CMU Trends

Trends: Global music data and blockchain – the story so far

By | Published on Friday 22 September 2017

Data has been a frequent topic of debate within the music community in recent years. Though data is a term that covers a multitude of things, so in fact the music industry’s data debate is actually a number of different discussions.

Firstly there is the fan data to be pulled from the streaming platforms, social media, ticketing services and direct-to-fan channels. Who has access to this data and what should artists and their business partners do with it to inform their business, creative and marketing decisions?

Second there is financial data, what monies have been generated by an artist’s different activities, and how is that money being distributed between the different stakeholders in the music industry?

In the streaming domain in particular, decent royalty information is too often lacking, in that artists know what money they are receiving, but they don’t always know how that figure was calculated. Though at the same time, with streaming being a micro-payments business, artists and their managers would be swamped if they got a line of data for every single stream.

Third, there is the impact that big data and so called machine learning could have on the future of the music industry, in terms of how artists interact with fans, how royalties are calculated and paid, how music is recognised and curated, and even how music is made.

And fourth there is the debate around music rights data. Pretty much everyone agrees that the music industry needs to do a better job of letting the world know who wrote each song, who performed on each record, and which artists, songwriters, labels, distributors, publishers and collecting societies have control over and/or a stake in any one piece of music.

Despite much debate on the music rights data issue, and plenty of initiatives attempting to address it, problems remain. Indeed, at the heart of developing news stories like Spotify’s mechanical royalties dispute in the US – and the claim that streaming services are ignoring the moral rights of songwriters by failing to credit them – sits the big unresolved music rights data problem.

As Music 4.5 puts the spotlight back on data in music this month, we review that latter challenge, and ask four of the people presenting at next week’s event about persisting problems and possible solutions.

ORIGINS OF THE PROBLEM
In most countries, copyright is an automatic right, which means no registration is involved. So, from a music perspective, as soon as song is written a copyright exists in it, and likewise as soon as a recording is made. There is no form to fill out, no registration fees to be paid.

The law then tells us who the default or presumed owner of the copyright is. Copyrights can be co-owned. With songs, the default owner is the writer, and where writers collaborate they co-own the resulting copyright. How the copyright is shared out is for the collaborators to decide.

Copyright can also usually be transferred, so a default owner can assign the copyright in current and future works to another party. Even where actual assignment doesn’t take place, a default owner may license some or all of their rights to another party, who will likely act as if they are the copyright owner until the licence expires.

When copyrights are transferred to other parties, it may be for life of copyright or for a set period of time, and may only apply in some countries. Plus different elements of the copyright – such as the ‘reproduction rights’ and the ‘performing rights’ – might be transferred to different entities. In the UK, songwriters commonly assign the reproduction (or ‘mechanical’) rights in their songs to a publisher, but the performing rights to their collecting society, ie PRS.

If a third party wants to make use of a song or recording, they must get permission from the copyright owner, or owners. Step one of that process is working out who owns the relevant rights. If it’s an existing track being used, there will be both song rights and recording rights to be licensed, which will likely be owned by different entities. The song rights may well be co-owned by multiple entities. And who you need to talk to may well vary depending on which elements of the copyright you need to exploit, what exactly you plan to do with the music, and what countries you’re operating in.

So music licensing can be complex. Though perhaps the biggest challenge is the working out who you need to a deal with in the first place. Because there is no statutory copyright register in most countries – copyright being an automatic right – and the music industry doesn’t provide a one-stop publicly-accessible database listing every song and every recording, identifying what song is contained in what recording, and listing who owns and controls each element of each copyright in each territory.

The growth of global digital music services offering tens of millions of tracks made the lack of decent music rights data an ever more pressing problem. With that in mind, the music publishing sector famously attempted to build a Global Repertoire Database, hiring Deloitte to do the groundwork. But the GRD project – which would initially have only containing song rights information, not recording rights – ultimately collapsed.

However, there are plenty of other initiatives under way to try to address the music rights data problem. Some are being led by the music industry’s collecting societies, which manage song and recording rights in collective licensing scenarios, and who are usually sitting on the best music rights databases that currently exist. Others are being led by commercial entities and start-ups, some within and some outside the music community.

Although, as the aforementioned Spotify mechanicals litigation and the controversy around the lack of songwriter credits in the streaming domain both prove, the music rights data problem still persists, and is having a tangible negative impact on the music industry, and especially the increasingly key streaming music business.

SHARING THE DATA
While there is no one-stop publicly-accessible music rights database, that’s not to say there aren’t any music rights databases out there. In fact there are lots and lots of music rights databases, which is partly the problem. No one database is complete. Many databases contain just song data or just recording data, and therefore don’t match tracks to songs. There is conflicting information between different databases. And most of the decent databases are private.

There is an argument that a good step forward to cracking the music rights data problem would be for everyone sitting on an existing database to make the top line basic copyright ownership information public, so that data-clever start-ups could grab that information, clean it up, and then build a bunch of decent music data tools from it.

Robert Kaye from MusicBrainz – the publicly accessible “open music encyclopaedia” that already aggregates, through crowd-sourcing, the music metadata that is public domain – reckons that such data-sharing by record labels and music publishers could help address a number of the issues – and they wouldn’t actually have to give up control of their databases to do this.

“If every label and publisher could set up an API that gives the public access to their repertoire data, copyright status, and so on, this would go a long way towards building a system where no one has to give up their cherished control over their own data”, he says. “People and companies wishing to license content could easily discover who owns the copyright to a particular song or recording; this API could also provide licensing terms for said content in an effort to streamline the whole process”.

For such data-sharing to work, there would need to be some industry oversight. “Conflicting copyright claims and policing the actors in this system would need to be carefully addressed”, Kaye concedes. “But I can see a non-profit ombudsman providing API standards, reference implementations and conflict resolution services. Bad actors carrying out bad acts could be financially punished in order fund the operation of the non-profit”.

In terms of logistics, Kaye reckons that “this system could be implemented quite cheaply as long as you engaged the right people – more the likes of [MusicBrainz operator] MetaBrainz and less the Deloittes of the world – without entities giving up control over their data”. Though, Kaye adds, while the potential solution could be relatively simple and cost efficient, that doesn’t make him optimistic that any solution is incoming. “Any sort of broad co-operation like this is still wishful thinking in the music industry”.

Fewer players would need to be involved in such a data-sharing enterprise if it was led by the collecting societies rather than individual labels and publishers. But Kaye doesn’t see any real appetite in the collecting society community to instigate widespread data-sharing in the public domain.

“Making data public forces people to do two things – first, do their job, and second, do their job correctly”, he says. “Societies love hiding in opacity. They love putting money into the pockets of the people who buy them loads of sushi and take them out for drinks and all that”.

Phil Barry of Blokur, a start-up seeking to help fix the music rights data problem, is more optimistic about persuading the music industry to get better at sharing its data, though adds that another issue alongside data availability is data standards. Indeed he reckons the latter might be the bigger issue.

“I think the problem is less the availability of data and more the challenge of piecing together all the different parts of the data puzzle that are out there and reconciling conflicts between different databases. The challenge there is a web of different standards and systems that don’t speak to each other and don’t update synchronously”.

While, as Kaye reckons, collecting societies may be resistant to sharing too much data because it would put the spotlight on inefficiencies in the collective licensing system, another issue may be that societies worry that – if there was a decent one-stop publicly-accessible music rights data platform out there – the power of the collective may be reduced.

One of the reasons why collective licensing has always made so much sense, especially with song rights, is that licensees can pass on the hassle of working out who controls the songs they’ve used to the society. Which may or may not then do a good job of getting the money to the right writers and publishers.

However, Barry reckons that the societies shouldn’t worry about there being more music rights data in the public domain. “Bringing the necessary data out into the open” he says, “is really a case of aligning incentives so that people can see the value they get from participation”.

“Until or unless there is a viable alternative to collective licensing, societies will continue to be an important part of our industry, and it is important to remember that they are mandated to defend the interests of songwriters and publishers”, he adds. “We work with societies to help them ensure they are very much part of the solution”.

YOU GET OUT WHAT YOU PUT IN
Beyond the challenge of getting the right data to the right people, there is the question of whether or not the data controlled by different labels, publishers and collecting societies is actually any good.

For legacy data, that means a big data cleaning operation, which some societies have already begun by comparing and integrating their respective databases. Though with more new songs and recordings being pushed out each week than ever before, there is also the task of ensuring that data associated with new works is correct and complete.

Given that it is the actual writers, performers and record producers who know best what happened in the studio and rehearsal room, the music creative should be participating in this exercise, especially with song rights. Which is what start-up Auddly, backed by top pop music makers Max Martin and Björn Ulvaeus, aims to do, by providing the tools that make it easier for songwriters to discuss, agree and log how the copyright in the songs they write is being shared out.

Songwriter and journalist Helienne Lindvall has an industry and songwriter relations role with the company. “The reason I joined Auddly is because it presented a solution to one of the main reasons we have a problem with rights data, ie ‘garbage in, garbage out’”, she says. “As a songwriter myself, I know first hand how crap we are at logging what happens in the studio. Matter of fact, I’d say that, for most of my career I’ve simply assumed that the publisher took care of it all, until I realised that they could only go by what I was telling them – which, to be honest, wasn’t much”.

For established writers, a manager and publisher will likely still be involved in managing copyright information, though if writers can be persuaded to take an active interest in that process, everything should be more efficient. And, Lindvall says, that means “we need to educate music creators” because “the level of knowledge of rights among songwriters is extremely low”.

“I usually start my presentations by showing my fellow musicians the four codes needed to get paid”, she says. Those are the Interested Parties Information (IPI) and International Performer Number (IPN), as well as the International Standard Recording Code (ISRC) allocated to each track and the International Standard Musical Work Code (ISWC) allocated to each song. “The vast majority of them don’t recognise any of them”, says Lindvall. “Let alone know that they have an IPI and/or IPN number”.

“Secondly, they need to understand that there is a metadata problem”, she goes on. “And that they play a part in it. When we show them that logging this info doesn’t have to take up more than a minute or two on your smartphone, the penny drops”.

“Their and their co-writers’ publishers and managers can access the info instantly without having to chase them – which, in turn, leads to earlier registrations – that are clean, right from the start. No multiple ISWCs for the same song, no conflicting data, everything in one place from the people who know the truth – the creators”.

Getting writers to be persistently involved requires making the process as simple as possible. Which is what Auddly is aiming to achieve. “It’s vital to make it easy for music creators to log rights data without interrupting the creative process”, Lindvall reckons. “That can be achieved by using mobile phone technology, as well as facilitating capturing metadata in the audio stems while recording, using software such as Logic and ProTools”.

While the industry needs to provide the tools to help artists and songwriters in this domain, music creatives need to take an active interest and role, Lindvall reckons. She concludes: “As a songwriter, I don’t think a service should be allowed to use my song without knowing that I wrote it – and therefore who to pay. Would it be acceptable for Tesco to sell a sandwich without knowing where they got the ingredients or even what’s in it?”

“However”, she adds, “I’m also of the belief that if you want something done you need to do it yourself, if possible. And we, as songwriters, need to take some responsibility here. All this rights information collection doesn’t happen magically – we need to make sure the correct info is there from the start, and that it is accessible to the streaming companies. We need to take away any excuse a service might use for it not to be displayed”.

WHAT ABOUT THE BLOCKCHAIN?
For much of 2016, the music rights data debate within the music community got very distracted by this thing called the blockchain which, some people argued, could play a key role in solving the industry’s big data problem. Though, it should be noted, the blockchain – still best known as the technology that powers cryptocurrency BitCoin – is simply a potential method for distributing and storing music rights data. The data being shared still has to come from somewhere.

Vaughn McKenzie is the founder of JAAK, a company that seeks to use blockchain technologies to simplify the music licensing process. “At a high level, a blockchain includes a few things”, he explains. Those are: “A software that connects a computer to a peer-to-peer network; a file stored on every computer connected to that network and software that keeps the computer in sync with the latest version of that file; software for broadcasting new transactions to the network to be added to the file; software that decides who gets to add all of the new transactions from the network – the block – to the file containing all previous transactions – the blockchain”.

In some ways, it feels like the big buzz around blockchain in the music community that started two years ago was something of a distraction. Though plenty of people still advocate that the technology has a role to play. Barry’s company Blokur is also employing the blockchain in its music data solution. “When blockchain technology first entered the discussion on music data, the technology was still very immature and the ideas were not fully developed”, he concedes.

But, he adds, “since then both parts of the puzzle have progressed enormously. I remain convinced that blockchain technology can form the backbone of a new set of rights infrastructure for data, licensing and payments, due to unique characteristics like the ability to automate processes outside the boundaries of traditional organisations”.

McKenzie, while being a champion of the blockchain, stresses that the technology is part of the solution, rather than being some kind of panacea. “Blockchain is a state machine”, he says. “It’s great for ensuring that everyone connected has the current state of information, without the need to trust any one party”.

“The data problem in the music industry has many facets, including availability, accuracy, authority, provenance, inconsistency, to mention a few. Some of these are addressable by blockchain, some aren’t. Blockchain’s role in the music industry will depend entirely on adoption – we have a view on this, but there are others. The outcome remains to be seen”.

At the CMU Insights conference on all things data at The Great Escape in 2016, Kaye was one of the interviewees playing down the excitement around blockchain. Why? “Because politics in the industry and this deep seated fear of losing any bit of control or opacity will likely prevent the blockchain from working”.

“The technology itself is sound”, he adds. “I quite like it and the thinking that is coming out of it, even if it is mostly hype so far, beyond Bitcoin. Accepting distributed systems and transparency is a really big and important step for the world”.

There are different blockchains, and Kaye name-checks Etherum – the blockchain being used by JAAK – as being “quite interesting since it doesn’t have some of the same limitations that BitCoin has”. But, for him, picking the right technology is still not the main challenge. Because no blockchain comes with “magical solutions for fixing the political issues in the industry”.



READ MORE ABOUT: |