2016 was supposed to be the year when Spotify reached 100m active users, en route to an IPO.
The former will happen soon while the latter may still be on the cards this year. But 2016 is now also the year when Spotify faces its biggest existential threat yet: a pair of class-action lawsuits in the US focusing on songwriter royalties
The lawsuits against Spotify were filed by two musicians, David Lowery and Melissa Ferrick, either side of Christmas, with media coverage splashing on the potential damages of $150m and $200m respectively.
It’s important to understand exactly what their claims are about, and especially important to understand that the issues go well beyond Spotify.
The first key point: these lawsuits are not just about unpaid royalties for songs streamed on Spotify. That’s part of it, but the issue is also over the licensing of those songs – or at least one of the two licences on the composition (songwriting) side.
For many people, this story’s public life began in October 2015 with a dispute between Spotify and independent music company Victory Records over mechanical royalties. However, it was actually Lowery who first opened the can of worms in a post on his blog The Trichordist in July.
“The past two years I’ve been trying to figure out how it is that Spotify has legally made available many of the songs that I have published under Camper Van Beethoven Music and Bicycle Spaniard Music,” wrote Lowery.
“In order to make my songs available on their service in the US, Spotify must enter into a direct license with my companies or an assigned agent. OR they must serve an NOI (notice of intent) to take advantage of the statutory compulsory license.
“After two years I find no evidence that they have properly licensed most of the songs that are currently available on the service […] Further I can find no evidence that they paid the US ‘mechanical streaming’ royalty to my companies or my agents.”
Three months on, the Spotify/Victory row brought the issue back into public view. Spotify removed Victory’s catalogue after royalty collection agency Audiam claimed that an audit of the label’s Spotify statements had identified 53m streams for which mechanicals had not been paid to its publishing arm, Another Victory.
In November, Victory’s catalogue was restored to Spotify after the two parties reached what Victory founder Tony Brummel described as “a good resolution that works for everybody”.
But the following month, Lowery’s class action lawsuit landed, alleging that Spotify had used his compositions “without mechanical licenses in an egregious, continuous and ongoing campaign of deliberate copyright infringement”.
Anyone familiar with the intricacies of US copyright laws can skip this next section: with these two publishing royalty streams being licensed separately in the US, these cases are not focusing on the performing right royalties collected from streaming services by bodies like BMI and ASCAP.
They are purely about the mechanicals, which are either paid directly to publishers, or via intermediaries like the Harry Fox Agency (HFA).
That’s one of the key differences in the US between music download stores and interactive streaming services. For stores like iTunes, the mechanical royalties for any track are paid to the label or distributor that uploaded it to the system; they are responsible for “passing through” the money to the relevant publishing rightsholder(s).
In the interactive streaming world – and we’re not talking about webcasters like Pandora or DMCA-protected YouTube here – that responsibility falls on the streaming service.
It has to identify the songwriters and their publishers for every song in its catalogue, then either agree a direct licensing deal; or send the “notice of intent” referred to by Lowery to secure a compulsory licence – the publisher or songwriter cannot decline.
In both cases, a payment relationship is established: the streaming service knows where to send its royalty cheque.
A third option, also referred to by Lowery in his July blog post, is that when the ownership of a composition is unknown, the streaming service can file a special notice with the US Copyright Office of its intent to use the work, but hold royalties until they identify the rightsholders.
Both sides in the current dispute agree on the central problem. For songs outside its direct licensing deals with publishers in the US, Spotify does not have the necessary information – the songwriter names, their share in the case of works with multiple authors, and contact details for their publishing representatives.
Without those details, a streaming service cannot negotiate a direct licence, nor can it send a notice of intent for a compulsory licence directly to the rightsholder – and it certainly can’t send a royalty cheque.
Spotify itself has admitted this problem, in testimony to a US Copyright Office licensing consultation in 2014 written by the company’s (then) head of licensing business affairs, James Duffett-Smith, who has since been promoted to head of publisher relations.
“In some cases, Spotify will be licensed for performing, but not mechanical rights in the same composition,” wrote Duffett-Smith. Note, that if the lawsuits get to court, Spotify’s defence may involve pointing to the mechanical rights as a statutory licence, making this a ‘paperwork’ problem rather than a licensing problem.
Critics will surely argue back that without the paperwork, there is a licensing problem.
What is the scale of this problem? Speculation varies. In October, Audiam told Billboard that of the 3,245 recordings that Another Victory has a share in, some 2,183 had not generated mechanical royalty payments for the company, despite 53m streams of those songs. That’s 67.3% of its catalogue, but this is one company.
National Music Publishers Association (NMPA) president David Israelite suggested in the same article that as much as 25% of royalty payments are either not being paid out – or are being paid out to the wrong publishers.
What is Spotify’s side of this story? There are three key sources: Duffett-Smith’s submission to the US Copyright Office in 2014; public comments made since the Victory dispute by its spokesperson Jonathan Prince; and a blog post by Duffett-Smith just before Christmas, announcing Spotify’s plans for a “publishing administration system” to solve the mechanical-royalties issue.
Prince’s statement in October set out Spotify’s public attitude: it is not trying to avoid royalties, but the music industry is at fault for the mess in the publishing market.
“We want to pay every penny, but we need to know who to pay,” said Prince. “The industry needs to come together and develop an approach to publishing rights based on transparency and accountability.”
Duffett-Smith’s blog post elaborated on this. “One of our core commitments is making sure that everyone involved in the creation of music is paid fairly, rapidly, and transparently,” he wrote.
“Unfortunately, when it comes to publishing and songwriting royalties, especially in the United States, that’s easier said than done because the data necessary to confirm the appropriate rightsholder is often missing, wrong, or incomplete.”
He went on to note that in the case of publishing rights: “in many cases, the ownership of the rights are not even finalised when a record is released; in many other cases, rights are held by multiple parties, rights change hands, and rightsholders remain entirely unclear.”
ROYALTIES HELD BACK
Duffett-Smith admitted, again, that Spotify does not have the necessary publishing information for every song in its catalogue:
“When one of our listeners in the US streams a track for which the rightsholder is not immediately clear, we set aside the royalties we owe until we are able to confirm the identity of the rightsholder,” he said.
He did not say whether or not Spotify had backed up this policy by filing special notices with the US Copyright Office – the third option listed earlier in this piece – but he did say that the royalties Spotify has set aside “amount to a fraction of one percent of all the royalties we have paid”.
(Based on Spotify’s $3bn of payouts so far, that means a fraction of $30m. The key, of course, is what fraction.)
Duffett-Smith’s earlier submission to the Copyright Office provided Spotify’s view on why it cannot get the necessary publishing data, noting the (infamous) lack of a global, central database of publishing rights; claiming that “third party rights administrators [HFA in Spotify’s case] often struggle with securing the necessary information”; and made a third claim worth remembering:
“Record label licensors of Spotify are not prepared to notify Spotify of such ownership interests and Spotify may not be able to identify the copyright owners from the sound recordings provided to Spotify.”
To recap: the reason Spotify lacks full licences for an unknown number of songs in its catalogue is that it doesn’t know who the songwriters or publishers are.
It says that this information is difficult to find, yet also claims that labels and distributors DO have this information – as, indeed, they are required to do if they are selling songs on iTunes – but that they are “not prepared” to give it to Spotify.
And no, there is no central publishing rights database for the music industry, after the collapse of the Global Repertoire Database (GRD) project in 2014 – and, indeed, the failure of several other attempts before that.
Much of this issue revolves around the decisions taken by Spotify in the US, particularly in 2011 when it launched there. Broadly speaking, for the issue of tracks for which it had not identified the owner(s) of the publishing rights, it had three options:
Plan A – it could have refused to make any of those songs available to stream, which would have meant a smaller launch catalogue, but a fully licensed one;
Plan B – it could have filed special notices to the US Copyright Office to use those works and hold the mechanical royalties for payment;
Plan C – it could have used the works without filing the special notices, leaving itself open to allegations of copyright infringement.
In Spotify’s defence, the Section 115 system was designed for labels trying to get clearance to release 10-12 song albums, rather than streaming services trying to identify the songwriting rightsholders for millions of works – until recently, the filing process was entirely paper-based too.
Whichever plan Spotify chose, it could also have started building its own database of publishing rights, drawing from any and all available sources. Spotify is doing exactly that, but it was only announced in December 2015.
“We are excited to announce that Spotify will invest in the resources and technical expertise to build a comprehensive publishing administration system to solve this problem,” wrote Duffett-Smith. “We are committed to solving it, but it is going to take significant time and effort.”
Audiam boss Jeff Price is one of the critics who think Spotify should have taken this step a long time ago, rather than complaining about the complexity of publishing rights.
“It’s a smokescreen,” Price told Music Ally. “The problem isn’t that the data doesn’t exist. The problem is that you don’t have a system to take the data. And by the way, you never asked for it.”
He continued, “It is a fallacy to suggest the data does not exist or it’s too complicated. If you pick up a phone and call the songwriter and say, ‘Hey, what songs did you write and what percentage did you control?’, 99% of the time, they’ll know. They may even have it written down in a spreadsheet.”
Spotify was trying to tackle this problem, but not in-house. In 2011, when it launched in the US, it hired the Harry Fox Agency to clear mechanical rights. With hindsight, committing resources internally might have been a better option, but at the time HFA was owned by the NMPA, and may have been seen as the best outsourcing option.
Lowery disagrees with claims that the data is hard to find. “The notion that this is somehow the ‘music industry’s’ mess in the first place is suspect. My songs are registered with the copyright office. it’s easily searchable,” he told Music Ally, pointing to a story by NPR where “their copyright expert found the rightsholders in 10 seconds”.
“The problem is that Spotify and its agents apparently didn’t look. Creating a database doesn’t help if a streaming service doesn’t bother to use it. Remember there is a cost to licensing each and every song out in the long tail. Follow the money. Everything else is a distraction.”
Contacting songwriters to ask about their splits requires knowing who those songwriters are. But Price and other sources suggest that if Spotify had chosen Plan A above, not only would it have been left in the strongest legal position – it might also have provided a wake-up call for other rightsholders to help it gather the data.
“Tell labels and distributors they are going to change the metadata schema, and in 60 days we’re going to require you to include the songwriter and publisher information,” suggested Price.
“I would guarantee you, if it was important enough, that information would be produced somehow,” says another source. “Labels aren’t going to take on the obligation of getting the [mechanicals] licence, but if some product manager who was trying to break an artist was told, ‘We can’t get your record out on Spotify’, they would find a way to make it happen.”
The source added that their sympathy for Spotify on the complexity of publishing rights would be greater if the company had chosen not to make songs available without the necessary licences.
“When you hear Spotify saying ‘If we could just find the data…’, that’s the argument you’d make if you could exploit a lot more music if you could just get the licences, but you can’t put it out,” they said. “That’s a good argument there, but it doesn’t make much sense if you already put the music out.”
Price was more blunt. “If you don’t know whose music it is, don’t make it live,” he said. “Whether or not you were licensed has nothing to do with the data being available. If it’s not available, it’s not licensed. Although in their defence, [several] other digital interactive service had done the same thing. It’s not just Spotify.”
Would Spotify alone have the clout to challenge labels in this way – especially those that are among its shareholders? Perhaps a coalition of Spotify, Apple Music, Google Play, Deezer, Rhapsody and others all adopting the same metadata policies would have more luck: but it seems a remote prospect.
At this stage, there are many unknowns about the two class-action lawsuits – including whether they will combine at some point. If the case(s) get to court, the verdict and (if it goes against Spotify) damages bill remain to be determined. With potential for appeals, this process may also drag on for years.
We can make some predictions about what happens next though. Spotify’s decision to build a publishing administration system may be belated, but few of Music Ally’s sources disagreed that it is the most sensible option for tackling this problem.
Why? Because the dream of a global repertoire database remains exactly that: a dream. Or more accurately a nightmare for anyone trying to get such a project up and running.
Even if such a database were to miraculously spring into life tomorrow, Price suggested it would not solve the problem.
“Really the solution has to be a one-to-one pipeline,” he says, of streaming services. “You’ve got to create a system that enables the delivery of information to you, and payments back.”
In Spotify’s case, some sources are calling for the company to populate its database by publishing its list of unidentified works – if it has one, which it should if it is calculating how much money to set aside for royalties – to help songwriters and publishers come forward to claim them. More on this later.
Lowery suggested another option to Music Ally. “One solution would be to allow US PROs to license the mechanical right. In Spotify’s 2014 filing with the US copyright office they (along with most digital services) opposed this move,” he said.
“Probably because the current rules for PROs benefit them marginally on the public performance side. In my view they have hoisted themselves by their own petard.”
NEW YORK EXAMPLE
Spotify says that it is working with “our partners and friends in the music industry” including the NMPA to “find the best way to correctly pay the royalties we have set aside to the right publishers and songwriters”.
The NMPA connection is important: multiple sources stressed to Music Ally that discussions between Spotify and the NMPA about how to address the mechanicals problem were in motion long before the lawsuits, citing this as evidence that the company is not trying to avoid its royalty commitments.
Several sources interviewed for this piece who were otherwise critical of Spotify’s policies in this area said they still see the company as a “good actor” that wants to pay.
That said, if the NMPA and Spotify do strike a deal to pay out the collected royalties, they can expect plenty of scrutiny of the methodology they choose, and how the process works for non-NMPA-members as well as members.
Back to that list, one source cites a 2004 agreement brokered by New York attorney general Eliot Spitzer after a dispute concerning labels and unpaid royalties to artists.
The $50m settlement saw the major labels agreeing to list the names of artists and songwriters who were owed royalties on their websites, while placing ads in trade media explaining how creators could apply for unclaimed royalties; working with music industry bodies to find artists; and sharing the resulting contact information.
“Why don’t they just publish a list? That happened with the New York attorney general years ago when the labels were holding all these royalties and didn’t tell anybody,” says one source.
“Why keep this information secret? Just publish it. You’re no more screwed from a liability point of view than you are anyway – and, if anything, you look better.”
Opinions vary on that point: Spotify may be reluctant to publish such a list while the class-action lawsuits are ongoing, as it would immediately become evidence. That, in turn, may be an incentive to settle the lawsuits as quickly as possible, in order for such a list to be published to help populate the new publishing admin system.
Publishing a list of unidentified works for songwriters to claim; requiring all new music uploaded to Spotify to include publishing metadata; and building out a database of this information that rightsholders can edit when works change hands looks like a positive way forward.
And not just for the streaming service that’s currently being sued. “It’s not just Spotify,” as Price put it. It is not hard to understand why Spotify was targeted first by these class-action lawsuits: it is the biggest player in the on-demand streaming market.
The lawsuits send out a clear message to all Spotify’s direct rivals, though: do they have the correct publishing metadata for every song in their catalogues, and are they paying out all the mechanical royalties that are due?
Those questions have not been asked loudly enough yet, but they will be in the coming months – regardless of the outcome with the Spotify lawsuits.