Metadata has been a buzz topic in the publishing industry lately, and as a metadata evangelist, I think this is great. In practice, however, the pleas of people like me often fall on deaf ears. Partly that's because we've failed to show how better metadata can increase publishers' revenues, and partly it's because we often advocate building an ideal world of metadata. But overhauling your organization isn't necessary.

In fact, small changes can make a big impact. It's a strategy I like to call the five degrees rule: make your changes five degrees at a time. Focus on very basic metadata practices to start. Encourage both marketing and editorial department input on metadata. Most importantly, get the data right for all your books, and you'll find that small steps can lead to big gains.

As CIO of BookSwim.com, a Netflix-like book rental service, I am a massive consumer of book metadata. What I've noticed so far is that the data I receive about bestselling books is usually great—descriptions are clean, short, and describe the content well, and the books are tagged with three or four well-targeted categories. But for every bestseller that has quality data, there are hundreds of other books—perhaps thousands—that have too little or too much data. Given the amount of time it takes to create a book, from writing to designing and marketing, is it so crazy to ask for a few hours more to ensure good metadata?

We know that good metadata can drive sales. What we sometimes forget is that a lack of good metadata is a sales killer. Look at it this way: if your book is shelved in the wrong location in a bookstore, it has little chance of selling. Online retailers, meanwhile, are becoming increasingly sophisticated, monitoring sales and customer data to create recommendation algorithms to push sales. Without good data, your books will either be dropped from those recommendations (as they aren't bought) or they will be incorrectly recommended (and therefore never bought).

Ensuring good data doesn't have to be a major operation. While I believe all metadata is important, and the better the metadata, the better the sales, you can start by simply ensuring the essential, most basic elements: title, author, description, page count (physical)/word count (e-book), reading level, and categories.

A one-sentence description regurgitating a book's category and author won't do anything. But a book with 25 categories will only get the book lost, as bookstores fail to find the right place to stock it, and Amazon buries it for lack of relevance.

When Little, Brown categorized Twilight, did it chose "vampires," which weren't the bestselling shoo-ins they are now, or was it just another angst-ridden teen romance? I don't know, but this brings us to another benefit of good data: analyzing trends.

Good data can help you find winners not only among your new acquisitions, but also among your backlist. While it may be easy to spot huge trends like vampires, niche trends are harder to see. But a five-degree shift in your data generation can make trends more visible. Imagine the impact on your bottom line if you could turn 5% of your losing books into winners.

Marketing departments are great at selling books—but we can't expect marketing to completely understand every book they need to sell. Quality data, however, can enable them to answer difficult questions and spot hidden trends. But this takes collaboration. Just as marketing has input into titles and covers, so should they have input into a book's description and categories. When marketing has a neat, single paragraph description, two to four quality categories, and a brief history of an author's work, they can spend more time generating sales.

Yes, there is effort involved, but no, you do not need to overhaul your organization. A five-degree shift in your metadata practices can pay off quickly. With minimal training, you can begin generating better data in three months, be analyzing trends within six months, see sales improvements in the first year, and enjoy a solid feedback loop in the second.

Industry estimates suggest that as many as 70% of books never earn out their advances. Some of them may just be bad books. But how many of them are simply never found? With so many new books being released every year, focusing on better data can ensure that your books show up where your readers are.

Nick Ruffilo is chief information officer of BookSwim.com.