The Myth of DRM-Free Music, Revisited

The most widely-read article on this blog these days one that I wrote two years ago called The Myth of DRM-Free Music.  So I thought I’d follow up on it.

I wrote that piece as a reaction to the popular story — which was coursing around the book publishing industry at the time — that Apple led the music industry away from DRM when iTunes went entirely DRM-free in January 2009.  This story was a myth, on two levels.

On the more superficial level, this story was one of the more enduring creations of the legendary Steve Jobs Reality Distortion Field.  Long before Apple’s iTunes announcement, there had been rumblings within the music industry about the major labels’ desire to find viable competitors to Apple, and their willingness to give up DRM in order to promote interoperability beyond Apple’s iTunes/iPod/iPhone walled garden.  Amazon emerged as the entity that could credibly compete with Apple, and it became evident that the majors were working on deals that would enable Amazon to sell DRM-free MP3s.  Steve Jobs heard the rumblings and decided to “get in front of the train” by putting out his now-famous 2007 “Thoughts on Music” open letter and positioning himself as the white knight who led the music industry away from DRM.  This view of Jobs’s PR opportunism was confirmed as recently as 2012 by an engineer who helped develop the DRM for iTunes.

On a deeper level, the myth was that DRM really went away from music at all.  David Hughes, Chief Technology Officer of the RIAA, liked to say that there were many ways of distributing digital music, and all of them used some form of DRM except one: paid permanent downloads.  Those of us who heard this treated it as little more than a rhetorical flourish.  But since sometime last year, it has taken on a new meaning: music delivered with some form of encryption now earns more total revenue than music delivered DRM-free.

My article in May 2015 predicted that “encrypted modalities will represent the majority of recorded digital music revenue by 2016.”  This wasn’t exactly rocket science, given the graph that showed two lines — percentages of digital music revenue delivered with and without encryption — converging towards each other.  But now it has actually happened:


Percentage of recorded music industry digital revenue from DRM-free vs. encrypted delivery modalities.  Sources: RIAA, Pandora Media, Sirius XM Radio.

Note that this chart is not exactly an extension of the one from two years ago.  On that chart, I counted a fixed portion of digital radio revenues (SoundExchange distributions), on the basis that such a percentage comes from services that encrypt their signals.  This time I got more accurate about that calculation.

SoundExchange collects digital performance royalties on five types of services: pure-play Internet radio (e.g., Pandora), AM/FM simulcasts (iHeartRadio), satellite radio (Sirius XM), cable/satellite TV radio (Music Choice), and business establishment music services (Muzak, DMX).  Most of these services encrypt their signals, the primary exception being many AM/FM Internet simulcasts.  Unfortunately, SoundExchange doesn’t break out its collections (revenue) by service type, so we have to do some calculations to estimate what percentage of its revenue comes from encrypted sources.

As it turns out, two services that encrypt their signals account for the vast majority (75-90%) of SoundExchange’s collections: Sirius XM and Pandora.  Both are publicly traded companies, so it’s possible to get a good idea of how much they pay SoundExchange from their annual reports. In Sirius XM’s case, they pay SoundExchange a percentage of their annual revenues that is set by law (and varies from year to year).  Pandora’s rates are also set by law and are more complicated, but Pandora does us the favor of reporting how much of its revenue it pays to SoundExchange in its annual reports.  (Neither Sirius XM nor Pandora pay royalties to SoundExchange on recordings either made before 1972 or on labels that have direct licensing deals with the service.)

In the above graph, I’ve taken the percent of revenue that SoundExchange gets from Sirius XM and Pandora and applied that percentage to the amounts that it actually distributes to record labels, i.e., the figures that it reports to the RIAA.  The result is a good estimate — a tight lower bound — of the percentage of SoundExchange distributions that come from encrypted delivery channels.

Apart from SoundExchange, assigning recorded music revenue sources to encrypted and DRM-free buckets is straightforward. “Paid Subscriptions” (e.g., Spotify Premium, Apple Music, Pandora One) and “On Demand Streaming (Ad-Supported)” services (e.g., Spotify Free, YouTube) are encrypted, while downloads are DRM-free.  I treat “Ringtones & Ringbacks” as encrypted, even though some probably aren’t, and in any case that revenue source has shrunk into insignificance (now less than 1% of total recorded music revenue) so it’s not material.

The fact is that sometime during 2015, digital music with some sort of encryption scheme became bigger than DRM-free — by consumer choice.  Sales of downloads have been dropping since 2013 and are now essentially in free fall, while all forms of streaming and subscription downloads (a/k/a “offline listening mode”) are on the rise.  This relates to the point I made recently about how consumers don’t seem to value ownership of pure digital files very much.  Certainly this is the case in book publishing, where e-book sales are stuck below 25% of trade book sales despite all the breathless predictions of the late 2000s that e-books would exceed print books by now.

The big difference between books and music in this respect — as my colleague Mike Shatzkin pointed out — is that while there’s a big difference in the user’s experience of the actual content between print books and e-books, there’s very little difference between the user’s experience of recorded music between physical and digital media.  That helps explain why print books outsell e-books by more than 3 to 1, while for vinyl and digital music it’s the other way around.

At the same time, physical music products — including CDs, which are physical digital products — seem to be finding a floor.  In addition to the well-known rebirth of vinyl LP sales, the decline in CD sales has slowed down, and cassettes are even starting to make a comeback thanks to their lower cost and faster turnaround time to produce.


Percentage of recorded music revenue from downloads, streaming, and physical products.  “Streaming” includes subscription downloads.  Source: RIAA.

This chart suggests that physical music products may bottom out around 20% of total industry revenue.  We’ll know more next month when the RIAA releases its revenue figures for the second half of 2016.




  1. […] Source: The Myth of DRM-Free Music, Revisited | Copyright and Technology […]

  2. Nice work Bill. I should add, ALL of the streaming music services use DRM esp the ones with downloads 🙂

  3. Actually that is not quite true. I just checked it carefully.

    First, “Streaming” covers a lot of territory. Interactive streaming (Spotify, Apple Music, Google Play, Rhapsody, Deezer, Tidal) – yes, definitely all encrypted. Satellite radio (Sirius XM), yes, also encrypted. And business establishment services like Muzak and DMX probably also encrypt because they are also paid subscription services. But so-called noninteractive streaming varies. It’s possible that pure-play Internet radio services such as Pandora encrypt their signals; I don’t know for sure, and they ain’t tellin’. On the other hand, AM/FM simulcast signals generally don’t encrypt; anyone can use a standard stream recorder and capture those signals. These signals are available through individual station apps and through aggregators such as TuneIn and the NPR app.

    Then there’s the problem of assigning recorded music revenue to each of those types. Interactive streaming services get broken out by the RIAA numbers pretty cleanly. Digital radio services with statutory licenses that pay royalties through SoundExchange are the problem. SoundExchange does not break out its royalty collections by type of service. It’s not possible to tell how much of SoundExchange’s revenue comes from pure-play webcasting (Pandora and its ilk), because Internet radio (noninteractive webcasting) royalties are calculated on a per-stream basis and not as a percentage of revenue, and that data’s not available. It’s not possible to tell how much of SoundExchange’s revenue comes from business establishment services because the largest one, Mood Media (owner of Muzak and DMX), offers various other non-royaltable services and doesn’t break out its revenue.

    The only one we can tell with any accuracy is satellite radio, which we know is encrypted and which actually brings in more than half of SoundExchange’s revenues, so it’s a decent lower bound. We can tell how much Sirius XM satellite radio pays to SoundExchange by taking advantage of the fact that it’s the only satellite radio (SDARS) service and using publicly available data on its revenues and its statutory digital performance royalty rates. From there we can calculate the percentage of SoundExchange revenue that comes from satellite radio. So I’m using that as a lower bound for the percentage of revenue to assign to the “Encrypted” category; the true number is somewhat higher.

    In other words, it’s possible that the true crossover point was up to a year earlier than 2016, but still recent.

    The other problem with these figures in general is that they only count recorded music revenues. They don’t count revenues from music publishing (performance and mechanical royalties collected by the likes of ASCAP, BMI, Harry Fox, and major music publishers). Total U.S. music publishing revenue is about 30% of recorded music revenue, though a) that’s rising slightly each year, b) the royalty rates are completely different from those of recorded music, like apples and oranges, and c) the music publishing industry (represented by the NMPA) doesn’t break out revenues the way the RIAA does for recorded music — all of which makes it impossible to calculate for these purposes.

  4. Jason Haberny · ·

    “digital music with some sort of encryption scheme became bigger than DRM-free — by consumer choice. ”
    Is it fair to describe this as “by consumer choice” if consumers are not properly informed?

  5. Ok, I’ll bite: informed about what?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: