To judge from the Da Vinci Codes and Harry Potters stacked high in bookstores everywhere, publishing is a classic hit-driven business. One megaseller is enough to make any publisher's year, possibly his career. The problem is that only a tiny fraction of books are hits, and predicting which they will be is nearly impossible. Unless you're willing to pay a fortune for the work of a celebrity author or proven bestselling writer (in which case all those sales are necessary just to win back your investment), the life of a hit chaser is one of frequent disappointment and constant uncertainty.

Here's the reality of the book industry: in 2004, 950,000 titles out of the 1.2 million tracked by Nielsen Bookscan sold fewer than 99 copies. Another 200,000 sold fewer than 1,000 copies. Only 25,000 sold more than 5,000 copies. The average book in America sells about 500 copies. Those blockbusters are a minute anomaly: only 10 books sold more than a million copies last year, and fewer than 500 sold more than 100,000.

So are all the rest of the books simply failed blockbusters? Of course not. As the book industry has known for decades, there's virtue in niches—books that aren't for everyone, but really thrill those they are for. The trick is finding a way to make a business in niches, rejoicing in the rare blockbuster if it comes, but not having to depend on it.

I call this mass of niches the Long Tail. It's the oft-neglected tail of the demand curve, the land of the relatively low-selling items. Publishers tend to focus on the head of the curve, where the hits are, and forget most of the rest. This isn't just myopia—it's demanded by the constraints of the physical world: limited shelf space, limited channels, limited screens. But now distribution methods have virtually unlimited capacity, from the digital products on iTunes to the Sears-on-steroids combination of online catalogues and overnight delivery pioneered by Amazon.com. Thanks to these extraordinarily efficient channels, mass markets are shifting to millions of niches.

What does that mean for books? After all, there are already books in every conceivable niche and nearly 200,000 new ones are published every year. The answer is to drive more demand to those niches by making such relatively narrow-interest titles more available and easier to find.

Some of this is already happening. The extraordinary transformation of the used book market over the past half-decade is an example of just how much business a little innovation can spur. Today, used books are the fastest growing part of the industry and make up nearly 10% of sales.

Another transformative technology is print-on-demand. The costs of obtaining the rights and reformatting books to make them available are still too high for many older niche titles, but the technology is at the core of the booming self-publishing industry. Just as in music and, increasingly, video, the tools of production are being democratized.

Scanning and search technology also promises to unlock unseen demand for niche books. Today, there are about 32 million unique titles in America's collective libraries. Only about two million of them—6%—are in print. Perhaps another 20%, mostly the oldest books, are in the public domain and legally free and clear to be scanned and put online. The rest—nearly three quarters of the knowledge, culture and history that lies in books—remains "out of print" but still under copyright. If the publishing industry can create a simple, cheap way to make those books available to be scanned and searched while protecting the interests of authors and publishers, we'll all be better off.

Finally, there's the favorite of futurists everywhere: the e-book. After years of disappointments, the technology of high-resolution/low power/paperlike readable screens is finally starting to arrive. Sony's latest e-book reader is a hardware marvel, and if Sony can avoid forcing clunky proprietary standards on consumers, people will be able to buy books the way they buy music from services like iTunes. The ability to have access to practically everything ever recorded is transforming the music industry, revealing that music consumers were far less hit-obsessed than we'd thought. Someday soon, the book industry may be transformed, too.