DRM and Ebooks: The Statistical Bikini

Cory Doctorow posted an interesting note on his blog. Basically, he noted that O'Reilly had dropped Digital Rights Management (DRM) on its ebooks, and since they did their e-book sales had risen 104%. The most interesting part to me was reading through the comments... in which they proceeded to eviscerate the note for a number of logical errors.

First, and foremost, as several commenters pointed out, correlation does NOT imply causation. Just because two events happened in the same time frame, even in the same location, that does not mean that one caused the other. Additional evidence of a link must be provided to draw that inference. Data which was completely lacking in this post.

Second, other commenters noted that the absolute numbers were not provided. One commenter noted that his books sold through O'Reilly had only sold 1 e-book last year, so if they manage to sell 3 next year it would be a 300% increase... but not really impressive when you're staring at your royalty check.

Finally, others made the important point that sales of ebook readers had taken off like a rocket in the past year, so one would expect sales of ebooks to rise. A more logical investigation would be showing the number of new ebook readers sold, the installed base, and comparing those figures to the rise in ebook sales for O'Reilly. One might even find that O'Reilly's sales of e-books dropped relative to the increase in the total market size, or to an O'Reilly competitor. Which would put the whole blog post completely backwards from reality.

However you feel about DRM and its utility or value, you'd do well to use rigorous logic and statistics to make your arguments. Cheerleading with isolated facts does not advance your argument; it hurts it.

Always remember that statistics are like a bikini; what they reveal is interesting, but what they conceal is vital.