Tor's sales average also indicates the boycott is failing

Over on File 770 Mike Glyer wonders if the numbers I presented about the Tor boycott failing actually show the opposite. It's a good question to ask, especially if you only look at a week or two of Tor's sales.

However, comparing a single week's sales to another week's sales can be extremely misleading, especially since book sales fluctuate greatly on a week to week basis. 

A better way to look at this is to establish an average for Tor's sales prior to the boycott, which is what I tried to show by tracking five weeks of sales prior to the boycott's start on June 19. This five-week sales average was 1652 copies per week for the sample of books I tracked.

In the graph below, this pre-boycott sales average is illustrated by the yellow line. The red line tracks the week by week total sales of the books I tracked. As you can see, there was a dip in the last two weeks of the boycott, but the dip has yet to go below the average set in the previous five weeks. This is one of the reasons I said the boycott had failed.

Red: Total Sales per week
Yellow: Average sales for five weeks prior to boycott
Blue: Weekly sales without including Ender's Games sales

But looking only at total sales doesn't give the full picture of Tor's sales because the books in my sample were published in different formats, from expensive hardcovers to trade paperbacks to mass market paperbacks. If you want to determine if Tor's sales are truly falling, lumping all these different formats in together for a cumulative sales number doesn't strike me as the best way to go. After all, is it better for Tor to sell one $25 hardcover or one $6 mass market paperback? I'm guessing Tor would go with the hardcover sale.

In addition, the book which dropped the most during the two week boycott was Ender's Game by Orson Scott Card, a strange title for the boycott to focus on since the author's politics are in line with many of the people calling for the boycott. 

It's because of variables such as these that I looked for overall patterns in Tor's sales with my analysis of their sales numbers. If the boycott was successful, you'd expect to see a sharp drop in sales across all titles, which is not what we're seeing here. Instead, what the numbers appear to show is random fluctuation which has yet to dip below the average for Tor's sales prior to the boycott.  In short, the pattern of sales we saw prior to the boycott is the pattern of sales we're seeing during the boycott.

Of course, it's possible the slight drop in sales the last two weeks will continue or increase. But that will only be evident after another few weeks of sales data come in.