Music Rights Data: A Complete Guide
The entities, identifiers and relationships behind every royalty payment — and where the model usually breaks.
Music rights data describes who owns what, where, and for how long. Get the model right and reconciliation becomes arithmetic. Get it wrong and every downstream number is suspect.
The core entities
- Composition: the underlying work, identified by ISWC, owned by writers and publishers.
- Recording: a fixed performance of a composition, identified by ISRC, controlled by the master owner.
- Release: the commercial product a recording appears on, often identified by UPC or EAN.
- Party: a writer, publisher, label or administrator, ideally identified by IPI or ISNI.
- Usage: an observed event — a broadcast, a stream, a sync — with a source, a territory and a timestamp.
The relationships that matter
One composition to many recordings. One recording to many usages. Ownership expressed as shares that must total 100% within a territory and a period. Rights that are always scoped: this share, in this territory, between these dates. Drop the territory or the date and the model silently stops reflecting reality.
Where it breaks
- Identifiers missing, duplicated or reused across entities.
- Shares that sum to more or less than 100%.
- Territory scope recorded as a note instead of a field.
- Effective dates absent, so historical ownership cannot be reconstructed.
- Party names stored as free text with no stable identifier.
Building something durable
Model entities separately rather than flattening everything into one row per track. Make identifiers required. Store shares with territory and date scope. Keep an audit trail of changes. Validate on import, not on discovery.
This work is unglamorous and it is the foundation for everything else: detection, reconciliation, evidence and valuation all read from the same graph.