Why do bestseller lists differ from source to source?

August 24, 2004 2:44 PM

What makes a bestseller? Something I've never understood is why different publications have different bestseller list. I mean, either the book sells X copies or it doesn't, right? How can it be on one bestseller list and not another? The only thing I can think of is that the NYT list is referring to sales in NY and WaPo concerns itself with it's own state, etc. But that doesn't make much sense. {more}

Second, is there a source for how many copies it takes in each country to make something a bestseller? Like, what's it take to be a bestseller in Canada? The USA? The UK, etc.

posted by dobbs to Writing & Language (7 answers total)

Here's a good article. It's not the most sophisticated system: "The New York Times editors select 36 titles they feel might be best-selling titles for the week and poll some 3,000 bookstores across the U.S. The stores are asked to fill in the number of books sold next to each title and to write in fast-moving books not on the list. "

Your question implies there is a national system to track book sales, like SoundScan for CDs. That may not be a valid assumption now, I don't know, and surely wasn't inexistence when these lists were founded.
posted by smackfu at 3:09 PM on August 24, 2004

It used to be much more of a black art (here's an old article that talks about the past state of such lists).

Over the past couple of years, though, while all the big newspaper lists still use a mix of different sources for their rankings, more and more, they use Bookscan statistics, which are becoming an industry-wide reporting mechanism for book sales.

There was a big shift about 2 or 3 years ago in not only book publishing, but CD publishing, as they started to use actual sales data that was being collated from most outlets. Most of the bestseller lists had to admit that they were deeply inaccurate, and have all also gotten much more similar as they all report from the same body of data.

There are still a couple of reasons why they differ. For one, like I said, they all still pull in other data from independent booksellers, etc., and each paper has its own sources. More importantly, they've all also reserved the right to fudge the lists if they see fit. When different books get parked at the top (like the Harry Potter books being parked at the top of the NY fiction list), they'll shove things around.
posted by LairBob at 3:10 PM on August 24, 2004

From the NY Times Book Section:

Rankings reflect sales at almost 4,000 bookstores plus wholesalers serving 50,000 other retailers (gift shops, department stores, newsstands, supermarkets), statistically weighted to represent all such outlets nationwide.

A good article in Slate offers a better view of things, though, and points out that the NY Times list might not be as accurate as it seems...

Very interesting question dobbs, thanks for posting it!
posted by NotMyselfRightNow at 3:13 PM on August 24, 2004

It's mojo. One of the questions we saw a lot on Google Answers were variants on "How can I tell how many copies of such-and-such a book have sold?" from authors who thought maybe their publishers were jerking them around, and other curious folks. The quick answer is "You can't"

Not only is it hard to get one number of how many copies sold even if you're the author [because of things like returns, freebies, destroyed copies, etc] but often publishers won't release this information because, at some level, how well a book is doing is sort of a trade secret. Sure the next copy of Famous Author's book will say "Famous Author's last book sold 100,000 copies!!" but try to figure out how a book that's not on the best-seller list is selling, or exactly how many copies of any particular best seller have sold and it's like pulling teeth.

This guy has a weird little page on how to calculate how many copies of your book have sold by looking at your Amazon ranking. Otherwise, as people move to more automated distribution systems for booksales, these numbers will get more and more ascertainable, but perhaps not more and more public.
posted by jessamyn at 4:07 PM on August 24, 2004

Although it does not answer your question in any way, in the interest of knowledge and information sharing: The Washington Post does not serve "its own state." Washington, DC, is a city that is not in a state; it's in an independent district. I think Canberra, Australia, might share a similar trait.

The Washington Post is also widely read in the DC suburbs, which are in Maryland and Virginia, and to some degree in West Virginia.

If the Washington Post had a non-national best-selller list, I imagine the paper would obtain it by surveying regional bookstores.
posted by croutonsupafreak at 8:21 PM on August 24, 2004

Response by poster: Thanks, all.

(And thanks, croutonsupafreak, for the geography lesson. As I was typing "state" I realized I didn't know what the hell it should say. They don't teach us that in Canadian school. :)
posted by dobbs at 9:36 PM on August 24, 2004

Corollary to all the above: blurbs referring to just "bestselling author" or "bestseller" (without qualifying which bestseller list) are meaningless. There are umpty zillion bestseller lists, offering lots of opportunity to pick your list carefully to make that statement.
posted by Zed_Lopez at 10:22 PM on August 24, 2004