We can’t really “prove” that either of these is true, but the approach we have taken, in which we compare paid sales before and after piracy occurs, gives us at least a window into whether piracy had an immediate impact on paid sales. What we saw with the O’Reilly titles is an increase in paid sales after piracy occurs. It’s not uniform, but on average sales increase.

That doesn’t make this true for all titles or even any other titles – we have to research them to say what happens. I do think that the estimates of the cost of piracy that the AAP and others have made may be high – they see each instance of pirated content as indicative of multiple lost sales totaling millions of dollars annually. That may ultimately be true, but we have not seen evidence that pirated content is the same as a lost sale.

Some authors see every downloaded copy of their books as a potential lost sale. In reality, it’s impossible to know how many of those who downloaded the book would have purchased a legitimate copy had it not been available for free. Do publishers make estimates of much money they lose through illegal downloads? If so, how do they go about this?

DRM is certainly not something readers ask for, and there is debate about how good it is at reducing piracy. We hope our research will help publishers understand what types of content are most likely to be pirated, and among those titles, which books are more likely to be hurt by it. If we can look at enough titles to draw some conclusions, it may be possible for publishers to be more selective in using DRM.