Music Downloads: Pirates—or Customers?

Professor Felix Oberholzer-Gee and co-author Koleman Strumpf floored the disbelieving music industry with their findings that illegal music downloads don’t hurt CD sales. Oberholzer discusses what the industry should do next.
by Sean Silverthorne

Internet music piracy not only doesn't hurt legitimate CD sales, it may even boost sales of some types of music.

Those were the counterintuitive findings released in March by Harvard Business School professor Felix Oberholzer-Gee and his co-author Koleman Strumpf, of the University of North Carolina at Chapel Hill. Their paper, "The Effect of File Sharing on Record Sales," caused a ruckus in the music industry not seen since the British invasion of the Beatles.

Many recording executives were not singing "Yeah, yeah, yeah," however. Convinced that illegal downloading and file sharing has robbed them of billions of dollars after four consecutive years of falling music sales, they criticized the team's methodology, which consisted of monitoring 1.75 million downloads over 17 weeks in 2002, scouring through server logs from OpenNap (an open source Napster server), and comparing the sales of almost 700 albums as reported by Nielsen SoundScan. Oberholzer and Strumpf concluded that there was almost no relationship between the two.

How could this be? The researchers believe that most downloading is done over peer-to-peer networks by teens and college kids, groups that are "money-poor but time-rich," meaning they wouldn't have bought the songs they downloaded. In that sense, the music industry can't claim those downloads as lost record sales. In fact, illegal downloading may help the industry slightly with another major segment, which Oberholzer and Strumpf call "samplers"—an older crowd who downloads a song or two and then, if they like what they hear, go out and buy the music.

Interestingly, the first half of this year saw the release of numbers seemingly supporting this theory: The number of illegal music downloads continued to increase—but so did music sales.

If in fact the research is correct, the strategic implications for the music industry are profound. Instead of conducting a high-profile campaign against pirates, should the industry instead target "samplers" to encourage them to buy more music? Should the industry consider peer-to-peer services as marketing tools rather than the enemy? Should online pricing be different from in-store pricing? What happens when broadband makes it as easy to illegally download an entire CD as an individual track or two? HBS professor Feliz Oberholzer-Gee recently spoke to Working Knowledge about these issues.

Sean Silverthorne: The draft of your paper with Koleman Strumpf came out almost three months ago, and caused quite a stir both inside the entertainment industry and out. What are your impressions of the reactions so far?

Felix Oberholzer-Gee: Two recent developments are important. Our study provides the first serious evidence that file sharing cannot explain the decline in music sales in the last couple of years. In addition, in the last two quarters, music sales increased while file sharing has become even more popular., an Internet monitoring firm, estimates that there are now up to 9 million simultaneous file sharers, up from about 4 million in early 2003.

In view of our evidence and these new trends, even the Recording Industry Association of America (RIAA) now states that file sharing is only "one factor, along with economic conditions and competing forms of entertainment that is displacing legitimate sales." The industry is rethinking its position, although change occurs slowly.

Q: Let's talk strategy. What have been the recording companies' strategies to date for combating their loss of property rights via illegal downloading? And how effective has that strategy been? For example, is it a good thing to sue potential customers?

A: Suing potential customers is not exactly a standard entry in the book of good CRM. More importantly, the RIAA's legal strategy is hopeless and smacks of short-sighted panic.

Our research shows that only 45 percent of music files downloaded in the United States come from computers in the U.S. More than 100 countries supply files to the U.S. file-sharing community, and many of these countries do not have strong records of protecting copyrighted materials. The RIAA does not stand a chance to implement an effective legal strategy in all these countries.

The RIAA's legal strategy is hopeless and smacks of short-sighted panic.

Those who dream of legal solutions do not recognize the truly global nature of the peer-to-peer (P2P) phenomenon. Even worse, the RIAA's legal strategy does not even seem to work here in the United States. Despite the lawsuits—the RIAA has sued about 2,000 individuals to date—file sharing is more popular than ever.

Q: Assuming your conclusion is right—that there is no evidence that illegal music downloads erode CD sales—and in fact might help top-selling record sales—what are the implications for the recording industry in terms of strategy?

A: Our research shows that people do not download entire CDs. They download a few songs, typically the hits that one would also hear on a Top 40 station. This suggests that P2P is much like the radio, a great tool to promote new music. The music industry has of course long recognized that giving away samples of music for free over the airwaves can stimulate sales. The same seems to hold for P2P.

The problem with radio as a promotional tool is that it can be quite expensive for labels to get radio stations to play their music. P2P networks are promising because they make the market for music promotion more competitive. From the perspective of the music industry, the more competition among P2P services, the less costly it will be to promote music.

Q: Apple's iTunes has seemingly validated the concept that people will purchase music online. But it seems the recording companies themselves have done little on their own to experiment with models here, such as tiered pricing (hits cost more) and bundling.

A: The classic business model was a teaser model: The music labels provided one or two hit songs for free by promoting them on the radio and on MTV. If consumers liked the samples, they purchased a dozen songs at a price of $15. We now have gone from one extreme to the other. While inflexible bundling was the rule, services such as iTunes now completely unbundle CDs and offer all music by the song. The difficulty with this approach is that the economics of producing music are characterized by significant fixed costs. It is not much more expensive to promote an entire album than to promote an individual song. With complete unbundling, the revenue streams generated by a new album are likely to be much lower. How many consumers will pay a dollar for song number thirteen?

Clearly, there is a profit-enhancing role for some type of bundling even with digital distribution. For example, consumers might be willing to pay full price for the core songs on an album if they get the rest at a discount. We need systematic experiments to find out which types of bundling are economically most attractive.

Q: What's the current state of your research? Where does it go from here?

A: A key uncertainty relates to our finding that file sharers do not download entire CDs. We do not know why they sample only a few songs. One possibility is that the current patterns of file sharing reflect consumer preferences. Consumers do not know the quality of new music and sampling one or two songs is good enough to assess quality and make a purchasing decision. If this view is correct, the radio model is well and alive, and P2P offers great opportunities to promote new content.

Clearly, there is a profit-enhancing role for some type of bundling even with digital distribution.

However, it is also possible that the observed behavior is due to technical difficulties. In our data, only one out of three downloads is completed successfully. File sharing is fairly cumbersome for many consumers with poor Internet connections. If this is the reason for highly selective sampling, we can expect consumers to download entire CDs when broadband connections become more common. This is a less rosy scenario for the music industry because downloads of CDs are likely to be closer substitutes for CD purchases.

If poor Internet connections explain file-sharing patterns, general access to broadband would have profound strategic implications, suggesting that music companies ought to pursue a strategy of selling complements to recorded music. We see some examples for this strategy even today: Apple sells songs to promote its iPods. Prince gives away his most recent release to promote his concerts. We need careful continuous monitoring of the effects of P2P to know which strategies are most appropriate in the digital age.