Friday, April 11, 2008


(Although the following appears with my name on it, ths is actually a guest-post by another IAG member: Michael S. Katz is an attorney, editor-in-chief of Strider Nolan Publishing, board member of the Independent Authors’ Guild, and author of the comedy novel Shalom On The Range.) recently announced a new policy requiring all Print On Demand authors to use Amazon’s own printing company, Booksurge, in order to be sold through Amazon. Many POD authors and publishers are understandably upset by this, as this can only serve to cost the authors money, and cost the printing companies business. But in terms of Amazon’s market share, how much business are we actually talking about?


Sales of books totaled $2 billion in 2000, at which time on-line sales made up between 7.5% and 10% of that total.1 Amazon and now account for more than 85% of online book sales.3 Recent data shows that Amazon’s book sales are approximately four times that of,4 and Amazon has a 70% share of the Internet book market, so this translates into a 15 to 17.5% market share for

Amazon’s total sales in 2006 were $4.63 billion, but this includes books, music, and various other items, including a lot of high-end electronics, jewelry, and the like. Barnes & Noble actually outsold them at $4.68 billion (and they were basically limited to books, music and movies), but their on-line presence had only $477 million in sales. Why are people flocking to Amazon over


A lot of it has to do with programming. Amazon has a reputation for being the best at tracking customer habits, having collected information longer and used it more proactively. Over the years they have collected detailed information about what its customers buy, considered buying, browsed for but never bought, recommended to others, or even wished someone would buy them.10 Amazon uses this information to calculate recommendations that boost sales.

In the entertainment industry, recommendations are a remarkably efficient form of marketing, as they enable films, music and books to more easily find the right audience.9 For example, the book Touching the Void, a tale of a mountain-climbing tragedy, was released in 1988 to good reviews but modest success. In 1998, the book Into Thin Air, about another mountain-climbing tragedy, was released and became a bestseller. All of a sudden, people began buying the older book again. Touching the Void began to be displayed side by side with Into Thin Air, and actually wound up outselling the newer book. How did this happen? Chris Anderson, author of The Long Tail, attributes this to recommendations. Amazon’s programs note buying patterns and suggest similar books to readers. Some people follow the suggestion, enjoy the book, and post excellent reviews. These purchases and reviews lead to more sales, more recommendations, and the cycle continues.9

Readers’ reviews also stimulate sales, although moreso on Amazon than One study (Chevalier and Mayzlin) examined how sales on both sites correlated with number of reviews and customers’ ratings.12 They determined that a good review will increase the number of books sold, although with much greater effect on Amazon than A bad review has a greater effect than a good one, based on the assumption that many 5-star reviews are believed to be “planted,” whereas 1-star reviews are seen as more legitimate.12


How do prices compare between the big two? A study (Chevalier and Goolsbee) collected Amazon and data for 18,000 different books during three different weeks in 2001. They determined that there was significant price sensitivity for online book purchases at both sites. But the demand at was much more price sensitive—both to its own prices and to Amazon’s prices—than at Amazon.4

A one percent increase in a book’s price at Amazon reduced sales by about 0.5 percent at Amazon but raised sales at by 3.5 percent, implying that (based on the 4-to-1 ratio in sales) every customer lost by Amazon instead bought the book at Conversely, raising prices by one percent at reduced sales about 4 percent but increased sales at Amazon by only about 0.2 percent.4 Therefore, a customer lost by Amazon would usually wind up buying the book at, whereas a customer lost by would not necessarily go to Amazon. If keeps its prices right, they can steal away a lot of Amazon traffic.


So how much traffic are POD authors talking about? Amazon’s book ranking system can be used to estimate the number of copies a book sells.

47.9% of Amazon’s sales consisted of titles ranked better than (under) 40,000. 39.2% of their sales were books ranked between 40,000 and 100,000.5 Titles ranked between 100,000 and 200,000 accounted for 7.3% of sales, while titles ranked from 200,000 to 300,000 accounted for only 4.6% of sales.5 Anything above that accounts for only 1% of sales.

Researchers at MIT (Brynjolfsson, Yu and Smith) studied publisher-provided data of one publisher’s weekly sales for 321 titles, and compared the figures to Amazon’s sales rankings for the same week. The observed weekly sales of these books ranged from 1 to 481 copies and the observed weekly rankings ranged from 238 to 961,367.5 Morris Rosenthal of Foner Books also analyzed performance based on a brand new book he published. 11 Combining the information culled from both studies, if a book is ranked 100,000 you’re looking at selling about 1 copy per day. At a ranking of 30,000 it’s averaging between 1 and 2 copies per day. The 10,000 ranking calculates to 2 copies a day. The 1,000 ranking is estimated at 11 sales that day.11 A book with a rank of 10 is estimated to get 700 sales a day.5

Keep in mind that a ranking at any single point in time is not indicative of actual sales. Selling two copies of a title, regardless of whether it has ever sold before, will propel it into the top 50,000 for at least a few hours. If the same book otherwise sells very rarely, or never, it will drop 100,000 rankings the next day, 400,000 rankings over the course of the week, another 200,000 rankings the next week, and so on. Eventually it will hover around 2,000,000.11

We can estimate sales figures for the majority of POD books using these rankings. Not to be insulting to anyone, but odds are that most POD books will fall within the over-100,000 rankings, which means less than one copy a day while the ranking stays close to 100,000. So it’s reasonable to assume that most POD books are selling very few copies on

Another calculation we can do is by using market shares. As of a few years ago, 2 billion books were sold overall.5 On-line sales accounted for less than 10% of the book sales market, so:

10% x 2 billion books = 200 million books sold on-line.

Since 70% of all books sold on-line were sold by Amazon,

70% x 200 million books = 140 million books sold by Amazon.

As previously mentioned, 13% of books sold by Amazon were ranked worse than (higher than) 100,000. This amounts to:

13% x 140 million = 12 million books sold ranked above 100,000.

Divide this by the number of titles that are available (there are more than 2.3 million books in print and many more that are out of print but still sold on Amazon5), and the average number of books sold by a POD author on Amazon are relatively small potatoes.


If sales of POD books are so small, why should any company bother to sell them at all? Because the sales add up and can bring in big money.

With no shelf space to pay for, no exterior costs, and minimal distribution fees, all books are equally worthy of being carried.8 Internet retailers have a nearly unlimited inventory, thanks to centralized warehouses and drop-shipping agreements with distributors.

And if 13% of Amazon’s book sales are in more obscure titles (ranked over 100,000), that’s still millions of dollars in sales. This is known as the long tail, referring to items available for sale that may not sell in large individual quantities but will still add up to some serious financing. The long tail has been estimated at 20%-30% of book sales.9, 11 Not a segment that would be economically wise to ignore.


Let’s face it. The big publishing companies are going to advertise books written by best-selling authors to the hilt, and people are going to continue to buy those books based on name recognition alone—even though they are often books that are ghostwritten. But a bookseller who will do right by the lesser-known authors whose books comprise the long tail can still make a pretty penny.

So what is the correct response to Amazon’s new policy of forcing POD authors to use Amazon’s printing company? Start pushing people toward Barnes & Noble and hopefully will take notice. It couldn’t hurt if they were to use their brick-and-mortar stores to draw attention to obscure titles. Amazon can’t really do book signings, now can they? Barnes & Noble could do multiple author signings to increase reader interest, like theme nights, and they’d sell a lot of coffee in the process.

Furthermore, this isn’t a two-horse race. There are lots of other on-line stores with reputations for excellent service. If competitors can make inroads in price, selection, service and other benefits (such as reviews and recommendations), they can make inroads into Amazon’s sales. If Amazon is going to put the pinch on POD books, then another retailer stands to make a lot of money if they do right by this segment. This is especially so were Amazon to stop carrying titles that would then only be available elsewhere.

(Well, you know he's a lawyer - so here's all the references and footnotes! - CH)

1. American Booksellers Association, Industry News, “Overall Book Sales Up Slightly for First Six Months of ‘01,” November 1, 2001,


2. Clay, Karen, Ramayya Krishnan, and Eric Wolff, “Pricing Strategies on the Web: Evidence from the Online Book Industry” (November 19, 2000).

3. Nielsen Net Ratings.

4. Chevalier, Judith and Goolsbee, Austan, “Measuring prices and price competition online: Amazon and Barnes and Noble” (April 2003).

5. Brynjolfsson, Erik, Jeffrey Hu, Michael D. Smith, “Consumer Surplus in the Digital Economy: Estimating the Value of Increased Product Variety at Online Booksellers” Management Science, November 2003.

6. Brynjolfsson, Erik, Michael D. Smith, Frictionless Commerce? A Comparison of Internet and Conventional Retailers (Management Science Vol. 46, No. 4, April 2000 pp. 563–585).

7. Smith, Michael D. and Erik Brynjolfsson, “Consumer Decision-making at an Internet Shopbot” (July 23, 2001).

8. Italie, Hilltel, "Amazon's Bottom 10: Not Exactly Page Turners," Chicago Sun-Times, August 17, 2001.

9. Anderson, Chris, “The Long Tail,” Wired (2004).

10. Linn, Allison, “ Knows, Predicts Shopping Habits,” Associated Press, March 28, 2000.

11. Rosenthal, Morris, and

12. Chevalier, Judith and Dina Mayzlin, “The Effect of Word of Mouth on Sales: Online Book Reviews,” September 2005.

13. Rosenthal, Morris,

No comments: