On the nullity of Amazon rankings

People tell me, as Pericles Commission nears release, to ignore the book's Amazon ranking.  Which I didn't even know existed until people told me to ignore it.

So this made me find out about Amazon rankings.  As far as I can tell, the advice is right; Amazon ranking appears to be a poor implementation of a bad idea.  The concept is that every single book is stack ranked based on sales.  And not only sales, but using a projection algorithm based on buying trends.

So my humble debut historical mystery is stack ranked in the same group as, for example, the latest Twilight.  Because, you know, it makes perfect sense that teenage girls who're into emotionally dysfunctional vampires will also want to read an ancient murder mystery, right?   Also, the same people want to read the latest textbook on calculus and Shakespeare's Hamlet.

You could make this work, I suppose, if you stack ranked within genres.  So that ancient mysteries were ranked only among themselves, and dysfunctional vampires only sucked each others' blood.  But even then there's a problem because the 7th book in a successful series comes with an existing audience.  A book late in a series could totally outsell a profitable newcomer and yet still underperform on expectations.

So what does this ranking system measure?  Actually, nothing, except perhaps a profitability projection for the bookseller.

Surprisingly, it seems nobody can tell how many copies of any book have actually been sold.  The publisher knows how many have been shipped to stores, but not how many have left the shop.  Books are sold on consignment; the store could in the future return some to the warehouse.   There's a system called BookScan which does measure sales in something approaching real time, but it doesn't monitor all stores.

So for months, until there's an audit, nobody knows the sale numbers!


(I stole the blog post title from a letter written by Roger Bacon in the 13th century, entitled On The Nullity Of Magic.)

10 comments:

C. N. Nevets said...

I sigh every time I see an author investing too much into this number. Thanks for dispelling some of the mystique with your authoritative voice.

And also for adapating Roger Bacon.

Andrew said...

I believe bookstores purchase from the publisher, then return unsold books for a refund. So, it's not really on consignment.

Gary Corby said...

Hi Nevets, the voice may be authoritative, but not the experience!

Roger Bacon is one of my heroes. Have you read Doctor Mirabilis, by James Blish?

Susanna Fraser said...

I know it's futile to worry about one's Amazon ranking, or even to try to interpret it. However, as a newly published first-time author, I can also testify that checking it is addictive when it's the only thing resembling sales data you have.

Gary Corby said...

Andrew, you're totally right. That was sloppy English on my part.

Yes indeed, stores pay for copies and are reimbursed if they return unsold copies.

Gary Corby said...

Hi Susanna. Good luck with it! So far I've managed to resist looking except out of morbid curiousity.

My ranking has changed by 80,000 places since I wrote this post! Which logically means Amazon has sold 80,000 different titles in under 24 hours. Or even more logically means the system is borked.

thepoisongarden said...

I had the same advice from my publisher and, of course, ignored it.

You realise things aren't right but, I think, it serves as some sort of indicator of activity.

In the UK, a couple of the big booksellers have online charts. I don't know if they are any better but if they all show movement in the same direction that might mean they're right.

Gary Corby said...

Hi John. Fortunately I can (mostly) resist the urge.

It's weird that nobody seems to collect this data. With movies and music they know it in fine detail. Maybe it's because books take longer to read?

Sean said...

I imagine Amazon rankings are only lightly more meaningful than blog stats. :)

Gary Corby said...

Hi Sean,

It occurs to me that book sales are a bit like small cap, low liquidity investments. It might be interesting to apply some standard financial analysis tools and see what they predict.

Of course, you'd have to get the detailed sales numbers to try it.