How do Internet markets for books create value for consumers? In the late 1990s, the answer probably would have been lower prices—the result of a reduction in operating costs and an increase in competition. However, today a larger source of consumer value has become apparent: the long tail, which describes the sale—made possible by online retailing—of small numbers of many different books, rather than the traditional approach, in which money is made on a few blockbusters.

Our early 2000 economic research showed that Internet consumers gain nearly 10 times more value for their money from finding just the right book to meet their needs, thanks to the long tail, than they do from getting lower prices on bestsellers that they could find at the mall. The value of the long tail has grown over time and is driven by the increased availability, improved discovery, and improved promotion of books sold online.

These are big changes, and not just for booksellers. For nearly 100 years, each of the major copyright industries—publishing, music, and film—has been dominated by a small number of big firms. These firms maintained power by controlling three key scarce resources, the financial and technical know-how necessary to make content, the physical channels necessary to distribute it, and an artificial scarcity in how consumers access their content, made possible by copyright enforcement.

None of the assets are as scarce as they once were, however. Technological change has democratized the production of content, long-tail markets allow everything to be distributed, and digital piracy creates a world where it is nearly impossible to control consumers’ access to movies, shows, music, and books. What is scarce today—and what major firms need to compete for—is customer attention, and a detailed understanding of customer preferences.

The problem is, to capture this value, you need Big Data—the sort that Amazon collects about its customers. Unfortunately, as is well-known in the industry, Amazon doesn’t share any customer data with its “partners” in the publishing industry, or any other industry for that matter.

So if you’re a traditional publishing firm, what can you do? Economic studies have provided an answer: start selling books in bundles. These studies have shown that selling content in large-scale bundles is much more efficient than selling the same content separately, and that by setting the right price for the bundle, firms can make more money. Bundling also creates significant economies of scale, which in the extreme can lead to a single “winner-takes-all” outcome for the company with the largest bundle. Not only that: the bundler is in a perfect position to learn more about customer preferences.

Think of the amount of information Netflix is able to collect about its individual consumers’ behaviors and preferences: it is able to note every show a consumer watches, at what points in those shows that they pause, what scenes they watch over and over again, and even on what device and at what times they are mostly likely to watch them. Spotify can record similar data for its consumers, and both services allow consumers to explore content they might not have tried if it were sold outside the bundle. This, in turn, generates value for content creators, who can use these services to find new audiences and new fans.

No dominant firms bundle books. Google (Oyster) and Amazon (Kindle Unlimited) are trying to create such bundles, but as of yet they just don’t have the sort of first-run popular content that can be found on Netflix and Spotify. We believe this creates a huge opportunity for major publishers, who have, thus far, ceded control of data about consumer attention and preferences to Amazon and other Internet companies that they have mistakenly deemed to be mere “distributors.”

Our message is both hopeful and simple. Publishers can invest in a platform that provides direct access to customer behavior and the ability to promote content directly to the right audience. Such a service will allow publishers to continue to create direct connections between the people who matter most: authors and their readers.

Michael D. Smith and Rahul Telang are professors of information systems at Carnegie Mellon University and are coauthors of Streaming, Sharing, Stealing: Big Data and the Future of Entertainment (MIT, Sept.). They codirect CMU’s Initiative for Digital Entertainment Analytics.