Emptying The Black Box

Photo by Emiel Maters

Last year I wrote about real issues to be found in the MLC’s dataset- an exploration of what was really sitting inside the industry's most transparent copyright dataset. The headline then: the MLC reported around 90% of royalty value matched, while our analysis showed only 22% of ISRCs had been attributed to any work- some 170 million of 217 million recordings sitting unlinked, with the long tail holding more recoverable money than the industry's hits-first workflows were ever going to find.

A year on it's time to revisit those numbers, keeping in mind that the data hasn’t changed dramatically, but because the rules are about to. From early 2027, the MLC begins its congressionally mandated market share distributions. In plain terms: the black box is about to be emptied, one month at a time, and every dollar in it will be handed to someone. The only question is whether it's handed to its owner.

Anatomy of the Black Box

The terminology matters here, because there is no standardised industry glossary, and music publishing is full of nuance that can be easily mis-interpreted:

• Unmatched royalties are royalties for recordings reported by DSPs that haven't been linked to any song in the MLC's database. Nobody knows (yet) whose they are.

• Unclaimed royalties are subtler- a recording may have been matched to a song, but less than 100% of the shares of that work have been claimed by members. The money knows where it belongs; the owner just hasn't turned up.

There are also two separate pots. Blanket royalties are those collected by the MLC itself under the blanket licence since January 2021. Historical royalties are the ~$397 million that 21 DSPs transferred over in 2021, covering unmatched usage all the way back to 2007. The 2027 distributions start with the blanket pot, 2021 usage first. The historical pot will follow later.

Then vs. Now

Credit where due: the MLC's progress is real. Over $4 billion processed. A 92% average match rate overall, and 94% for 2021 usage specifically. More than 90,000 members and 55 million works in a publicly searchable database. On the historical pot, $317 million of the $397 million they’ve received (nearly 80%) has been matched, and $229 million distributed.

And yet- look at what remains, because this is where "matched doesn't necessarily mean matched" (point #2 of our original newsletter) comes home to roost.

For 2021 usage alone, roughly $77 million remains unclaimed- month by month, between $6.0m and $7.1m, a consistent ~11.4% of each monthly royalty pool. Sit those two numbers side by side: a 94% match rate, and 11.4% of the dollars still without an owner. Both are true. Match rates count usage matched to works; the unclaimed pool also contains money for works that are matched but where shares sit unclaimed. The gap between those two figures is precisely where publishers should be looking.

On the historical side, $168 million of the original $397m remains undistributed- and $88 million of that is matched but unclaimed. The industry knows which songs that money belongs to. It is sitting there, named, waiting.

The MLC's framing is that what remains is small change- fewer than 5% of remaining unmatched recordings have accrued more than $1, and the rest average 5-6 cents each. Though accurate it is also exactly the point the original edition made: small claims at scale are the goldmine. Pennies multiplied by millions of instances is how you arrive at $77 million. No human team can claim 5-6 cents economically. Tooling can.

The Great Redistribution

Here is how the emptying is going to work. From early 2027, the MLC will distribute one month of remaining 2021 blanket royalties each month, starting with January 2021 (expected to be just under $7 million). The money is allocated pro-rata based on actual distribution data, with interest at the statutory rate. Every song that earned royalties that month receives its share of the leftovers.

Read that again from a different angle: this is a transfer of wealth from rightsholders with data problems to rightsholders with market share. A publisher with 2% of the market will passively receive something like $129,000 a month during 2027- money that, by definition, belonged to someone else and couldn't find its way home. This also bypasses for now the fact that rightsholders with market share are hence incentivised to be rightsholders with data problems…

Critically- once a month's distribution is made, claiming for that month closes forever. Unclaimed shares can still be claimed today, with all back royalties and interest attached. From 2027, that stops being true one month at a time. A six-year window (twice the statutory minimum, to the MLC's credit) becomes a hard deadline on a rolling schedule.

What We See in the Wild

The dashboards tell you the size of the pool. They don't tell you whether your catalog is a net receiver or a net donor. Here's what we find when we actually look.

A recent client with a significant catalog had 9.5% of their entire catalog carrying material work-level issues at MLC across underclaimed works (as measured against their own internal catalog expectations) and duplicated work registrations. The more startling figure: 58% of their works had no ISRCs matched at the MLC at all- works earning nothing from US mechanicals.

Duplicates deserve a special mention ahead of 2027. A duplicated registration doesn't just create admin noise- it fragments matching and muddies exactly the distribution data that pro-rata allocations will be calculated from. Fixing duplicates was always good hygiene, but now is directly financial.

What's it worth? As a general rule we see a minimum 10-15% collections uplift from catalog cleanup work. One large PRO recently told a client their revenue had doubled over the period of our cleanup- and being honest, as they were, not all of that can be attributed directly to our work. But by their own admission the direction of travel is not subtle. This isn't growth from new repertoire or better deals. It is money that was always owed, finally arriving.

The Clock Is Now Ticking

A year ago, our conclusion was that the value was "hiding in the gaps between datasets", waiting to be connected. That is still true, but the deadline is becoming more real. The gaps are about to be monetised- just not necessarily in your favour.

The playbook before 2027 is not complicated:

• Reconcile your internal catalog against what the MLC actually holds (registrations, shares, duplicates)

• Claim the matched-but-unclaimed shares- with interest, they are the cheapest revenue you will ever recover

• Investigate the unmatched recordings at scale, because 5-cent claims only make sense with automation

• Resolve duplicates before they quietly reshape your share of every distribution to come.

The MLC has done what almost no other collective in the world does: shown everyone the money, published the data, and built the tools required to affect change. After six years, it is reasonable for the leftovers to be redistributed. The rightsholders who lose out in 2027 won't be able to say they weren't warned- the warning has been sitting on a public dashboard the whole time.

In 2027, every publisher will be on one side of the market share distributions: collecting their own money, or quietly collecting everyone else's.

Are you certain which side your catalog is on? I'd genuinely like to hear how prepared (or not) the industry feels for this one, especially from those who've already dug into their own unclaimed picture.