describe and quite fiddly to execute. A song gets written once and recorded many times, and the royalties from all those recordings should flow back to the original composition and its writers. In practice the links between covers and originals break, or never get made, and the money pools up against recordings that aren't properly connected to the writer.
Last week we launched our top Spotify songs model. The very sharpest end of the market. The most valuable, most streamed, most scrutinised compositions in the world. If the long-tail story were right, this is exactly where we'd expect to find nothing.
We found mistakes.
Not in the obscure corners. At the top.
The worked example
Take a catalogue we work with. We'll keep it semi-anonymised, but the writer is a working songwriter with a real body of work.
When we came in, we merged the writer registration, which sounds like admin and is actually where a lot of the value leaks out. Then we ran the catalogue through our cover matching model and linked previously unlinked cover versions back to the originals. Each one of those links is a flow of royalties that was either sitting in a hold pool or accruing against a recording that wasn't connected to the writer. Those covers now account for around 20% of the catalogue's annual returns.
That's the methodology. Merge the registration so the writer is one entity rather than several fragments. Run the cover version matching against the model so you catch the recordings that should be flowing back to the composition. Link them. Then the money has somewhere to go.
It's not glamorous. It's a catalogue we cleaned up, song by song, recording by recording.
Why this is the thing we keep coming back to
We invest in catalogues. But the investment isn't the whole job, is it? You buy a stake and then you do the work. You audit, you clean up, you link the cover versions to the originals, and you find the royalties that were never reaching the right owner.
At the sharp end, that work has run into the hundreds of thousands of dollars in unclaimed money on individual catalogues. Not theoretical money. Money with a name on it that wasn't getting there.
So when a catalogue owner or their representative asks what we actually do to increase catalogue value, this is one of the answers. We're not theorising about an industry problem from a distance. We have a model, we have a method, and we have the receipts to show what it found at a level of the market everyone assumed was already clean. The MLC is the most transparent PRO on the planet, and if the mistakes can be discovered there, then you'd be naive to assume the mistakes weren't rife in other territories. This is a global issue.
The long tail is real. But it was never the whole story. The interesting money, the surprising money, is sitting higher up than the comfortable version admits. We've spent enough time looking now that we are confident of that, and we have the receipts to prove it.