A split sheet is a short written record of who wrote a song and in what proportion, agreed by everyone who wrote it. That is the whole idea. It is not a contract in the intimidating sense, it does not need a lawyer to exist, and it takes about two minutes when everyone is still in the room.
It gets difficult later, which is the entire argument for doing it early.
What it needs to contain
- The song title, including any working title it was written under
- The date it was written
- Every writer's full legal name, not the artist name they go by
- Each writer's percentage, adding up to 100
- Each writer's performing rights organisation and their member number
- Each writer's publisher, if they have one
- Contact details for each writer
- Some form of agreement from everyone: a signature, or in practice often a message thread
Legal names and PRO membership numbers are the two people skip, and they are the two that cause the delay later. A society cannot pay a stage name, and tracking down a member number from a phone contact two years old is genuinely hard.
Writing splits and master splits are different
A split sheet covers the song: the composition, the thing that exists as melody and lyric regardless of who recorded it. The recording is a separate asset with its own ownership, and producers are often involved in that side as well as, or instead of, the writing side.
Conflating the two is one of the most common sources of confusion, because a conversation about "points" and a conversation about "percent" can sound identical in a session and mean two different things. If a producer contributed to the composition, that belongs on the split sheet. If the agreement is about the recording, that is a separate note and should say so explicitly.
How to decide the percentages
There is no legal formula. The most common default is an equal split between everyone in the room who contributed to the writing, and it is popular because it is fast, hard to argue with, and avoids relitigating who suggested which chord.
Some people split by contribution type instead, treating music and lyrics as halves. Both are normal. What matters far more than the method is that it is agreed and written down on the day, by everyone, rather than assumed.
If a session includes someone who did not end up contributing to the writing, write that down too. A note saying a person is on zero percent and agreed to it is worth more than their absence from a list.
Why the same day matters so much
On the day, everyone remembers what happened and nobody has a financial reason to remember it differently. Six months later, memories genuinely diverge, and if the song has done something in the meantime the conversation is no longer a friendly one. Nothing about the split changes; only how hard it is to agree.
The other reason is registration. You cannot register a work properly without knowing the writers and shares, and unregistered works do not collect. Money accrues and sits waiting, and claiming it retrospectively is slow.
Where to keep it
Anywhere you will still have it in five years, and where it is attached to the song rather than filed by date. A photo of a handwritten sheet in a camera roll technically counts and is better than nothing, but it is not findable, and findability is the whole point.
The practical test is simple: if someone asked you today who wrote a song you worked on three years ago and what the splits were, how long would it take you to answer with confidence? If the answer is more than a minute, the record is not doing its job. This is the problem Locker was built around, keeping credits and splits attached to the song itself, but the discipline matters more than the tool.