Faced with the growing popularity of ‘pirate’ libraries such as Z-Library and Anna’s Archive, book publishers have ramped up their anti-piracy efforts.
This year alone, Google has processed hundreds of millions of takedown requests on behalf of publishers, at a frequency we have never seen before. The same publishers also target the pirate sites and their hosting providers directly, hoping to achieve results.
Thus far, little is known about the effectiveness of these measures. In theory, takedowns are supposed to lead to limited availability of pirate sources and a subsequent increase in legitimate sales. But does it really work that way?
To find out more, researchers from the University of Warsaw, Poland, set up a field experiment. They reached out to several major publishers and partnered with an anti-piracy outfit, to test whether takedown efforts have a measurable effect on legitimate book sales.
The study only takes printed books into account, since the Polish e-book market is rather small, and statistically reliable sales data is difficult to obtain.
One-Year Takedown Experiment
The research methodology for this study is quite straightforward. In total, 12 prominent publishers took part, of which three dropped out. The publishers shared between 5 and 53 book titles, which were assigned to a takedown group or a control, for which no takedown notices were sent.
The group assignment wasn’t completely random. Instead, books were matched in pairs based on characteristics such as price, format, and previous sales figures, to remove as much noise as possible.
After the books were assigned, those listed in the takedown group were shared with the Polish anti-piracy outfit Plagiat.pl, which started protecting these works. For the remaining books, no action was taken. In total, Plagiat.pl found pirated copies on 53 sites and continually issued takedown notices for a full year.
Takedowns Were Effective, Sales Trend Up
After the research was completed, three assistants were asked to conduct online searches to confirm whether it was harder to find pirated copies ‘protected’ books online. That was clearly the case, as fewer copies were spotted, and it also took longer to find them
The effect of these successful takedowns didn’t clearly translate to sales of print books, however. While the researchers found a small positive effect in the takedown category, it wasn’t pronounced enough to be statistically significant.
Only after the researchers tried a Bayesian analysis, adding data from previous research, did they find an uptick in book sales.
“We were able to substantially curb the unauthorized distribution, which resulted in a small, positive effect on sales,” the researchers write.
“While using classical analysis we found it not to be significantly different from zero, a Bayesian approach using previous ‘piracy’ studies to generate a prior led to the conclusion that protecting from piracy resulted in a significant sales boost of about 9 per cent.”
Open Ending
The paper attributes the lack of a stronger initial result to the relatively low sample size. With more books, it would have been easier to get a more definite answer. However, the positive direction doesn’t contradict earlier research.
For example, a previous study found that takedown notices can lead to an increase in sales of e-books, which are digital, and therefore a more direct substitute for pirated copies.
Strong results or not, the researchers believe that their work is an important contribution to the existing literature. Piracy research typically relies on before-after comparisons, while this study allowed results to be compared during the same time period, with a full control group.
In the end, however, many questions remain unanswered, so follow-up research is warranted. It would be interesting to see the same approach in countries where e-books are more prevalent too.
Since the major publishers are now heavily involved in takedown efforts, they might be interested to see if they get a decent return on their investment? After all, sending takedown notices typically isn’t free.
—
A copy of the paper, published in a recent edition of the Journal of the Economic Science Association is available online here.
Hardy, W., Krawczyk, M. & Tyrowicz, J. Internet “piracy” and book sales: a field experiment. J Econ Sci Assoc (2024)