Note: This analysis is in no way sponsored or approved by Amazon.com. I have another article for Amazon book sales compared to Barnes&Noble and Borders. Way back on October 14th, 2004 Amazon made the first major change to their ranking system that I'd seen in five years. The new system is actually more transparent than the old system. The new rank is preceded by a # and appears right on the sales page for any book. There are two main differences between the old system and the new system. First, the new system for Amazon sales rank includes sales of both Marketplace books (used and new), but not Kindle e-books. Second, the new system is based almost entirely on "what have you done for me lately." Historical sales only have a small impact on the decay rate. If your primary interest is increasing your book sales on the Internet, read my free guide to online book marketing.

The following graph is based on January, February and March data. The graph addresses books with an average Amazon sales rank between 1,000 and 1,000,000. Actual sales vary with day, week and season, and the January/February data period includes Spring semester textbook sales. The sales equivalencies for lower ranks (ie, bigger rank number getting into the millions) mean little in terms of average sales unless you can average them over years. I cannot stress enough that for any sales rank, checking the rank on Amazon twice and looking at this graph means nothing. You have to get an average rank for at least a week on stronger sellers for it to have any meaning at all. Amazon's overall share of industry book sales has grown so much in recent years that the bestsellers are disappearing from the main graph! Amazon has sold more than 7.5 million unique titles at this point, and rankings indicate a title must sell at least one copy a year to remain above a rank of two million.