SplitGraf Settlements (Beta)

Royalty reconciliation that ends with a statement you can prove.

SplitGraf Settlements takes the CSVs your distributor sends, matches every row to the right asset, applies your locked splits, and produces statements that are traceable back to the source agreement, not just the source spreadsheet.

Currently in beta. Ingestion and matching production-ready. Statement generation and verification records in active development.
Workflow
01
Revenue CSV
02
Normalised rows
03
Asset match
04
Split applied
05
Statement generated
06
Record preserved
Problem

The reconciliation problem has outgrown the spreadsheet.

A finance team or royalty administrator opens a quarterly cycle. Reports arrive from distributors, sub-publishers, and platform-direct programs in different formats. Each row needs to be matched to an asset in the catalog. Each match has to apply the right split, in the right version, at the right time.

Reports arrive in different formats.

CSVs from Spotify, Apple, YouTube, TikTok, and a dozen sub-publishers each arrive with different column conventions.

Identifiers are inconsistent.

ISRC fields are sometimes blank, sometimes wrong, sometimes formatted inconsistently.

Matching happens in spreadsheets.

Asset matching gets done in VLOOKUP across catalog spreadsheets that haven't been audited in years.

Statements ship without the record.

Statements get generated in Excel, exported to PDF, and sent with no audit trail.

Before / after

What changes when reconciliation has a record.

The difference isn't faster accounting. It's a continuous truth. The reconciliation answer and the agreement answer are the same answer, and they live in the same system.

Without SplitGraf

Distributor CSV arrives -> matched manually against catalog spreadsheet -> splits looked up in document drive -> calculated in Excel -> exported to PDF -> sent to artist -> six weeks later artist asks "why this number?" -> finance team digs through email archives to reconstruct the answer.

With SplitGraf

Revenue CSV ingested -> rows normalised and matched deterministically -> matched events linked to locked splits from the original agreement -> statement generated with the full record attached -> artist receives statement they can trace back to the moment of agreement -> questions answered in minutes, not weeks.

Solution

How it works.

You upload a revenue report. SplitGraf normalises the rows automatically, matches against your catalog using deterministic resolution, and routes ambiguous rows to review rather than guessing. When an artist or auditor asks "why this number?", the answer is already in the system.

Rows are matched against your catalog.

Unique ISRC matches first. Unique title matches as fallback. Ambiguous matches are flagged for review.

Matched events connect to locked split snapshots.

The system applies the split version that was active at the time the revenue was earned.

Statements preserve the full record.

The chain from CSV row through asset match through applied split through statement line is available for audit.

Workflow

The workflow, step by step.

Step 1

Upload a revenue report.

CSV or XLSX from Spotify, Apple, YouTube, TikTok, sub-publisher statements, distributor reports. Format variations handled automatically.

Step 2

Rows normalised and matched.

Each row is cleaned, standardised, and matched to your catalog. Unique identifiers win. Ambiguous matches are flagged for review.

Step 3

Exceptions queued.

Unmatched rows, ambiguous matches, and rows with missing required fields route to a structured review queue. You see exactly what couldn't be matched and why.

Step 4

Splits applied from locked snapshots.

Matched events link to the agreement version that governed them at the time of earning. Historical accuracy is preserved.

Step 5

Statements generated.

Trusted events group by recipient, calculate against applied splits, and output as statements with full record references.

Step 6

Operating record preserved.

The complete chain from raw CSV through to final statement is held as a verifiable record for audit, dispute resolution, or artist questions.

Benefits

What this changes for your operations.

Reconciliation cycles that close cleanly.

Deterministic matching handles the routine 80-95% of rows automatically. The exception queue surfaces only what genuinely needs human judgment.

Statements that survive questions.

When an artist or auditor asks how a number was calculated, the answer is in the system. The chain from CSV row to applied split to statement line is preserved.

A reconciliation history that holds.

Quarterly cycles produce records that don't degrade. The audit trail from this cycle is still usable two years from now.

Built for

Built for the operations carrying real reconciliation weight.

This product is built for teams where reconciliation is a budgeted operational function, not a quarterly inconvenience.

Royalty administration firms

Teams processing statements across multiple label and publisher clients, where format variation, matching scale, and defensible audit trails are the daily operational challenge.

Mid-market and major-adjacent publishers

Teams reconciling distributor reports, sub-publisher statements, and platform-direct revenue across multi-territory catalogs.

Catalog funds and rights holders

Teams managing income from acquired portfolios where agreements are sometimes inherited, rarely clean, and always being questioned.

Publishing operations teams at scale

Teams handling mechanical and performance reconciliation across multiple territories, source systems, and counterparties.

Trust

Why the matching is trustworthy.

Refusing to guess is the right behaviour for a financial system. A false match is worse than a missed match because a false match enters the books as fact.

Matching is deterministic.

When the system can prove a match, unique ISRC or unique title against a single catalog asset, the event is trusted.

Ambiguity routes to review.

When the system cannot prove a match, the event is routed to review, not silently guessed.

Historical splits remain historical.

The snapshot version applied is the one active at the time of revenue earning, not whatever version happens to be current.

The record is append-only.

Every ingested report, matched event, applied split, and generated statement is part of an append-only operational record.

Beta status

What's production-ready and what's in active development.

We're being explicit about what works today and what's coming next.

Production-ready today

Revenue report ingestion (CSV, XLSX) with format normalisation.
Deterministic asset matching with unique ISRC and title resolution.
Exception queues for ambiguous and unmatched rows.
Trusted monetization event creation linked to locked splits.

In active development

Statement generation against trusted events.
Externally verifiable settlement records at the statement level.
Verification links for artist and counterparty access.
Pricing

Pricing.

Reconciliation Pro

For active finance and rights operations processing reports from multiple sources at meaningful volume.

Enterprise

For royalty administration firms, publishers at scale, and catalog funds with high-volume, multi-tenant, or custom integration needs.

All plans include the full reconciliation workflow, deterministic matching, exception queues, and operating record. Pilot pricing is available for selected reconciliation operations. Book a walkthrough to discuss what fits your team.
FAQ

Common questions.

Is this royalty accounting software?

Not in the traditional sense. We are not competing with Curve or similar systems on royalty calculation depth. We handle reconciliation: the layer between receiving revenue reports and generating defensible statements.

Will this force us to migrate off our current systems?

No. SplitGraf is designed to run alongside your existing accounting and rights systems during evaluation and onboarding. Most early customers run parallel for a quarter or two before deepening their commitment.

What revenue sources can you ingest?

Any CSV or XLSX format. We handle Spotify, Apple Music, YouTube, TikTok, Amazon, distributor consolidated reports, sub-publisher statements, and most platform-direct programs.

How accurate is the matching?

For rows with valid, unique ISRCs, matching accuracy is near-100%. For rows requiring title resolution, accuracy depends on catalog metadata quality. Ambiguous matches are flagged for review rather than guessed at.

What about historical reconciliation?

You can ingest historical reports and reconcile them against locked splits as long as the agreement records exist in the system. The system applies the split version active at the time of earning, not at the time of ingestion.

How does this integrate with existing accounting systems?

Trusted events and generated statements can be exported in formats consumable by external accounting systems. Customers running parallel systems during evaluation is the norm, not the exception.

Is this safe for sensitive financial data?

Yes. Workspace data is isolated at the database level. Revenue data is encrypted at rest and in transit. Role-based access controls limit who in your organisation can see what. Full audit logging applies to financial operations. See the trust page.

Can we evaluate before committing?

Yes. Book a walkthrough first to scope your reconciliation volume and workflow. We typically offer 30-60 day pilots for serious prospects, with success criteria agreed in advance.
Next step

Reconciliation that connects to the agreement underneath.

Most reconciliation tools answer the question "what did the artist earn?" SplitGraf Settlements answers a different question: why did the artist earn that number?