I recently published a piece about music rights data and the challenges of delivering timely royalty payments to artists on Cherie Hu’s Water & Music publication. I wanted to share a bit of the background research I conducted to write this piece. Here’s an exclusive interview I did with Phil Barry, Founder at Blokur. We discussed how his company addresses one of the toughest issues of the publishing industry regarding royalty payments: data management. If you’re not familiar with these music copyright data issues, I advise that you read the Water & Music piece first for an introduction to these wonders 🙂
Before founding Blokur, Phil produced and co-wrote music for artists signed to BMG, Universal, Warner and EMI; performed in front of millions; founded an indie label and led the development of several innovative music business models and experiences for Radiohead singer Thom Yorke; as well as pioneered the implementation of blockchain and smart contracts for creative rights and royalties with the grammy-winning artist Imogen Heap.
Julie Knibbe from Music Tomorrow: Can you tell us what Blokur is about?
Phil Barry from Blokur: In a nutshell, we work on improving music rights data. From the publishing perspective, we work on the data that goes into royalty systems and make it as accurate as possible, so that people who are making decisions about who to pay can make sure that they’re paying the right person. We make it as complete as possible and reduce gaps or errors as much as we can. Then, on the way out, we make sure that the correct money has been paid to the right holders so that there’s no usage that has failed to be detected by other parts of the system.
If you are a music user, like a TV company or a film company or something like that, you can use Blokur to identify who owns the song, so that you can go and get your license and send a license request for them. We also help bigger music users like digital platforms to perform their responsibilities efficiently and cheaply to get people paid when the music is used on their platform.
JK: Feeding accurate metadata about works and recordings to royalty systems is critical to get paid. Unique identifiers like ISWCs (International Standard Musical Work Code) are supposed to help reconcile what gets in. On your blog, you mention that “only about 30% of the songs in publisher databases have an ISWC.” How do you explain this?
PB: What that means is that even though these identifiers are meant to be unique, they’re not. You might have lots of ISWCs that are by mistake assigned to more than one work. It should theoretically not be possible, but it does happen.
At Blokur, we don’t use the [ISWC] identifier as the definitive answer. It’s just one piece of metadata that might help you to identify the song, along with other pieces of information like the title, alternative titles, songwriters and their alternative names, publishers’ internal IDs, etc. We build a web of relationships surrounding the song, and we use all of that context to identify what is a unique song, or what is a match. In the end, I’m not sure that the idea of a single true definitive identifier practically could exist.
JK: How do you define data authorities? For example, if you receive conflicting information about the same work, how do you decide who has the final say?
PB: We use a hierarchy of data authority. It basically means that the closer you are to the information you describe, the more weight you will get. So if you are talking about yourself, you would have the highest score. If you were talking about somebody else, you would have a lower score.
Most problems are not deliberate. They are due to the difficulty of record-keeping, keeping things up to date. If you can automatically resolve data management issues, you solve most of the problems. Real genuine copyright disputes – in a legal sense – don’t happen so often.
JK: Do you have an idea of how often real disputes happen?
PB: I don’t have actual statistics, but let’s say I’ve never seen one. It’s always a data management problem.
JK: How do you work with publishers to resolve these data management issues?
PB: Before Blokur, lots of companies didn’t have a conflict resolution mechanism at all. Some do, but they use email or the phone to reach out to CMOS. It’s very slow. The idea is that Blokur is more efficient even when publishers have to get involved and do the conflict resolution work. Our system updates many places in one go, so you don’t have to go repeatedly, check your data in this or that country, and update everything.
JK: I bet that before using such conflict resolution tools, most publishers were not even aware of most conflicts.
JK: Given your relationship with publishers, what insights are the most critical for them that you can provide?
PB: On the copyright side, we combine usage with conflict information to say: here is a really valuable song with a problem. This piece of music has a billion views, and you didn’t get paid for it, because somebody failed to match the usage to your composition. So, there’s money missing for you to claim. Once you have such data, there are all kinds of interesting things you can learn. Maybe the titles are not right, or the writer is not linked properly. We are able to do it because we have the data from so many different sources. If you register with one name and someone else with another, then you both know both names. Everyone gets the benefit of this application.
JK: You’re working on resolving conflicts about copyrights. We’ve seen that most problems stem from data management issues, so much so that you can automate resolution. So what’s a good level of accuracy? What do you think is the best you could achieve?
PB: Good question. If everything was perfect and we perfectly did our job, there would be no conflicts left… but this will never happen because the data is not static. Even if it were true today, it would change in the future. The job will never be finished. We measure things like, for all of the songs in digital charts, for example, for how many do we have complete information? How many of them are conflicting? We want those things to be as high as possible. We start and focus on those because they are the songs that are earning the most money overall.
JK: You’re already tracking digital charts to estimate consumption and value to focus your efforts. Do you plan to take it a step further?
PB: Yes, for example, we work on TikTok. We look at the music that’s used on TikTok, how many videos are created for each song, and report that back. So it’s not just based on charts, but on what is actually the usage of this song? We calculate the market share based on the number of videos created for each track (or total streams if you would like to go one step deeper) and the publisher’s share in each of the underlying songs.
We also work on other use cases like YouTube. We can help where there weren’t successful matches by the standard process with content ID. We were able to identify usage of music on YouTube that wasn’t identified before. In the end, we want to do more of that because the more details we have, the more valuable it is to publishers.
JK: Can you share more about your roadmap? What are your next steps?
PB: Blokur has until now been focused on a key foundational building block, which is the right database. The more complete the database is for our publishers, the more coverage we have of all of the errors and problems, the more valuable it is to them. The core has been about scaling all of that up. Music copyright data is complicated in unexpected ways in terms of maintaining the publishing chain along with lists of contracts.
We’re now more focused on issues relating to usage, like what I mentioned about TikTok and Youtube. We are building that up in terms of scalability and user experience to be more efficient.
JK: The industry overall has been fantasizing about getting a centralized source of truth about rights data. I think people now have accepted that such data will never be perfect. However, which changes do you wish to see in terms of royalties and data?
PB: In reality, I think it’s better if we aim towards an ecosystem where data can be shared efficiently, so that it is possible to take the best bits from the best places and combine it together. The value is not really in the raw data itself. It’s in the way that it’s matched and the insights you gain from it. If we open the borders between the different organizations, tell them what the data is and allow us to compare… It will be a big step.
One other point is, I don’t think immediate payments is necessarily the goal to aim for. If you could get paid the same month, that would be already fantastic. That’s what most people in the world manage with their jobs. From the musician’s point of view, they don’t want to be spending their time thinking about money. What they want is to have the predictability of the money coming every month or every two weeks. That solves the problem for them, instead of getting paid one year a lot, then nothing the next year.
That’s the difficulty in my own experience as a musician. Sometimes I would open a royalty statement, and I would be like, “wow, I never knew I was going to get paid this much for this”. And then the next time it would say zero, and I’d be like, “what does it really mean?”. It’s this kind of unpredictability. That’s partly because you’re getting paid in these long time periods. Recently I’ve been paid money from something from 2010! Music publishers might even pay you once a year now, so once every month will be a lot better.
There is competition for music publishers to offer better service, but I don’t think they have to pay you the same day for it to make an impact. You know, you’re a product person. Overall, it’s about the value for the customer, not the feature. The value is being able to predict when I’m going to pay the rent, whether I’m going to be able to get a mortgage, whether I’m going to be able to do those things, not the feature of being paid every day.
JK: Before wrapping up, I wanted to talk about recording data. Do you think that recording and publishing data should be coming to one place?
The digital world and streaming in particular made it essential for publishers to know about the recordings of their songs. It’s such a lot of volume, and it’s all about tracking all these different recordings. A big part of the work is about matching between the composition and the recording. However, it doesn’t mean that if you’re doing lots of matching, that you’ll end up solving all the problems.
Blokus is a publishing data platform. There are millions and millions of recordings in the database that are linked to the works. It’s linked to the work, rather than separately as a recording. So the data actually is coming to be in one place. I think that the data has to be in one place for it to be efficient. I don’t necessarily think that means that there will be one database for both parts of the industry, but at least publishing needs to have recording data alongside it. No doubt about that.
We have had recently more inquiries from companies to have both sides, because they want to have a clear idea of what they could control independently. That gives more freedom in terms of licensing for novel applications of music. If you could say, I’ve got a million pieces of music in my database, but there’s a hundred thousand I control myself entirely (recordings and composition), and that’s enough for me to get started with some initiative. So, yeah, let see what happens!