A split sheet records what everyone agreed. It does not record what happens when two people file different numbers for the same song — and that is the moment a co-writing dispute starts.
A split sheet is a one-page agreement identifying the ownership percentage each collaborator holds in a composition. It is signed, it is binding from the moment of signature, and it is the single most effective step any co-writer can take. Every writer, producer and topliner should complete one for every session. Nothing below argues otherwise.
What a split sheet does is settle the outcome. Sixty to a writer, forty to a producer, agreed and signed. Where one exists, everyone signed it, and everyone registers the same numbers, it governs, and most collaborations never need anything more.
What it does not do is guarantee that everyone registers the same numbers. A split sheet is an agreement between the people in the room. Registration is a separate act performed later, often by different people at different publishers, from memory or from a copy that may not match what another co-writer files. The sheet can be perfectly signed and still produce a conflict, because the conflict lives in the gap between what was agreed and what was submitted.
The sheet also depends on separate parties acting on it correctly. A producer's publisher registers a track; a songwriter's administrator registers the same track independently. If the two filings disagree — a rounding difference, a mis-transcribed percentage, a writer credited or omitted in error — the sheet may never get consulted, because the collecting body sees the conflict first.
For most songs this never surfaces. The registrations match, the song is administered, the royalties flow. A song can be registered correctly by one party and incorrectly by another for years before either side checks the other's filing, and by the time a sync placement or a catalogue sale draws attention to the numbers, nobody remembers which registration was right.
Split disputes are rarely about theft. They begin with two administrative filings that do not match.
One publisher's system shows a fifty-fifty split. The other shows sixty-forty. Neither party necessarily did anything wrong — a spreadsheet updated after a late change and never reconciled, a track registered before the final sheet was signed, a team registering from memory. Each filing is internally consistent. Neither was checked against the other until the collection society flagged them as incompatible.
Three situations move this from an administrative mismatch to a formal problem.
The first is the song that becomes valuable. A discrepancy nobody would chase on a song earning nothing becomes worth pursuing once the song earns, and a party who never noticed the mismatch has every reason to notice now.
The second is delayed registration. Publishers do not always register a work the day it is finished. Where one party registers promptly and another months later, using an updated split, the later filing conflicts with the earlier by default, whatever the agreement said.
The third is the missing sheet. Where none was signed, each party registers what they believe the split to be, with nothing to check either registration against. That conflict is not a disagreement about what happened — it is the absence of anything to disagree about, dressed up as a dispute.
Where no split sheet was signed at all, a separate body of law decides the shares, and the co-write problem that produces — pre-existing material, joint authorship, who contributed what — is covered separately in our post on proving song authorship before a co-write. This post concerns the narrower administrative failure: a sheet that was signed, and registrations that fail to match what it says.
The consequences arrive through the collection system rather than the courts, and in the United States the mechanics are set out in the Mechanical Licensing Collective's published dispute policy. Where two parties claim overlapping shares in the same composition, the MLC flags a conflict and places the affected shares into suspense. Income stops for everyone attached to the work, including collaborators who are not party to the disagreement and whose own registration was correct.
The MLC's own 2024 filing with the United States Copyright Office puts a figure on the practice. Royalties held from distribution pending legal claims, ownership disputes, or review of eligibility for payment amount to less than one per cent of total royalties reported since the organisation began operating. Those categories are broader than registration conflicts alone, and the filing does not break them down. But across a body handling the mechanical royalties of an entire national streaming market, a fraction below one per cent is still a substantial sum held back from writers, and conflicting filings are one reason it stays there.
Disputed splits are one of several reasons royalties go on hold, and not the largest. They are, however, one a writer can prevent — not by writing better, but by registering in a way that cannot conflict with itself.
That is the practical shape of these disputes. Not a courtroom, but a frozen account and a stalled reconciliation between two publishers' systems, resolved by whoever can show which registration matches the agreement.
Reconciling a conflicting registration is rarely expensive outright, but it is slow relative to the sums involved, because both sides must confirm which record is right before either publisher amends a filing. A writer who can point to a single dated agreed document resolves that in one exchange. A writer who cannot is relying on two administrative teams working to their own timetable.
Co-writers are not undocumented. A split sheet gets signed, emailed, filed with a publisher, sometimes forwarded between two or three administrators before it reaches a collection society. Where a conflict is genuine rather than a simple data-entry error, this trail frequently resolves it, and a writer who can produce the original signed sheet is in a far stronger position than one who can only describe what was agreed.
The limitation concerns which copy is authoritative rather than whether a sheet exists. A split sheet forwarded by email exists in several inboxes, each a copy rather than the original. A version saved to a publisher's system may have been re-typed by an administrator rather than uploaded as signed. Where two registrations disagree, each side can usually produce a document supporting its own filing — the question is which reflects what was signed.
Where both sides accept the same document as authoritative, none of that matters. Where a dispute turns on which version is correct — the sheet as originally signed, or a later verbal adjustment one party registered and the other never agreed to — the copy each side holds is a copy they have handled, forwarded, and in some cases amended themselves.
The gap is specific: not whether a split sheet was signed, but whether the version each party is relying on is provably the one that was agreed, unaltered, from the moment of signature onward.
What closes the gap is depositing the split sheet itself, at the moment it is signed, so that every party's registration can be checked against one fixed original rather than against whichever copy each side is holding.
When a file is deposited, it is hashed using SHA-256, the hash is timestamped by an accredited Trust Service Provider under RFC 3161, and the timestamped hash is anchored to the Bitcoin blockchain via OpenTimestamps. This is a single automated sequential process, each step depending on the output of the preceding one. The result is a portable certificate recording the file name, the cryptographic fingerprint, the timestamp, and the blockchain anchor.
A free RFC 3161 timestamp from an unaccredited service produces a technically valid record but carries no legal presumption, leaving the holder to argue for its accuracy in any proceeding. With optional eIDAS Article 41 qualification from an accredited Qualified Trust Service Provider, the timestamp carries a legal presumption of accuracy across all 27 EU member states.
That presumption stops at the EU border, and most of the disputes described here are American in character. The United States has no statutory equivalent: under Federal Rule of Evidence 901, a documented and reproducible process supports admissibility, and the record's weight is assessed on ordinary evidentiary principles rather than presumed. UK law affords the certificate strong evidential standing without the statutory presumption. For a US writer facing a PRO or MLC conflict, the value of the record is practical before it is legal — it answers the administrator's question about which version is authoritative long before any question of admissibility arises.
This maps onto how a conflict gets resolved in practice. Under the MLC's published dispute policy, when claims conflict the affected shares go into suspense and claimants are given a fixed window to update their claim or submit substantiating documentation. The document that substantiates a split claim is the signed split sheet. The process does not ask who remembers the session more clearly; it asks which party can produce the agreement.
So the practical application follows the shape of the disputes. The moment a split sheet is signed, it is deposited — before it is emailed to a publisher, before an administrator re-types it into a registration system, before any of the copies that will later disagree come into existence. Where a conflict arises over which percentages were agreed, the certificate settles it against a fixed original rather than whichever copy each side happens to have kept.
The same approach covers amendments. Where a split is renegotiated or a late writer added, each revised version deposited on the day it is signed joins a dated sequence, so a later dispute over which version is current has an answer neither party must recall from memory.
None of the material leaves the collaborators' hands. Only the fingerprint is submitted, so the signed sheet stays under the writer's control while its existence in that exact form on that date becomes independently verifiable. Over a career, the accumulated certificates amount to a dated record of every agreement reached, checkable against whatever a publisher's system later shows.
None of this displaces the split sheet. The sheet remains the document that governs the percentages, and it should still be signed at the end of every session. The anchored record does something the sheet alone cannot: it fixes which copy of that sheet is the real one, before anyone has a reason to need to know.
For a songwriter or topliner, the anchored record settles a conflict at reconciliation rather than argument. When one publisher's registration does not match a co-writer's, the certificate identifies which filing tracks the signed agreement.
For a producer, the same record guards against being registered out of a split by someone else's administrative error, since the signed document — not a re-typed copy — is what any conflict gets checked against.
For a publisher or administrator holding a contested registration, an independently dated copy of the original sheet changes the position immediately. A dispute that would otherwise run on two internally consistent but conflicting filings acquires a fixed reference point neither publisher's system produced, which is frequently enough to resolve the claim and release the suspended income.
For a catalogue owner assessing an acquisition, the same records answer a question that ordinarily has none. Unreconciled splits are a known discount factor in valuation, because a buyer inherits every unresolved registration conflict attached to the works. A catalogue whose split sheets carry dated, independently verifiable originals carries measurably less of that risk.
Co-writing runs on trust, and most of the time trust is sufficient. The disputes that cost real money are rarely about who wrote what. They are about two publishers' systems disagreeing over a document nobody thought to pin down. Sign the split sheet. Then make sure the version everyone registers from is the one that was signed.
This post provides general information about the role of cryptographic evidence in music copyright disputes. It is not legal advice. For advice on a specific matter, consult a qualified lawyer in the relevant jurisdiction.
Related Reading
Co-Writing Splits and Split Sheets: A Practical Guide — Songpact
Proving Song Authorship Before You Walk Into a Co-Write
Sample Clearance Does Not Cover the Part You Wrote Yourself
James Snell is the founder of Provlyn, a platform providing cryptographic prior proof of IP ownership. provlyn.com