Numbers for Book Sales and Nielsen Bookscan

I've written a lot about publishing industry numbers and how to estimate title sales using Amazon sales ranks. The one program I've never had direct access to is Nielsen Bookscan, due to the price. I have seen Nielsen reports thanks to low friends in high places, but I've never systematically used Nielsen Bookscan for research as practically everybody else with an interest in publishing numbers does. Nielsen aggregates results from the biggest North American book retailers, including Amazon, Barnes & Noble and Borders, but they lack data from the mass merchandisers like Walmart and supermarkets. Bookscan results are usually given as representing about two thirds of the industry (65% to 70%), but this obviously varies by title. For bestsellers that are widely stocked in mass merchandisers, you'd expect them to be way off, but for books that are sold almost exclusively through Amazon and the chains, which represent a huge part of the trade, they should be pretty spot on.

In an Op-Ed published in the New York Times last week, Tim O'Reilly pointed out that according to Nielsen Bookscan "only 2 percent of the 1.2 million unique titles sold in 2004 had sales of more than 5,000 copies." That amounts to 24,000 titles selling more than 5,000 copies a year. To put that into context, Amazon lists over 24,000 titles published by just one giant trade, McGraw-Hill, in the last five years. Obviously, some of those books are out of print and most of the rest aren't selling over 5,000 copies a year, but it should give you an idea about the relative smallness of the pie. According to Nielsen, they log sales of over 300,000 individual titles in a typical week, but by the time you get to the end of the year, only 1,200,000 unique titles have been sold. Those 900,000 additional titles that have to get smeared out over 52 weeks can't be averaging more than a few sales a year.

If you make a dollar or two a book as the author of a moderately successful title, selling a few thousand books isn't going to change your financial life. If you make five or ten dollars a book as a publisher, you might earn a living with a few average books, or get rich with a few hundred. Years ago, before print on demand publishing added its contribution to the explosion of new titles, I worked out from various sources that the average trade published book sold 2,000 copies. If anything, I suspect that number may have gone down. Bookscan data gives acquisitions editors a quick and painless (or brainless) way to estimate the market size for a particular title. As a self publisher, if you don't go through a similar exercise using whatever data you can, you're just kidding yourself about being in the publishing business.

Another fun number that I can't credit because I haven't a clue where I first saw it, is that Barnes & Noble and Borders turn over their inventory around twice a year. Just imagine, of the 100,000 plus titles in your local superstore, a large proportion won't even sell one copy before getting returned to the publisher, another big chunk will sell at such a slow rate across the chain that they get dropped within a few months, and of the remaining books, the average title will sit on the shelf for six months. Barnes & Noble and Borders have around 1300 superstores between them that represent the majority of the retail bookstore trade, so two copies per store a year only gets you to 2,600 books for a success! If we intentionally overestimate and say that the two chains only represent half of the bookstore market and that a given book is actually stocked everywhere else (which isn't possible due to bookstore size limitations), it would give us just over the magic 5,000 number.