You hold rights. You expect collection. You file what you can. And year on year, the picture repeats: royalties accruing in places you cannot see, against works you cannot prove, in territories you cannot reach. The black box is not a metaphor - it is a measurable gap. Closing it is our work.
Spelling differences, punctuation differences, sub-title additions, language variants. To a human ear, the same song. To a collection society's matching engine, four works with no obvious link. The royalty pools fragment. The statements arrive incomplete.
ISRCs are supposed to be unique. In practice, collisions happen - the same identifier issued to two different recordings, or the wrong identifier attached at distribution. The downstream effect is silent: usage is logged, but credited to the wrong rights holder.
A work performs in a territory whose society has no registration on file. Usage accrues. The pool builds. The money sits, year after year, against a work that does not formally exist in that territory's records.
Mergers, acquisitions, catalogue sales, and inherited estates all change the rightsholder of record. Society databases lag. Statements continue to flow to legacy addresses for years after the actual rights have moved.
An estimated $561M sits unclaimed at the MLC alone. One country. One rights type.
Black Box applies the rigour of forensic software to the business of royalty collection.
Our reference catalogue spans 207 million tracks. The cross-reference engine works across 105 royalty-collecting territories. The output is not an estimate - it is a documented case file built to the standards a tribunal will accept. We do not chase royalties. We establish the facts, build the evidence pack, and file the claim.
The CWR Scoring Engine reads catalogue registration data, matches it against our 207M-track reference catalogue, and applies the published payout models of every monitored territory. Output: an earnings expectation per work, per market - and the gap between expected and actual collection. This is not a search. It is a structured comparison.
The Match Queue presents human reviewers with proposed matches scored across seven factors - title proximity, ISRC alignment, performance window, writer attribution, geographic context, recording metadata, and collection-society fingerprint. Colour-coded confidence. Bulk actions. Audit trail on every decision.
The ISRC Health Checker runs batch verification across the catalogue. Five status levels: Clean, Collision, Unregistered, Mismatch, Not Found. Surfaces the identifier problems that would otherwise cause a society to reject the claim or credit the wrong rights holder.
The Evidence Pack Generator compiles the documentary record per claim: registration history, performance logs, usage timelines, listener geography, society correspondence, contradiction summaries, and a written verdict statement. Output formats meet the documentary standards of MLC, ASCAP, PRS, and tribunal-grade rights bodies. CWR, CSV, or PDF.
The Case Manager handles lifecycle from filing through to acknowledgement, dispute, and settlement. Per-case timelines. Per-territory recovery breakdowns. Estimated royalties versus actual recovery. Cases close with a documented sum recovered - not an estimate.
Six months inside a back-catalogue recovery: the writer, the gaps, and the claim that settled in ninety-one days. Read the full case in Field Notes.
Black Box engagements are scoped against a real catalogue and a real recovery question. We do not sell a subscription to a tool. We deliver an outcome - the evidence pack, the filed claim, the recovered sum.
Three engagement shapes covering most catalogues:
A forensic royalty investigation product for publishers, writers, and estates. We run a catalogue against a 207M-track reference catalogue, identify revenue gaps across 105 territories, build the evidence pack, and file the claim with the relevant collection society. Output: documented recovery, not an estimate.
Recording-level metadata, ISRC and ISWC identifiers, registration histories from the major societies, performance and broadcast logs, streaming usage where licensed, and writer/composer attribution chains. Updated continuously, partitioned by territory.
Initial sweep against a 1,000–5,000 title catalogue: two to four weeks. Evidence pack generation: days, not months, per claim. Settlement timelines depend on the receiving society - most close within ninety to one hundred and eighty days. The case study above closed in ninety-one.
MLC, ASCAP, BMI, PRS, PPL, GEMA, SACEM, SOCAN, APRA - the major mechanical and performance societies across our 105 monitored territories. Evidence packs are formatted to the standards each society requires.
Three shapes - SCOUT (sweep + estimate), FILE (sweep + evidence packs + filings), and RECOVER (full lifecycle to settlement). RECOVER engagements can be priced as a share of recovered revenue. SCOUT is fixed-fee. FILE is time-and-materials with a documented evidence pack per claim.
The methodology is documented and reviewable under NDA. The cross-reference logic, scoring weights, and evidence-pack composition are not public - that is the moat - but the documentary outputs are auditable and built to be defended in front of a tribunal.
Yes. Disputed estates, contested writer credits, catalogues mid-acquisition. The forensic standard is the same: establish the fact against the documentary record. Where contested matters require legal representation, we work alongside the instructed firm.
Initial enquiries are confidential and conducted under NDA. The first conversation is methodology. The second is the size of the gap. After that, you decide whether to file.