If you ever write a book (or more) you'll find yourself very interested in how it's selling. In part because good sales indicate that you did a good job writing the book, but in perhaps a larger part because your royalties are directly correlated to the number of units sold. Sure, every month or three (depending on the publisher) you'll get a royalty statement that has the precise sales figures for your title(s), but who wants to wait a month or more for such information?
I recently read a blog post by Steve Anglin on the APress Blog where Steve talked about the use of the Amazon.com sales rank as a useful metric for measuring a book's sales, liking Amazon.com's sales rank to the big board on a stock exchange:
... the Amazon.com rankings are sales rankings that can be utilized as the "share price" of a book where 1 is the best on up to the worst. The 5 star reviewer ratings are done by reviewers who could be likened to equity analysts who grade the book like a stock.
Unfortunately, it's not all inclusive. Most books that are bought and sold on AMZ are also bought and sold in other markets or on other sites, etc. whereas stocks on the NYSE are not traded on the NASDAQ. Likewise, stocks on the NASDAQ are not traded on the NYSE here in the US.
There was even once a site, AmazonScan.com, that allowed you to enter a book's Amazon.com ID (ASIN) and it would track the sales rank over time. (The site appears to no longer be up and running.)
The Amazon.com sales rank provides one metric of a book's success, but I have found it not to be a very accurate metric. While Amazon.com has a much broader audience than, say, Nerdbooks.com, the audience, I believe, still isn't as wide as the brick and mortar stores. For example, imagine that you are not the most computer savvy user. Sure, you may have used Amazon.com to buy the latest potboiler, or to buy a book from an author you're familiar with, but if you needed a book on a topic that was new to you and you wanted to learn, would you head to Amazon.com, or the Borders down the street? I'd wager the latter: you'd want to flip through the pages, browse the rows of books in a much more intuitive and easier way than Amazon.com could make possible.
Furthermore, how many people buy the latest potboilers or “best-sellers” from Amazon.com vs. buying those more obscure, audience-specific books (like books on ASP.NET)? In a recent interview, Amazon.com founder Jeff Bezos admits that “relative to the industry as a whole, we're disproportionately weighted toward harder-to-find titles.” I think this further skews the accuracy of the Amazon.com sales rank.
Finally, some emperical evidence. While I've found that there is a strong, positive correlation between actual sales and Amazon.com sales rank among my ASP/ASP.NET books, I've noticed that there's not a correlation between my ASP/ASP.NET books and my latest book, Creating Your Own Website (Using What You Already Know), a book geared toward computer newbies who want to build their own website. Create Your Own Website... has a horrid Amazon.com sales rank - 213,754 as of the time of this writing - while ASP.NET Data Web Controls Kick Start, ASP.NET Tips, Tutorials, and Code, and Teach Yourself ASP.NET in 24 Hours all have higher sales ranks but either have sold fewer copies or on pace to sell less.
I imagine the Amazon.com sales rank is a decent metric when comparing books that are very tightly aligned in audience demographic, but quickly falls apart if trying to compare books aimed at different audiences or trying to guage the success of a book whose intended audience is less likely to use Amazon.com over a brick and mortar outlet.