CNN had an interesting report yesterday on a respected but understandably obscure biology volume by Peter Lawrence called The Making of a Fly: The Genetics of Animal Design that had an asking price of $23,698,655.93 on Amazon (plus shipping). So perhaps the usually prescient Herman Melville was wrong when he wrote in Moby-Dick, “To produce a mighty book, you must choose a mighty theme. No great and enduring volume can ever be written on the flea, though many there be who have tried it.” Was the seller crazy? And was Amazon crazy for letting the seller put that price on any book, let alone a book about a fly (not a flea, but close enough)? Not exactly.
Welcome to the world of algorithmic book pricing, where many bookseller computer programs vie with each other to have alternately either the lowest price or the highest price for any given book in their inventory. Just to be clear, before I continue, here at Between the Covers, despite an inventory of hundreds of thousands of books, we prefer to determine our asking prices the old fashioned way by doing it by hand when we catalog the book. So what follows (until I get back to what we do at BTC) is based on observations of the book market, not first hand experience.
Back in the mid-1990s, when Internet bookselling of antiquarian and used books started, pricing behavior was merely an extension of the long-existing methodology of the book trade: partly knowing the book you had (including edition and condition), partly knowing the sales history of the book (from auction records, dealer catalogs, going to bookfairs, scouting, word of mouth in the book trade), partly knowing your customers (“I’ve got a fly-book collector, but he never goes above $200”), and large parts intuition. Publishers of new books set a price and, when they print it on a book, commit to it with very few exceptions. Prices of used books are understandably more fluid. And pricing of antiquarian books was (and I’ll argue still is) something of an art. Or at least a skill. Before the Internet, by definition you could not survive for long in business if you could not 1) buy books at a lower price than you could sell them, 2) understand the reasons for the difference between the two prices, and 3) build up some kind of reputation for pricing reasonably. But by the millennium there were so many used books for sale, and so many people trying to sell them (especially people who had never sold books before the Internet), that it was inevitable people would write computer programs to speed up the process.
There are two pricing shortcuts that many bookselling novices try to follow: underprice or overprice. If as a novice seller you know nothing about books, nothing about condition, nothing about editions, but you see that someone is selling an obscure book on genetics for $50 and you have a copy of the same title (without regard for edition, condition, etc.), you might price the book at $49 and so look like the best bargain, or price the book at something above $50 (say, $55) and imply that yours is a better copy or that you are a $5 better class of bookseller. The underpricing part most folks can understand right away. The overpricing part might seem a bit counter-intuitive, but it can work and there are sellers who rely on it. So, let’s say you don’t know much about books but you own a business that can find great heaps of them at low prices (I won’t name names). You don’t want to price these books yourself – if you tried you wouldn’t really know what you’re doing most of the time anyway. “Let the market adjust itself” you say, believing this is what nature intends. What do you do for your thousands of books?
You write (or more likely hire) a computer program to take the pricing shortcuts for you. Tell the program to underprice everybody on everything most of the time (what do you care, you’re selling so many cheap books that you get discount shipping rates from the postal service and make your meager book-by-book profit on the shipping), but sometimes, perhaps under a different selling name, overprice everybody as well in the hopes of fooling some buyer into thinking that your copy is special, or you know something that no one else does. And, because your computer is doing all the work, tell it to reprice the book several times a day as the book “market” changes – let it raise or lower the price of each book on an hourly basis.
The most obvious problem with these automatic pricing programs is that they are widely used and so compete with each other, creating crazy prices such as the example found by CNN. Do you really think there is a person out there deliberately pricing a book at some dollars and .93 cents? Of course not. Computer programs react to an oversimplified look at what they perceive to be the book market (usually this means simply assessing other copies available on Amazon and stopping there) and adjust the clients’ prices accordingly. Often this results in books costing a penny, and then what the customer is really paying for is the shipping (again, this is where the so-called penny sellers make their profit so it only seems fitting). In order to sell a large volume of inexpensive books (large volumes meaning many hundreds or thousands of books a day) you wouldn’t last 2 seconds without such a pricing program. And so, many used booksellers utilize them. Unfortunately it isn’t just huge, book churning behemoth companies that use the programs – many small, used booksellers with inventories of only a few hundred titles try to use them too. These small-inventory, used booksellers are under the impression that using these pricing programs is necessary to sell books (“because your competition is” – Amazon recently used this line with us!).
So many used books are sold this way that when it comes to many books, there really is no one steering the ship of market values (not Amazon, not the bulk of sellers, not the computer programmers because they don’t know books and can’t take into account things like edition and condition, and not the computer programs themselves because they are just blindly going about the business of following their very simple instructions). Most book prices take a rapid race to the bottom. Sometimes they take a weird jump to an extremely ungainly mountaintop. Examples such as the 23 million dollar fly book demonstrate the fallacy that algorithm driven markets self-adjust without oversight. I won’t make any statements about algorithm driven financial markets and their problems because I’m not the least bit qualified to do so. But I understand crazy computer-driven pricing happens there too.
Of course, there is an alternative. It helps to actually look at the books and know something about them. Pricing your books by hand, as we do at Between the Covers, certainly has its drawbacks. Obviously, it is slow and labor intensive. Our low price is $10 because we can’t afford to sell books under that price given the time we put into them. So, we try very hard not to offer books that aren’t worth at least $10. Another disadvantage is that because we price by hand but have hundreds of thousands of books, it isn’t that hard for these programs to come along and underprice us on the title, if not on the edition and condition. We revisit our prices as we can on an on-going basis, but we’re not going to let a computer do it for us.
But pricing books by hand, especially with genuinely rare or unique material, is one of the fun parts of the bookselling tradition. When it comes time to set a price on a manuscript, or archive, or truly rare antiquarian book, we have spirited discussions in the office arguing on behalf of one price or another based on the merits of the object itself and the sales history, sometimes going back many decades, of any comparable material we can identify. Back in 1998 we offered, in our first Classic Book Cards set, a pristine copy of The Whale, the true first edition of Moby-Dick, for $150,000. You remember that one? “Call me Ishmael” and “No great and enduring volume can ever be written on the flea.” At the time it seemed to many an ambitious price. Needless to say, the book sold in no time, no copy anywhere near as nice has been seen since, and I know at least one collector who, to this day, laments not getting to the phone fast enough. I wouldn’t be telling you this story if we’d priced it $149,655.93.
– Dan Gregory