It’s a crisis that has left, by some estimations, billions on the table unpaid to musicians.
a musician signed to a major indie label told me they were owed up to $40,000 in song royalties they would never be able to collect. It wasn’t that they had missed out on payments for a single song — it was that they had missed out on payments for 70 songs, going back at least six years.
The problem, they said, was metadata. In the music world, metadata most commonly refers to the song credits you see on services like Spotify or Apple Music, but it also includes all the underlying information tied to a released song or album, including titles, songwriter and producer names, the publisher(s), the record label, and more. That information needs to be synchronized across all kinds of industry databases to make sure that when you play a song, the right people are identified and paid. And often, they aren’t.
Metadata sounds like one of the smallest, most boring things in music. But as it turns out, it’s one of the most important, complex, and broken, leaving many musicians unable to get paid for their work. “Every second that goes by and it’s not fixed, I’m dripping pennies,” said the musician, who asked to remain anonymous because of “the repercussions of even mentioning that this type of thing happens.”
Entering the correct information about a song sounds like it should be easy enough, but metadata problems have plagued the music industry for decades. Not only are there no standards for how music metadata is collected or displayed, there’s no need to verify the accuracy of a song’s metadata before it gets released, and there’s no one place where music metadata is stored. Instead, fractions of that data is kept in hundreds of different places across the world.
As a result, the problem is way bigger than a name being misspelled when you click a song’s credits on Spotify. Missing, bad, or inconsistent song metadata is a crisis that has left, by some estimations, billions on the table that never gets paid to the artists who earned that money. And as the amount of music created and consumed continues to increase at a faster pace, it’s only going to get messier.
critical that metadata is distributed and entered accurately, not just for a song or album’s discoverability, but because metadata helps direct money to all the folks who made that music when a song is played, purchased, or licensed. Documenting everyone’s work is also important because, “That attribution could be how someone gets their next gig,” says Joshua Jackson, who leads business development for Jaxsta, an Australian company that authenticates music information.
There are multiple ways this process can go awry. The first is that, because there’s no standardized format for metadata, information often gets discarded or entered incorrectly as it’s written down or moved between people and databases.
A label’s database is likely different from Spotify’s database, which is likely different from the databases of critical collection societies, like ASCAP and BMI, which pay public performance royalties to musicians. “Part of the problem is the fields everyone has chosen to write into their software to populate these credits are all different,” says entertainment lawyer Jeff Becker of Swanson, Martin & Bell. “So if a credit is sent to a database that says ‘Pro Tools engineer,’ but that database doesn’t have that field, they either choose to change it or ignore it altogether. Typically they ignore it, and that credit has nowhere to go.”
Each database has its own set of rules. If Ariana Grande, Nicki Minaj, and Jessie J collaborated on a new track, and it was delivered to Apple Music with all of their names in the same artist field, that would cause what Apple Music and Spotify call a “compound artist error.” Entering an artist’s name as “last name, first name” would also result in a rejection. There are ways to embed metadata in a song file to ensure everything travels together, but distributors generally request that it be removed since it can cause “issues with the upload.”
The second big problem is that the information being entered in the first place is frequently wrong. A song can pass through multiple songwriters, producers, and engineers before it gets released by an artist, and every new contributor adds the potential to screw things up. The longer the chain of custody for the data, the greater chance a portion of it will be incorrect. A songwriter could fat-finger a name inside one of these databases, or a producer who briefly worked on the track could be left out, or a faulty merge between two databases could cause a technical error that erases information.
Even on one song, metadata can get complicated in ways you might not expect. In a guest post for HypeBot, Annie Lin, senior corporate counsel at Twitch, uses Katy Perry’s “Firework” to show how messy a song’s data can be. Capitol Records owns the recording for “Firework,” but five different songwriters with five different music publishers own percentages of the composition rights, and all their information needs to be included in the metadata so they can get credited and paid.
Having this many people working on one track is not uncommon, says Niclas Molinder, founder of music metadata company Auddly (now Session). In 2016, the average hit song had over four songwriters and six publishers. That creates a lot of opportunity for metadata to be submitted incorrectly. And if someone’s credit is missing, spelled wrong, or doesn’t match a streaming platform’s style guide, that can muck up payments for everyone involved. All these little errors add up. It’s estimated that as much as 25 percent of royalty payments aren’t paid to publishers at all, or are paid to the wrong entity.
“You may get your data correct in your database,” Molinder says, “but if you don’t get the others’ 100 percent correct too, and if they don’t get yours, no one gets paid.”
In an ideal world, once a song is finished, the metadata would be crafted by the artist or the artist’s producer, and they would submit that data to the record label, distributor, or publisher(s) involved for verification and distribution. In reality, the process is frequently more rushed and haphazard — artists and labels hurry the process along in order to get songs out, and metadata is frequently cleaned up later as mistakes are noticed. ”A lot of these credits and negotiations don’t happen on a single piece of paper, and also happen after the fact,” says Joe Conyers III, co-founder of digital rights management platform Songtrust.
It’s possible to correct metadata errors afterward, but that’s reliant on someone catching that error and then correcting it in every database where it appears. Even if it does get fixed, that doesn’t mean an artist gets all the payments they’re due — every company and collection society has different rules about how long they hold on to unclaimed royalties. The musician who was owed $40,000 missed out because a glitch between two databases removed many of his credits. It wasn’t the musician’s fault, but too much time had gone by before anyone noticed. The companies involved declined to pay him.
“We take it for granted that we can look up movie or TV credits on IMDb and see everything, down to production assistants,” says Jackson, who recently hosted a standing-room-only panel on metadata at the Music Biz conference in Nashville. “But the changes to music metadata and the standards are so slow.”
a centralized database and set standards for music metadata — Jackson’s idea of an IMDb for music — sounds like a straightforward goal, but getting there has stumped many of music’s largest and most powerful entities for decades. There are many reasons for this, but the tectonic shift to streaming is a major contributor. “There was not only an explosion in the number of releases, but the unbundling of the album,” says Vickie Nauman, consultant for music tech firm CrossBorderWorks. “We went from 100,000 physical albums released in a year to 25,000 digital songs uploaded a day to the streaming services.”
Additionally, songs are now being consumed and monetized in many different ways that weren’t available just decades ago. “If you think back to when people primarily bought CDs, the only version of a major song that mattered was the major song itself,” says Simon Dennett, chief product officer at Kobalt. Today, a major hit could have hundreds of different versions, like remixes, covers, sample packs, YouTube lyric videos, recordings in other languages, and more, all of which can, in total, generate “trillions and trillions of transactions” that each bring in fractions of a cent. “The volume of data that now has to be managed has unfolded into a massive problem,” Dennett says.
Not only is there way more content to catalog, music rights are very fragmented to begin with, and so slices of a song’s metadata are often kept across a variety of databases. Labels, publishers, collection societies, and others all maintain their own databases, none of which come close to having all of the information about all the works that exist in the music industry. (To see how truly complicated music data is, here’s a horrifying flow chart from The Music Maze and an explainer from Sonicbids on how to track down song ownership, which ends with “consider paying for research.”)
The creation of a global centralized database for song metadata has been attempted multiple times, but has always ended in failure. Among the numerous reasons: in-fighting between different arms of the music industry, international governance challenges, reluctance to share information, and funding issues. There are other, more practical roadblocks as well, like varying languages, differing copyright laws, and music industry cultures and traditions across the globe, which are often at odds with each other.
There isn’t much agreement on if any particular arm of the music industry should lead the way or be responsible for fixing music metadata. Some think digital music distribution companies like TuneCore or DistroKid could do more to educate artists, as it’s often an artist’s only touchpoint before their music is live on streaming platforms. Others think the streaming platforms themselves could set an example for better metadata by displaying more credits, which would encourage everyone involved to make sure the data is right. Some, like Jackson, suggest educating songwriters and producers to keep metadata records at the point of creation. “I imagine in the long term that’s only going to make all our jobs a lot easier, when we’re getting this [metadata] from the source as early as possible,” Jackson says.
But a lot of artists don’t even know they should care about metadata, or that possible metadata issues could be affecting their paychecks, because royalties are so complicated. One Grammy-nominated artist I talked to said, “Honestly, I wouldn’t even know where to look to find out.” Lots of startups are trying to make artists more aware of metadata, but it’s an uphill battle. Splits, a free mobile app, lets artists create a digital agreement that manages a song’s collaborators and their percentages of ownership. There’s also Creator Credits, a technology that works within music production software Pro Tools to embed song credits within the Pro Tools files themselves.
What everyone does agree on is that while things are starting to get slightly better, there’s a long way to go. “I remember putting things out on TuneCore, and it didn’t ask you for any metadata. Maybe a song title and that’s it,” says Doug Mitchell, director of customer success at music tech firm Exactuals. “Now it asks you for more info like genre. As the stores are displaying more metadata, then [TuneCore] asks for that information. That’s a start.”
Although the idea of crafting centralized and standardized metadata is daunting, many say it’s not something to give up on. Aside from cleaning up record-keeping errors, it would help prevent other musicians from “dripping pennies,” and connect them with the money they’re due. “The process of taking hugely dispersed geographic data, hugely dispersed ownership data, and hugely erratic data quality, and pushing that together into a coherent aggregated global view is a challenging, but incredibly noble mission,” says Dennett. Conyers III puts it even simpler: “It’s a good dream.”
Originally published on The Verge, 29 May 2019.