In May, I sat in a session at Music Biz in Atlanta called “From Spreadsheets to Systems: Rebuilding the Publishing Back Office.” The technology conversation was interesting. The room dynamics were more interesting.
The younger people in the room, many of them running music tech platforms or working inside modernized royalty operations, talked about the back office like it was a solvable engineering problem. Clean the data, standardize the formats, automate the reconciliation, pay artists faster. Done.
The veterans in the room, people who have spent decades making the current system work, were noticeably more measured. Not obstructionist. Measured. Their message, sometimes stated and sometimes implied, was some version of “you don’t understand how deep this goes” or “we’ve tried this before.”
Both groups are right about something. And the tension between them is the actual story, because the technology to fix music data interoperability largely exists. What doesn’t exist is the will to adopt it, and that has everything to do with whose job depends on the mess.
Why is music data such a mess in the first place?
Every song generates data across dozens of systems that were never designed to talk to each other. A single track touches a distributor, a label’s catalog system, a publisher’s rights database, multiple PROs, mechanical licensing agencies, dozens of DSPs, sync agencies, and neighboring rights societies. Each one stores the same core information, titles, writers, splits, ISRCs, ISWCs, in its own format, with its own field names, its own validation rules, and its own tolerance for missing data.
Standards exist on paper. DDEX handles recording-side data exchange reasonably well between major players. CWR handles works registration for publishers. But here is the uncomfortable truth: standards that are inconsistently adopted function as just another format to translate. When one PRO accepts CWR 2.2, another requires 3.0, a mid-size publisher still emails Excel templates, and an indie label doesn’t know what CWR is, you don’t have a standard. You have five dialects and no interpreter.
The result is familiar to anyone who works in this business. Unmatched royalties sitting in black boxes. Writers waiting 12 to 18 months for money that was earned in a single stream. Small publishers employing full-time staff whose entire job is reformatting spreadsheets. The estimates on unattributed royalties float in the hundreds of millions of dollars per year, and the honest answer is nobody knows the real number, because the data is too fragmented to count.
What would a “Zapier for music data” actually look like?
Zapier didn’t succeed by convincing every software company to adopt one standard. That’s the insight most people miss. Zapier succeeded by accepting that standardization would never happen and building a translation layer on top of the chaos instead.
Apply that model to music data and you get something like this:
A neutral middleware layer that connects to each system in its native format. It ingests a CWR file from one side, a proprietary CSV from another, a DDEX feed from a third. It normalizes everything into a common internal schema, matches records using identifiers where they exist and fuzzy matching where they don’t, flags conflicts for human review, and pushes clean data back out to each system in whatever format that system requires.
Nobody has to change their internal database. Nobody has to migrate. The translation layer absorbs the incompatibility so the companies don’t have to.
Is a shared metadata API possible? Technically, yes, and it has been for years. The components are proven: schema mapping, entity resolution, API-based data exchange, audit trails. Companies like Zapier, Plaid, and Segment built billion-dollar businesses doing exactly this in other industries. Banking data was arguably messier than music data, and Plaid still built a universal translation layer across thousands of financial institutions that never agreed on anything.
So if the technology exists, why doesn’t the solution?
The real barrier is not technical. It’s human.
Here is the part of the Music Biz session that stuck with me, and the part most panels dance around.
If your job for the past 20 years has been manually matching royalty statements, reconciling registration conflicts, and knowing which contact at which society can fix a stuck payment, then a system that automates data translation is not exciting. It’s threatening. Your institutional knowledge is the workaround for the broken system. Fix the system and your value proposition changes overnight.
I want to be clear: this is not a criticism. It’s a completely rational response. Nobody volunteers to modernize themselves out of a paycheck, and the people slowing these initiatives down are often the same people who kept artists getting paid at all through decades of duct tape and heroics. They earned their skepticism. They’ve watched grand interoperability initiatives launch with fanfare and quietly die more than once.
But rational self-interest at the individual level is producing an irrational outcome at the industry level. Artists wait a year or more for money they earned. Songwriters lose royalties permanently to matching failures. Small companies burn payroll on data janitorial work instead of growth. The people the entire industry exists to serve, the artists, absorb the cost of everyone else’s job security.
The artist-first reframe: modernization creates jobs, it doesn’t just eliminate them
This is the argument I want to put in front of both generations in that room.
The premise that modernization equals job loss assumes the amount of work in music rights administration is fixed. It isn’t. The current system doesn’t just create bad jobs. It prevents good ones from existing, because so much capacity is consumed by translation and cleanup that there’s nothing left for the work that actually grows the pie.
Look at what happened in adjacent industries. When banking data got a translation layer, bank back-office jobs didn’t vanish. They shifted from manual reconciliation to fraud analysis, product development, and customer-facing roles. When marketing automation matured, the people who used to manually build email lists became lifecycle strategists and operations leads. The job titles changed. The employment didn’t disappear. In most cases the new roles paid better than the old ones.
Music rights administration would follow the same pattern. A functioning translation layer doesn’t eliminate the need for people who understand rights data. It eliminates the spreadsheet reformatting and multiplies the value of the judgment. The person who spent 20 years learning why registrations conflict becomes the person who designs the conflict-resolution logic, audits the edge cases, and handles the disputes no algorithm should touch. That knowledge becomes more valuable in a modern system, not less, because it’s finally applied to decisions instead of data entry.
And the new categories of work are easy to imagine: metadata quality analysts, catalog data auditors, integration specialists who onboard companies to the shared layer, royalty experience roles focused on artist communication instead of artist apology. Faster, more accurate payment also means more money circulating to artists, and artists with reliable income invest in teams, marketing, touring, and yes, administration.
The honest version of the pitch to the skeptical veteran is not “your job is safe.” It’s “your job is changing either way, and the version where you help design the new system is better than the version where it gets built around you.” The generation that understands the mess is the only generation qualified to fix it. That’s leverage, not obsolescence.
What would it take to actually get there?
A shared metadata API won’t come from a standards committee. Twenty years of committee efforts have proven that consensus-first approaches stall. It’s more likely to come the way Zapier and Plaid came: a neutral third party builds the translation layer, wins adoption from small and mid-size companies who feel the pain most acutely, and grows until the majors can’t ignore the network effect.
Three things would accelerate it:
Start with the underserved middle. Small publishers, indie labels, and boutique management firms lose the highest percentage of revenue to data friction and have the least invested in legacy systems. They adopt first. The majors follow markets, they don’t lead them.
Make the incumbents partners, not casualties. Any serious effort should recruit veteran rights administrators as designers and auditors of the system, publicly and early. Their fingerprints on the architecture is both better product and better politics.
Anchor everything to artist payment speed. Interoperability is an abstraction. “Songwriters paid in 60 days instead of 400” is a rallying cry. Every design decision, every adoption pitch, every panel discussion should route back to that single measurable outcome, because it’s the one goal no one in the industry can publicly oppose.
The technology stopped being the bottleneck years ago. The question in that Atlanta session, and the question for the whole industry, is whether we can make modernization feel like a promotion instead of a pink slip. If we can, the shared metadata layer isn’t a moonshot. It’s overdue infrastructure.
The artists have waited long enough.
Frequently Asked Questions (FAQ)
Doesn’t DDEX already solve this? DDEX solves data exchange for companies that fully implement it, which is mostly larger players on the recording side. Publishing data, works registration, and the long tail of small companies remain fragmented. A standard with partial adoption still requires translation, which is the gap a middleware layer fills.
Who would own or run a shared metadata API? The most realistic models are a neutral commercial third party (the Plaid model) or an industry-funded utility with independent governance. History suggests the commercial route moves faster, because it doesn’t require consensus before building.
Would this eliminate royalty administration jobs? It would eliminate specific tasks, primarily manual reformatting, rekeying, and reconciliation. The roles evolve toward data quality, conflict resolution, system auditing, and artist-facing service. Comparable transitions in banking and marketing technology grew net employment in those functions.
What can a small music company do right now? Clean your own house first. Standardize your internal data, enforce identifier discipline (ISRCs, ISWCs, IPI numbers), and document your splits at creation, not at registration. Companies with clean internal data will plug into any future translation layer in weeks. Companies with messy data will wait in line for years.
