As Amazon celebrates its 10th anniversary, the real story isn't what the company has accomplished so far, but how the Internet pioneer promises to change the book business in the coming 5—10 years. Chances are, you haven't spent a lot of time thinking about Amazon.com's recent acquisitions of on-demand book printer BookSurge and e-book company MobiPocket. But you should. They may signal a coming transformation of the publishing business, one that includes an end to the industry's biggest problem: returns.

Since practical POD technology became available in 1998, it has offered the vision of a book business driven by demand-and-supply rather than the current consignment model. Who better to turn that vision into a reality than Amazon? The Web retailer still owns well over four million square feet of warehouse space, no small portion of which is devoted to books; it employs 9,000 people to process orders. Imagine how Amazon would benefit if it could forward orders to a printer to drop-ship books directly to customers.

Imagine, also, how publishers could benefit. One strategy might be for Amazon to print pre-sold books in its own plant. Aside from shifting printing and shipping costs from publishers to the retailer, this approach would sharply reduce the guesswork for publishers setting print runs. Given current economies of scale for large print runs of big books, it's likely publishers would, at least for the foreseeable future, continue to print books the traditional way for brick-and-mortar accounts. But even a mix of POD and traditional printing makes more sense than the current reasoning that you can make more money by printing a million copies and selling half of them than you can by printing half a million and selling all of them.

Though POD manufacturing costs are currently far higher than those of traditional long print runs, longer POD print runs—and lower unit costs—will become more common if the number of preorders on the site continues to rise, and as the technology continues to improve.

That brings us to Amazon's acquisition of the e-book retailer MobiPocket. E-books represent virtual publishing in its purest form, eliminating hard copies. Perhaps Amazon CEO Jeff Bezos has glimpsed the vision that has driven so many e-book pioneers for the last 10 years: a book distribution network consisting of nothing more than a writer, a reader and a server.

As Amazon masters these technologies and delivery systems in the coming years, perhaps becoming a publisher in its own right, it will be harder and harder for traditional publishers to support the outdated consignment model. Indeed, trade publishers may find themselves shifting to a system in which most books are pre-sold, regardless of the channel. True, the effect on brick-and-mortar bookstores could be turbulent. But could it be any more threatening than continuing to accept a 40% returns rate as normal?

With the means for eliminating returns at hand, the old system of flooding stores with copies to guarantee bestseller rankings, then taking them back and pulping them, will no longer justify the terrible losses of money, time, labor, paper and fuel, to say nothing of author initiative, literary quality and editorial pride and morale.

If this all sounds like a pretty dramatic change, it is. But if it seems far-fetched, remember that it was a only decade ago that Bezos introduced the idea that the Internet could become a mainstream sales channel for books. While Bezos hasn't revealed exactly how he's planning to use BookSurge and MobiPocket, the possibilities are clear. If Amazon, the Nine Zillion Pound Gorilla, recognizes the need—and the means—for change, the rest of us would be smart to follow its lead.