Since Nielsen BookScan published it's list of the top-selling books of 2005, a number of publishing commentators have discussed what the numbers mean (see also, here and here). I won't repeat what they've said. Here's what struck me.
Of the 200 slots on the list, only 113 are occupied by editions with a 2005 publication date. (I'm counting three of the books with 2004 dates as 2005 books, because by far most of their sales occurred in 2005.) More significant, only 60 of these 113 are hardcovers. With the exception of a few paperback originals, the other 73 are paperback or mass market editions of previously published hardcovers.
Of the 55 authors who account for these 60 books, only 11 have not had a previous bestseller. Of those, only 4 have not previously published a book.
So what? As time allows, I'll have more to say on this blog on publishing in general, but for the moment, this is simply further evidence of how imprudent it is to drop seven figures on an unknown, first-time author (such as John Twelve Hawks). There is a 99.998 percent chance that this sort of gamble won't pay off. My guess is that Doubleday is at least $500,000 in the red on Mr. Twelve Hawks' book, which, of course, was only one of many, many such gambles in 2005.