Decrypting the Metadata Landscape

If you follow this blog, you know that FMC spends a lot of time thinking about metadata, a shorthand term that can mean a lot of things including the information about who wrote a song, or who played on a recording. We’ve looked at the problems from different angles, examined the wide range of possible solutions, and attended metadata conferences. This month’s Future of Music Policy Summit will have some special sessions that tackle metadata, so we asked FMC’s director of programs Jean Cook six questions about the topic.

1. What does the term “metadata” mean? 

In the broadest sense, metadata is the information that identifies and describes a piece of recorded music. Sometimes metadata is text – the songwriter’s name, the album title, for example; sometimes it is numbers – the UPC barcode or the ISRC code. This information travels with your music as it’s delivered to services, and powers discovery, attribution and proper monetization.

2. What does metadata have to do with discovery or the flow of money?

A lot. Think about it – there are millions of sound recordings that have been released commercially in some way over the past 50 years. Digital music services that have sprung up in the past 15 years have tried very hard to license as much of this music as possible to make their online catalogs robust and complete.  That means there are millions of pieces of music information that these services need to display or make available in some way. But screen real estate is precious, so it means that services have made some tough choices about what information is displayable, and what information is searchable. Clearly, all the services display artist name, song title, album title. But what about key side players, or composers, or producers? This additional information – even if it’s not displayed or searchable – is critical for the music’s creators to get proper attribution and, in some cases, proper payment.  

3. Why do metadata problems happen?

When services are missing important information about a recording (like who played on a record or who wrote a song, for example), or when an artist doesn’t get paid the money that is due to them for usage of their composition or their recordings, it’s probably because someone along the chain of “people who get music from artists to fans” and “people who get money from fans to artists” is missing the information they need. 

4. Who is responsible when the information isn’t there?

Let’s take a closer look at those two chains of people between artists and fans. 

First, here is one oversimplified* example of how information about who played on a record and who wrote a song might flow. It starts with the composers, the performers, the session and background musicians, whose information is collected and sent to the record labels, publishers, and unions.

The information flows from the creators, to the labels and publishers, to the aggregators, to places like All Music Guide, and to the music services like iTunes and Spotify. Eventually some of that information is also shared with fans by the music services. 

All of the entities in the chain are responsible in some way for a piece of the information. But it’s a delicate chain and there is a lot of room for error. If information doesn’t show up where it’s supposed to be, it won’t be passed along and often no one will try and fill the gaps. 

What does that mean? Well, for example, lets say for some reason we live in an insane alternate universe where a label decides the featured artist for the track “The Girl is Mine” from Thriller is only Michael Jackson, that is what information they will pass on to the services. But Michael Jackson is not the only performer on the track; it’s a duet with Paul McCartney. If the services want to include Paul McCartney’s name when they present “The Girl is Mine” but they don’t get that information from the label, they will have to do the research to find that information themselves. What is the likelihood they will do it if it’s one of a million tracks they get? Hard to say. And that is for one of the most visible, influential, and popular musicians on the planet.

While I think everyone agrees that theoretically fans should be able to know who played on a record or who wrote a song, it’s hard to get the various entities along the chains to agree whose responsibility it is when information isn’t where it’s supposed to be.

As mentioned before, there are two chains of concern. The one described above shows how information gets from musicians to fans. The next chain describes how payments get from fans to musicians. 

Here is another oversimplified* flow chart that shows how the money trickles back from the fans to the artists. Along the way, each entity that touches the money takes a piece of the income. Sometimes it’s a percentage of the gross income in the case of the services, a distributor’s fee in the case of the aggregators, a percentage of the royalties in the case of the labels and publishers or producers, or a modest admin fee for the PROs and unions.

For every licensed use of music there is a payment generated, and this payment has to go through each link of this chain before it gets to the composers and performers.

Along this chain, each entity needs to know how much of a cut they take, and the next stakeholder in the chain to pass the remainder of the money on to. Services need to know what labels and publishers to pay, whether directly or via an aggregator. Record labels need to know who the songs’ composers are to distribute mechanical royalties. ASCAP, BMI and SESAC need to know the composers of any particular song, and the songwriters’ royalties splits, to be able to distribute public performance royalties. The unions need to know what background musicians played on a session to send them Sound Recording Special Payments Fund or AFTRA Contingency Fund payments to qualified players for record sales and streams.

5. Are certain links in the information chains weaker than others?

You bet. The development of these data chains was ultimately driven by business and industry, which invested a tremendous amount of money and time to build them. So their priorities (usually profit and market share) are reflected in how well each link of the chain was built. Let’s look at the oversimplified* payment chain as an example.

The strongest links in the chain are the ones where the most money flows. Generally, the closer you get to the fans or the users or the places where the money originates, the stronger the links are and the fewer the errors. Services spend millions and millions of dollars to optimize and idiot-proof the link between them and their customers (A) in the payment chain. Their business depends on it. Similarly you’ll see a tremendous amount of effort, time and money on the part of labels and publishers to ensure that the maximum amount of money will flow from the services to the copyright owners (B).

The closer you get to the artists, (C) the more likely you are going to find missing information or errors. That’s because artists don’t have the same kind of leverage as labels and services do to demand better or more transparent payment systems. They live with whatever system is convenient for those with more leverage: the labels, publishers, and the services.

And there is also a hierarchy between the two chains for sure. While there has been tremendous investment in figuring out how to make and split up the money generated from music uses – a business critical issue – making sure the performer and composer information makes it to the fans and users is often considered a “nice to have” issue, and not business critical. So that chain is not nearly as well developed.

6. So what exactly do people mean when they talk about “metadata problems”?

If you pick at random any two people in the music industry railing about metadata issues, chances are high they may not be talking about the same thing. That’s because each person’s perspective depends on which link on which chain they are concerned with, and what part of the music community they inhabit. Metadata problems come in all different shapes and sizes. To illustrate this, consider these eight different types of metadata challenges you might encounter along the two oversimplified* information chains.

Each of these challenges exists because there is no single right answer to the questions listed above. The answers are different depending on the particulars of each situation. And in the case of Challenges 1, 2, 5 and 6 the number of different “right” answers increases exponentially because you are dealing with an incredibly diverse artist population that includes everyone from Stevie Wonder to the Los Angeles Chamber Players, from Amanda Palmer to Gillian Welch – each of whom have a different way of addressing these issues depending on if they have a manager, if they own their own copyrights, if they are a member of a union, etc.

The complexity of the information gathering and verification process can be incredibly frustrating for artists who find themselves in the position of trying to figure out how to fix missing or incorrect information about who played on a record or wrote a song. Equally overwhelming can be where to start if you want to make sure important information gets entered in the right way the first time.

That is the motivation behind a special two-part lunch session as part of the Future of Music Policy Summit called Metadata for Musicians: a hands-on workshop about the data attached to your music that powers discovery, attribution and royalties on October 27 and 28. Designed for artists and managers, these sessions will explain how this data gets used within the various parts of the information and payment chain, and some basic things artists can do to ensure their information is correct and accurate.

The Summit will also include a session called “Life with the Devil: Who Minds the Details? Artist Metadata and Royalty Payments” on Monday October 27, where artists, managers, and other interested parties can hear about the structures that are being built to make reporting, verification and attribution easier for artists in the future.

Interested in joining us in DC for the Summit? If you are a working musician you can apply for a scholarship to attend the conference

If you’re interested in learning more about the metadata issue, there is a ton of interesting work being done by MusicBiz  and DDEX, and anyone from the industry is welcome to get involved.


* To illustrate these general concepts, we’ve simplified some details of information flow, and grouped some things together that may not normally be grouped together. That’s because when it comes to specific recordings, the music industry is complex, so there are many different ways the information flow for a particular recording could go. Most of the time it will not include all of these players, and often it will be more than three or four steps. For a more detailed and precise representation of how the money flows check out Music and How the Money Flows.

Metadata image via Shutterstock.