While writing a paper on methodology of empirical studies on file sharing, I ended up reading an article by Rob and Waldfogel (2006). At one point, the paper starts listing the short comings of another paper, which argued that illegal downloads and file sharing can boost album sales. Rob and Waldfogel approach it like this:
Their contemporaneous-album-sales approach also faces the handicap that the variation in a particular album’s popularity over time would tend to induce a positive relationship between purchases and downloads. Purchase and download are simply two ways of obtaining an album. To see this clearly, consider a different pair of channels for obtaining an album, from Sam Goody or from HMV. Virtually no one purchases a particular album from both outlets. Thus, the displacement is one for one. Yet the weekly sales of a particular album at Sam Goody are surely highly positively correlated with the weekly sales of the same album at HMV (as a result of aggregate temporal shocks that affect both).Do HMV sales stimulate Sam Goody sales? Probably not. Looking at downloading and purchases has the same problem. An album’s popularity has a time component to it, and when it is popular through one distribution channel, it is popular through other channels as well. But this does not mean that availability through one of the channels stimulates demand through other channels.
I had to ignore this and continue writing my paper, but seriously, my brain needs to address this right away.
I suppose my main problem is: Wrong metaphor! Illegal downloads are not substitutes in the same way HMV and Sam Goody would be. You can download an album and very well go and buy it. This is mostly because downloading illegally or sampling on Spotify is for free and because actual CDs have additional value material such as the cover of the album or the notes in the booklet, or sometimes other additional material. Most people do download first and buy the album after listening. Because when there is so much choice and it is quality is always so unclear (until you have listened to it anyways) (the good old nobody knows principle), that you need a service to help you choose. When there are free options (preferably legal) why go buy the album without listening to it first?
Empirical data from 2002 to 2005 still was talking about Napster. That is so sad. In 2012, Napster is a distant memory. I’m sure there is plenty of things happening in research related to digital downloads and sampling etc, I need to yet catch up with those. But it is always nice when I read about these things and get excited. May have more to come as I dive into social media and museums for my thesis soon.