By the end of 2016, Dewar changed Paperchain’s strategy. Rather than help streaming services identify rightsholders, the company would help rightsholders sort through the NOIs to find money they were owed. He wrote a program to download and query the NOIs filed with the Copyright Office, then began to study how much money all of these NOIs represented. What he found was that, although streaming services aren’t doing a very good job of matching compositions with rightsholders, the number of NOIs don’t reflect the number of compositions for which rightsholders aren’t being paid. Spotify was sloppy enough that Wixen Music Publishing filed a lawsuit for the service’s unlicensed use of music by Tom Petty, Neil Young, and other household names, and a smart intern ought to know who wrote some of the songs the service infringed. But services also seem to be filing NOIs for compositions they’re also paying for: Although Spotify filed an NOI for Ed Sheeran's "Perfect," Sony/ATV Music Publishing confirmed that the service had been paying for its use of the song. “It’s a misnomer that Spotify can’t find Ed Sheeran,” Dewar says.
Services shouldn’t be filing NOIs for songs that can identify and pay for – and some observers have suggested it’s illegal to do so. But the only thing that’s easy to understand about this entire mess is why they’re doing so: NOIs has become a process for CYA, which is to say they’re covering their asses. Faced with the potential for stratospheric statutory damages, streaming services are burying the Copyright Office in an avalanche of paperwork. This is dubious (unnecessary), dishonest (because an NOI is essentially a statement that Copyright Office records don’t identify the rightsholder, when in some cases they do), and yet exactly what many businesses would do in the same situation. It’s also scary for music publishers: An NOI offers a temporary cost-free license, and enough of them together could potentially undermine the ability of publishers to enforce their rights. It’s death by a thousand paper cuts!
The problem is hard to quantify because it rarely affects big songwriters like Ed Sheeran, many of whom are signed to major publishers. It mostly impacts underground or even amateur songwriters, many of whom upload their music through services like TuneCore. Since services make so much of this material available, most experts believe that streaming services can’t accurately account to rightsholders for a relatively large percent of the compositions they offer – some suggest it could be more than 20 percent – but also that these compositions make up a relatively small percentage of music consumers are listening to. In a study he presented at the early 2018 Copyright and Technology conference, Dewar suggested that NOI "black box" royalties are worth between $50 million and $300 million. Some of that money was probably generated by Tom Petty hits, as the Wixen lawsuit suggests, but a lot of it almost certainly comes from hundreds of thousands of songs hardly anyone has heard.
Dewar believes that up to 90 percent of these compositions could have generated less than five dollars in mechanical royalties. But since U.S. copyright law assesses statutory damages per work rather than per use, each of them could still theoretically cost Spotify up to $150,000. (A jury would almost certainly assess much less than the maximum in damages, but even a fraction of that would be devastating.) Streaming service executives could make the case that this is absurd: Why should hobbyists be eligible for more in damages than they could possible make on music? Then again, if these songs are worth so little, why are services streaming them in the first place?
In the end, the way publishers and songwriters feel about Spotify and other streaming services has much less to do with how accurately they’ve paid rightsholders in general than how they’ve paid them in particular. Think about it. If a company pays 75 percent of rightsholders accurately, including you, it’s doing a pretty good job in a very tough situation. If it pays 95 percent accurately, and you’re part of the remaining 5 percent, it’s either run by thieves of morons. Thinking that way is illogical, unreasonable, and also inevitable. It’s just human nature.
The debate about how to fix this problem has been complicated, since no one entirely agrees on exactly what problem they’re trying to fix. Publishers and songwriters want to get paid accurately; streaming services want to reduce their legal risk; and publishers as a group would like to get more control over the mechanical royalty payment process. The Music Modernization Act, which was formally introduced on Tuesday, is supposed to address that first problem, and to an extent it does. But it will be far more effective at addressing the second and third -- which are important for the future of the music business overall. Like Blockchain, the bill addresses a problem that can’t really be solved without an accurate database of rightsholders that matches compositions to sound recordings -- which simply doesn’t exist yet.