Sound, fury

[I almost served on an FTC panel on file sharing. Made the first cut, but didn't fare so well in the callbacks; I'm generally not as smart-sounding in real time. I'd put a lot of thought into the questions, so I was going to leave a comment, but the notes didn't really gel into an essay with a beginning, middle, and end, and I have to get back to working on my actual book on software policy. *Sigh*. So here are notes which could have added up to something but didn't. If these things inspire you to leave your own comments on the FTC site above (feel free to cut 'n' paste), then I guess they've provided some benefit.

OK, having given you lots of caveats that these are unedited notes, and having cited the appropriate Shakespeare, here they are.]

Different networks (in the sense of humans connecting with other humans) have different characteristics. A wide network, where everybody is connected to everybody, provides utility for different situations compared to a friend-based network, where people are connected to only a few others. Similarly, different media benefit from different types of networks.

Wide networks are potentially detrimental only to certain types of media. It is difficult to argue that friend-based networks would be significantly detrimental to any media relative to the status quo. This comment will conclude that the FTC and media providers should make efforts to encourage friend-based networks, and at the least not discourage them by threats or sabotage.

Network types

Downhill Battle is developing a Gaim plug-in to permit filesharing, so that friends can share media with friends. This mitigates many of the problems which the panel has convened to discuss, such as viruses or spyware being transmitted via file sharing, since one can expect that if a friend has put out a file, it is more likely to be trustworthy. The method is a continuation of how music has been shared for decades. The media industries have thrived with friends trading tapes with friends, and has thrived with friends handing dubbed CDs to friends; the online friend network is the next incremental step, but does not fundamentally change the speed at which goods are transmitted. I expect that friend-based networks for file sharing will become increasingly more public and more common with time.

Compare to traditional file-sharing networks. The typical anonymous Internet file sharing system comprises a complete network, where everyone on the system is connected to everyone else. Humans rarely form complete networks in any groups of more than a few people. For example, I recently did an analysis of junior high school classes in LA; the average class size was about thirty students, yet few students had more than about three friends in the class.

Music

[As an aside, I would like to clarify that although many look down on music which is fashionable but has little merit as music, that is not my intent here. First, the entire debate is about music, not cures for cancer; even the music of the highest merit is still a consumption good for personal enjoyment. Second, there is clearly a massive market for fashion goods, and these goods drive a not-insignificant portion of the economy. But regardless of rhetoric on either side, music as fashion good is a different type of good from music for headphones, and an analysis of the industry as a whole needs to consider the two roles separately.]

Music is a `networked good', meaning that people consume more of it when others are consuming more. But fashion-oriented music is a networked good for different reasons than headphone music.

One who seeks to consume a fashion good simply seeks the goods consumed by the maximum number of people. Even those who claim to eschew the popular still tend to gravitate toward a few strictly delineated genres which have their own sub-fashions; notice how so many fans of alternative (aka indie) music wear the same trucker caps and t-shirts. But the great mass of people are not so subtle: they seek only the most popular overall. For them, the complete network is preferred to the friend network, since it gives fast information about the most popular.

The informational approach to networks assumes that people make decisions based purely upon their private tastes, but they have limited information about the goods available. The network provides that information. The standard example is of two unfamiliar restaurants: if the first is crowded and the second is not, then a passer-by may presume that others have information that the first is better.

Similarly with music, but the twist is that there are a multitude of genres and tastes, so the tastes of a Metallica fan simply have no informational value to a fan of Baroque and early liturgical music. Therefore, for the network to provide information to consumers who will be listening privately on their headphones, it must be subdivided into groups.

For those consumers and musicians who are in the fashion-oriented segment of the market, the complete network is the ideal means of fast distribution (whether desired or not), since it is easy to quickly obtain what is currently the most popular. For those who seek the best music for private consumption, the friend network provides information that the complete network does not; the friend network is therefore more ideally suited to headphone music.

The fashion business

For a fashion-oriented good, the value produced is an image, which is expensive to produce but intangible; instead, vendors sell a tangible good, music, which has low marginal cost. In other words, the business model produces value in one place, but then attempts to recoup profits from another. Such models exist throughout the business world, but they are often tenuous and run risk of failure. A good example would be the sale of printers and toner cartridges: printers add most of the value but are underpriced, while the cartridges are overpriced to provide profits. It is no surprise, therefore, that a small cartridge-producing industry has sprung up to undercut the producers of overpriced cartridges.

Nor is it a surprise that the printer manufacturers have begged the courts to shut down its competitors and protect its business model. From the perspective of market design, is there any reason that we should protect the ability of printer producers to shift prices away from costs? Simply put, no. There exist situations where the normal market model simply would not work, but it is difficult to argue that the printer market is such a place.

Back to music: the question is whether the intangibility of music makes it impossible for the producers of fashion goods to continue to operate. It does in the sense that if the only means of producing a product from intangible fashionability is in music, but this is clearly not the case: the names of successful performers sell apparel, perfumes, music-playing devices, movies, and anything else to which a name can be affixed.

In such a case, the vendors of fashion-oriented music may do well to shift away from providing a purely informational product. Since the core of the product's value comes not from the music itself but from the performer and the advertising, hype, and style, the managers of such performers can vend physical goods such as apparel and notebooks as easily as vending music and images.

But given that their primary product is currently information, their best bet is to slow the speed at which the information can be transmitted. The best way to do this, given that people will certainly trade music among themselves, is to encourage systems based on friend networks instead of complete networks.

[P.s.: Want more writing on the subject that doesn't quite lead to a conclusion but is interesting anyway? Have a look at this PDF.]