Seasonality in Book Sales

I've been motivated by a current correspondence about Long Tail estimates for Amazon to look again at the seasonality of book sales. The best numbers I've found on the subject come from the US Census Bureau. Their report on 2004 Bookstore Sales by month, including college and specialty bookstores, yielded the following:

Jan - $2,070 billion
Feb - $1,072 billion
Mar - $1,036 billion
Apr - $0.987 billion
May - $1,071 billion
Jun - $1,159 billion
Jul - $1,128 billion
Aug - $2,043 billion
Sep - $1,493 billion
Oct - $1,032 billion
Nov - $1,044 billion
Dec - $2,089 billion
Total - $16,224 billion

The main drivers of seasonality are school textbooks and Christmas gifts. I used these ratios to adjust my Amazon Sales Rank graph, but the dynamics at Amazon are likely very different than at the brick-and-mortar stores, so I took a look at the last 4 quarters of Amazon North America media sales.

Apr to Jun 2005 - $632 Million
Jan to Mar 2005 - $699 Million
Oct to Dec 2004 - $751 Million
Jul to Sep 2004 - $502 Million

Just to see the rations, I grouped the store sales from the Census Bureau into similar quarters:

Apr to Jun 2004 - $3.22 Billion
Jan to Mar 2004 - $4.18 Billion
Oct to Dec 2004 - $4.17 Million
Jul to Sep 2004 - $4.66 Million

So what stands out from this comparison, allowing for the year to year difference? The best quarter of the year for the group including college bookstores is the Jul to Sep quarter, which is Amazon's worst quarter by far. Amazon's best quarter, Christmas, is only middle of the road for the group with the college bookstores. What I've quickly learned is that I've got to find another way to adjust Amazon sales for seasonality, because these groups just aren't apples and apples!