Last month Google announced it had reached an agreement with US authors and publishers that will allow Google to scan books, keep a copy of the scan, and allow users to search for phrases within the scanned books. The agreement further allows Google to sell downloads of out-of-print books and give the money to the author and publisher (unless the author and publisher specifically opt out.) Public-domain books will still be free and in-print books will not be sold by Google.
Unless the copyright holder logs into Google and sets the price Google is free to decide the price using an algorithm to 'maximize revenue for each rightsholder'. It will do this by dividing books into price categories called bins with prices of $1.99, $2.99, $3.99, $4.99, $5.99, $6.99, $7.99, $8.99, $9.99, $14.99, $19.99, and $29.99. When Google starts selling books it will somewhat randomly assign books to various bins. Google aims to eventually put similar books in similar price bins, and to move popular books to the more expensive bins and unpopular books to the cheaper bins. No books will cost less than $2 or more than $30 although Google reserves the right to add more bins.
This is all explained in the settlement agreement [PDF].
Setting the highest price at $30 seems insanely low. For example, right now I am lusting after Robert Göbl's 1973 German-language classic Ostkeltischer Typen Atlas (Atlas of Eastern Celtic coin types). I've found a used copy for $150. I would pay $75 to download the work. To maximize the revenue for the publisher (Klinkhardt and Biermann, a non-US publisher, so this deal doesn't apply to them) Google has to estimate how many people will buy at each price point. Almost no one wants a book in German on Eastern Celtic coins, so perhaps the market is five people no matter if the price is $2, $10, $30, or $100.
As a thought experiment, consider the imaginary book Humorous sayings of Bill Gates' Grandmother. To maximize revenue one could set the price at $50,000 and sell one copy. At $100 perhaps two copies would sell.
That sounds pretty rediculous but have you seen the pricing for back issues of medical journals these days? Or consider Sebastian Heath's example, Brickstamps of Constantinople, with a list price of $750. A human priced it! WorldCat reports 58 libraries buying it at that price. Perhaps $750 is the true revenue-maximizing price, and at $125 only a few more libraries would acquire it.
It will be interesting to see what titles fly off Google's virtual shelves at the $30 price-point. It will also be interesting to see where Google's algorithm prices books with few readers but insanely dedicated ones. Books about expensive obscure collectibles (ancient coins, knives, etc) fall into this category, as do books on stage magic.
Pagination of physical specimens and CSV downloads - I have made some minor modifications to the coin type pages in OCRE, CRRO, etc. A relatively small number of types across these corpora have more than 100 ...
3 hours ago