Following up our piece on music collections and the apparently surprising result that P2P users buy a lot of digital music, and also our piece discussing some pushback from NPD–the survey firm used by the RIAA–we offer a response to comments by the record industry association IFPI.  Here’s what they said:

While previous studies have shown that some unlicensed P2P network users also pay for music, and a few are serious fans who pay a lot, they are far outnumbered by the bulk of unlicensed P2P network users who pay little or nothing for music. Research by The NPD Group during 2010 in the US found that just 35 per cent of P2P users also paid for music downloads. 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. That finding was corroborated by a study in Europe by Jupiter Research in 2009.

A more recent review of studies on the impact of piracy, Assessing the Academic Literature Regarding the Impact of Media Piracy on Sales, by Michael D. Smith and Rahul Telang of Carnegie Mellon University, concluded that the vast majority of these studies found that piracy harms media sales.

Let’s break this down.  The first point is an interesting one.  In effect, it says that we’re wrong for just looking at P2P users ‘on average’ because if you break this group down further, most are moochers, not buyers.  That’s possible—though again they’re relying on unreleased NPD findings, which is annoying and makes comparison difficult.

(On which point: like many others, I’ve given up asking RIAA, IFPI, BSA, MPAA, and others for more transparency about their research and methods.  None of the major industry  groups will provide it–except, for some reason, the International Chamber of Commerce.  Good show guys.  For our part, we’ll be releasing top line results, cross tabs, and the dataset with the final report.)

Anyway, let’s take their criticism seriously and break down our P2P user data down further.

Among P2P users with digital music collections:

  • 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).

So like most groups, this one is pretty diverse in its practices.  Not everyone lines up with the average.  And even those with less than 10% purchased collections can be heavy buyers when those collections scale beyond 10,000 songs (as they do for 5% of digital music collectors under 50, in our survey).   Our results do not support the NPD claim that P2P music buyers “are far outnumbered by the bulk of unlicensed P2P network users who pay little or nothing for music.”  Quite the contrary.  We stick with our position that P2P users are—on average—both the heaviest pirates and the heaviest buyers of new music.

As for the part about whether piracy harms sales—that’s a bait and switch.  We make no claims about whether piracy harms sales, and we are not saying that piracy boosts sales.  We’re saying that many of the biggest music fans now want access to everything because that’s what digital technologies have enabled.  And many of them pursue this goal through a mix of legal and illegal strategies.

My view is that the debate about sales has begun to resemble one in which economists look for their house keys under the street lamp because that’s where the light is—while their homes are being foreclosed by the banks.  And in this version of the story, the RIAA and IFPI are in the business of pretending that the main problem is lost keys.

So, per Smith and Telang, it is true that most ‘substitution studies’ find that piracy substitutes for some portion of legitimate sales.  But I’m very skeptical that these studies tell us much of anything about the larger ecology of cultural consumption of which music is a part.  The problem, for recorded music sales, is not just that they ‘compete’ with piracy, but also with the vast array of other discretionary expenditures, forms of leisure, and alternative forms of legal music consumption that have emerged in the past decade.  So the goal isn’t to model the record label fantasy of a world without Napster (and its successors), but also a world with a boom in DVD sales, video game sales, and concert revenues that compete for discretionary income. A world in which the $15 album is has been replaced as the main unit of sale by the $1 song.  A world with the emergence of many adequate legal substitutes for buying songs, such as music streaming via YouTube or Pandora or industry-approved solutions like Spotify.  A world in which leisure time is routed through Facebook or Twitter–and in which those forms of community are built in part around the sharing of media. A world in which vastly increased access to old music and TV undercuts the value of the new.  A world with erratic enforcement practices.  And–finally–a world in which median incomes have stagnated while other media costs, such as cable TV, Internet access and phone data plans, have risen.   (On this last subject, Will Page is doing some pioneering and—so far—lonely work that puts substitution in its proper wider economic context.)

Content industry groups used to argue that a pirated song was directly equivalent to a lost sale.  This ran counter to common sense so researchers began to develop substitution studies that tested that assumption.  Results varied.  For music, they said: maybe the acquisition of a pirated song displaces 6% of a sale, or 20%, or 30%.   A few said sampling or discovery effects outweighed substitution effects.  It took a while for this perspective to penetrate the industry groups.  But over the last few years, they’ve made peace with substitution studies because (1) they’ve been pushed to engage with credible research and (2) at least those studies keep the focus on piracy vs. sales.

But the wider cultural ecology is very different than a decade ago, and the forms of competition for music purchases are more numerous and compelling.  Piracy is part of this but only a part—and so far, in my view, impossible to isolate.  Our data suggests that the issue isn’t whether someone will stop buying music altogether but whether, in this competition for time, attention, and resources, the practice of purchasing recorded music can command the same share.  If we follow the industry in taking peak CD sales as the benchmark, the answer is no.



One thought on “NPD Confidential II: Die, Substitution Studies, Die

Comments are closed.