Some weeks ago, we published a lengthy blog post called Where do Music Collections Come From?  which discussed findings from our Copy Culture survey.

Some of the data demonstrated that P2P file sharers (who own digital music files) buy more music than their non-P2P using peers (who also own digital music files).  Here’s the chart again:

To me, this was a fairly innocuous finding, well in line with other studies.  For my money, the more important findings were that personal sharing ‘between friends’ is about as prevalent and as significant in music acquisition as ‘downloading for free’, and that together they are outweighed by legal acquisition.

But the public spoke and the P2P finding went viral: the biggest pirates are the best customers.   Headlines like this generated pushback from record industry groups RIAA and IFPI—mostly centered around the work of NPD, their survey firm in the US.  The exchange, I think, is an interesting window on the state of the empirical debate around file sharing.

Here’s the chronology:

Oct.15: We argue that P2P users are the biggest buyers of recorded music.  The story jumps from TorrentFreak to Gizmodo to many many other sites.

Oct.16: Russ Crupnick, Senior VP at NPD tells NBC News’ tech blog:

We hear this argument all the time and it makes no sense…. Peer-to-peer users tend to be younger and more Internet-savvy, so the likelihood that would be buying digital files makes perfect sense. But you can’t compare that to the entire population.

Oct. 17: We point out that we didn’t compare P2P users to the general population, but to digital music file owners (50% of the US population; 42% in Germany).  We acknowledge that our labeling was a little ambiguous on this point, so we fixed it.  We noted that “if NPD has data that suggests otherwise, perhaps they could share it.”

Oct. 17: IFPI weighs in, arguing that NPD says that most P2P users are moochers, even if a few skew the average by buying a lot:

P2P users spent US$42 per year on music on average, compared with US$76 among those who paid to download and US$126 among those that paid to subscribe to a music service. The overall impact of P2P use on music purchasing is negative, despite a small proportion of P2P users spending a lot on music.

Oct.18: We say, OK IFPI, that’s not super clear.  Those categories don’t seem mutually exclusive. But we take your general point so let’s break down the P2P users with digital music collections.  Here’s what our data says:

  • 16% bought no music files.
  • Another 9% said that 10% or less of their music file collections were purchased.
  • The median music file collection, among P2P users, is around 50% purchased.
  • And 15% said that their whole collection was purchased (suggesting that they used P2P for other purposes).

It’s a diverse group, but the vast majority buy music.  We stand by our claim.

Oct.19: Then Russ Crupnick at NPD writes a piece that accuses us of publishing while drunk and also lacking a license to make proper sense of data (not joking about this).  He repeats that you can’t compare P2P users with the general public, and then notes that we’re right about P2P users–but also wrong because it’s dumb to be right about this.

P2P music downloaders do indeed buy more music than non-users.  We’ve known that for about 10 years. It’s a dumb, illogical and irritating argument.

He then brings out his presumably non-drunk, licensed findings and, well, there are a couple things to say.

First, he gets his math wrong by including the subtotal in the grand total (h/t Michael Geist).  Possibly this is advanced ‘licensed’ math.  I wouldn’t know.

Second, when corrected, the numbers are pretty similar to ours! P2P users do buy more legal music than non-P2P using music buyers. And if you add in concert and merchandise, they spend quite a bit more on music.

As near as I can tell, Mr. Crupnick has no actual disagreement with us on the P2P findings.  That’s just smoke and mirrors.  Rather, he want to shift the focus to two other claims.

First, that even though  P2P users buy more than others music buyers, they buy less than they used to.

The average P2P user spent $90 per capita on music in 2004- now they spend $42 (CDs, downloads, subscriptions). This was during the same period when the number of files illegally downloaded per capita was rising.  

Our spending numbers would look higher, but we agree with the basic story. Here’s how we put it:

[P2P users] are better digital consumers.  But is also clear that this investment has fallen vis à vis large CD-based collections.  The survey offers ample evidence of this shift in the way music aficionados relate to music–no longer organized around large CD collections or measured in terms of individually priced songs or albums, but rather defined by a mix of legal and illegal strategies for accessing everything now.

Then he gets to what he really wants to talk about:

Celebrating P2P users for their contribution belies the fact that the paid component of the music that they acquire, aka their acquisition mix, is 50% less than the average music consumer.

Ah.  The moral order is restored.  Or is it?  On any normal reading of the post, this makes no sense: P2P users can’t simultaneously spend more and 50% less than other music buyers.   (Admittedly, I’ve had a few drinks and should probably leave this to the metaphysicians at NPD.)

But I’m willing to go the extra mile and assume that Mr. Crupnick is just being unclear, rather than contradictory.   Maybe the “paid component” refers to the percentage of overall collections, not to the annual “spend” on music.  This would have the virtue of making the statement true, in the self-evident sense that P2P users acquire more music than they buy. In our formulation above, the median music file collection, among P2P users, is around 50% purchased.

But it wouldn’t make the statement relevant. At this stage of the game, knowing who supports the music ecosystem and what their expectations are matters a great deal. The fact that P2P users pirate, on the other hand, only matters if your main strategy for increasing sales is enforcement.  Boiled down, Mr. Crupnick’s point is that it’s more important to stigmatize the pirate than understand the customer.

Nov. 12:  The RIAA’s Joshua Friedlander steps in to endorse this view:

In reality, the comparison is unfair – what it’s comparing is people who are interested in music with people who might not be interested at all.  Of course people interested in music buy more.  But as research firm the NPD Group (which has been studying these issues for a decade) points out here, this data is neither new, nor illustrative.  In their words, “Celebrating P2P users for their contribution belies the fact that the paid component of the music that they acquire, aka their acquisition mix, is 50% less than the average music consumer. Yes, that’s half the average.”

For what it’s worth, I think piracy does play a role in declining purchases of recorded music, but I also think there are so many forms of disruption in the market that it’s impossible to isolate that impact.  Here’s how we put it in a post called Die Substitution Studies, Die II: Well, Maybe Some Should Live.

We’ve argued that the media ecology has become so complicated that nobody has a handle on what substitutes for what.  Does a pirated MP3 file substitute for a $1 purchased file, a $12 CD, some number of listens on YouTube or Spotify or radio? Does Spotify substitute for MP3 purchases?  Or YouTube listens? Should we take stagnant discretionary income into account, and rising costs for other media services, like cable TV, Internet access, and data plans.  Do national differences matter–including major differences in digital markets and services (In Germany, CD sales represent over 80% of the market; in the UK and US, under 50%)…. Which of these factors get priority?  How do we model their interaction?

Increasingly, we don’t think it matters.  For younger music fans, the primary connection to music no longer passes through carefully curated CD (or MP3 ) collections but through the universal jukebox approximated by overlapping services–iTunes, YouTube, Spotify, The Pirate Bay, and your friends’ collections.  The total spend is shaped not just by the availability of pirated music, but also by the close complementarity of other free and cheap music services and by the greater competition for discretionary income and attention from other media–games, DVDs, apps, data plans, concerts, and so on.

So what’s at stake in all the misdirection and cheap shots?  In a generous mood, I’d say carelessness.  In a less generous mood, I’d say it sounds like resentment that Mr. Crupnick has to debate this stuff at all.  Ten years ago, he didn’t have to.  Send out the press release, watch it get picked it up, and call it a day.  NPD and RIAA simply owned the discussion.  Now they have to nitpick with academics.

Companies like NPD make money not just by surveying people about their habits, but also by ensuring that the data that they release leads toward conclusions their clients like.  This is the noxious side of an advocacy-driven research culture.  And for many research firms, it produces occasional schizophrenic moments: the social scientist warring with the company man.  Maybe that’s what we’re seeing here.  The P2P results may have been obvious and ‘known’ for years but I can find no trace that NPD thought them worth mentioning before this exchange flushed them into the open.  NPD has tons of data and could make a huge contribution to public understanding of these issues, but that’s not their job.   Dissonant findings stay confidential.

Which is too bad, because in the end, Mr. Crupnick arrives at many of the same conclusions we do.  From earlier this year:

“There are always going to be those who look for bootlegs and songs you can’t find on sites like Spotify and Rdio, and there will always be people who see illegal downloading as a sort of game, but I think that number will just get smaller and smaller as other options become more convenient with all your devices,” says Russ Crupnick, senior entertainment industry analyst for NPD.

The reason for this, as Crupnick and others note, isn’t because of potential legislation that mirrors SOPA so much as the growing number of cheap, legal alternatives to illegal downloading combined with the decline of many well-known file-sharing sites.

So what’s he defending? Not different data or even significantly different findings, but–by all appearances–just his client’s failed monopoly on interpretation.  But that drunk horse has left the barn.