Here is the press release:
Here is the headline:
"Print and Digital Books Reflect Seasonal Sales Patterns According to AAP Publishers' April 2011 Report"
Here is my commentary on March's report that I posted in May:
Here is the table I posted in May, so you don't have to go extract it, with April added:
Jan 32 70
Feb ~30 90
Mar 28 69
Apr 28 73
If it isn't perfectly clear, that is, in rounded millions, AAP's reported e-book sales by month with columns by year. While you can make a case for a seasonal pattern (and the press release indicates publishers are making that case), it's not the first argument that springs to my mind.
Downloaded audiobooks had a good month year-over-year with over 20% growth.
It's hard to know how one might go about comparing the April AAP dollar numbers with Amazon's statement about unit sales for April (100 v 105, paper v ebook). If I think of something particularly clever, I may add to this post. Right now, the prospect of trying to guess average pricing and converting the dollars to units is making me feel tired.
ETA: Apparently not that tired. Also, A. is napping and R. took T. to his therapeutic riding lesson and I'm bored with the TV machine.
I took the table of April numbers and assigned each category an average unit price to convert the revenue in dollars to units. If you think that the numbers I picked are far enough off that it matters, please feel free to chime in with better suggestions.
Adult HC 111 / 15 = 7.4
Adult PB 96 / 10 = 9.6
Ebooks 73 / 8 = 9.125
Relig 48 / 9 = 5.3
Child HC 41 / 12 = 3.4
Child PB 37 / 9 = 4.1
MMPB 29 / 7 - 4.14
This gives me a total of 34 million trade excluding ebooks and 9 million ebooks. That's very different from Amazon's 100 all-paper vs. 105 ebooks.
105% of 34 is about 36. (Please don't snivel about rounding at this point. I'm barely holding it together as it is.) What would an addition 27 million units of ebooks imply at Amazon? If those were .99 ebooks, it's probably not something that a publisher should be worrying about. Much. If those were 2.99 ebooks, again, not too scary. But if Amazon's average unit price per ebook on that stuff that's not getting counted by AAP because it's from a non-member (and maybe a very nontraditional publisher) is getting up in the range of the mass market paperbacks, I think they should be getting a little concerned.
It's possible my unit difference is off. High price estimate on ebook in conjunction with low price estimate on paper books would have the effect of making paper sales look more substantial than they are (as a unit calculation). If the price estimate on ebooks is too low and the price estimate on paper books is too high, then paper sales look lower vs. ebooks than they are. Anything that is part one and part the other (consistently high estimates or consistently low) should mostly tend to cancel out.
Criticisms of the calculation (whether error in doing the math or in the assumptions) are welcome and I will try to incorporate them into the analysis.
ETA: It occurs to me that I've picked average unit numbers reflecting what I think the total revenue was, but the publisher numbers are probably revenue to the publisher NOT revenue to the publisher plus revenue to the retailer. It's not completely unfair to characterize the publisher/retailer cut as 50-50; if you want to adjust for that, the implication is that the AAP numbers represent approximately 53 million fewer ebook sales than Amazon had during the same month, vs. approximately 27 million fewer ebook sales.
ETAYA: Actually, I think the publishers should be concerned even at the $2.99 price point. That would imply they were leaving the equivalent of their total children's/YA revenue for someone else to collect instead of them.
2.99 * .5 * 53 is a shade under 80 million to the publisher. I wonder if this is how Amazon convinced the publishers to sign up for that Sunshine Deals promotion?
ETA Still more:
It looks like this guy is estimating revenue sales for ebooks at 95 million:
"When I include my estimate of non-AAP sales, ebooks are almost (no quite) quadruple MMPB sales (94.9M vs. 28.5M)." That would be 22 million revenue over the AAP figure (about 73 million). If my "missing units" is correct, it would imply a very, very low price point on the sales (under a dollar price for the customer). He explains his methodology and believes he is undershooting on the ebook estimate.
"I do an estimate on 'non-AAP' ebook sales. 10% in December 2009 and 20% in December 2010. I then just extrapolate a linear trend... I know. Two very approximate data points and then an extrapolation."
It's hard to know exactly how to incorporate this thinking into my approach (two super ridiculously simple and almost certainly wrong models go into a bar . . .).