In his book Free, Chris Anderson lays out the arguments for a Free pricing strategy and gives some good examples of how it can succeed. (A list of 50 of these examples can be found at the back of the book, organized by business model). Relying primarily on these examples, Anderson strives to prove that products, and especially digital products, can be (and many should be) free because the marginal cost of providing them is low and getting lower all the time. While this sounds nice (especially to consumers), Anderson eventually points out that Free works in only limited situations. Free has a place in the economy -- alongside paid products and services.

Despite the stories of success and the evidence that Free can be profitable, Anderson fails to give the reader any real framework with which to understand these models and evaluate any further applicability. The most egregious example of this is his discussion of two-sided networks. In his introduction, Anderson boldly declares that economics has nothing to say of Free. In the following sentence, he backtracks (but only slightly) and says:

In fairness some [theories] do exist, as later research would reveal. But they were mostly obscure academic discussions of “two-sided markets” and, as we’ll see in the economics chapter, nearly forgotten theories from the nineteenth century.

Anderson’s later discussion of two-sided networks barely runs over a page. Instead, Anderson focuses on Bertrand marginal cost pricing, concluding that in competitive markets characterized by abundance, price tends to fall to marginal cost. Based on this, companies selling information products should seriously consider Free. Best yet, these companies should be the first to do so.

Anderson takes examples that run counter to this theory (e.g. Microsoft) and explains them away by stating that they fall outside the restrictions set by the definition (Microsoft does not operate in a competitive industry). While that may be true, it does not explain why Windows beat out the Mac OS in the early history of computers without charging consumers the marginal cost. (The theory actually states that there need only be 2 companies). More fundamentally, if Anderson is willing to discount this example because it falls outside the theory’s assumptions, he would need to discount any application of the theory as no market meets the strict restrictions set forth by Bertrand.

Microsoft’s success was due to a number of different factors, many of which can be understood in the context of two-sided networks. Bertrand economics, while undoubtedly useful, is too simple an explanation to be used in isolation and to the extent that Anderson recommends. Anderson’s heavy reliance upon Bertrand economics focuses too much on price and too little on network value.

Free worked in many of Anderson’s examples because it increased the size of a platform and that size has real value.

With the rise in popularity of ebooks, many publishing companies have begun to focus closely on the cannibalizing impact that ebook sales have on print sales. Sourcebooks, Inc. has even announced that it will delay the release of “Bran Hambric: The Farfield Curse” as an ebook in order to prevent any loss of hardcover sales. However, the discussion of cannibalization may be less relevant than once thought. While the introduction of the ebook format necessarily brought about some cannibalization, the market may have reached the point that the issue is no longer cannibalization, but missed sales.

The argument that ebook distribution will result in print cannibalization rests on a number of consumer behavior assumptions. In order for print cannibalization to occur, consumers would need to seek out book titles first and select the desired format second. This consumer behavior is consistent with many current and past retailing practices (both in physical stores and e-stores) where consumers browse in sections and select the format (hardcover versus paperback) only after the desired content is found. This consumer behavior is a primary reason why paperback editions are released at a later date.

However, ebook distribution challenges this behavior. Consumers who want ebooks now go to dedicated ebook sites to browse and buy. This type of purchasing behavior is possible because ebook distributors have amassed enough content to satisfy the majority of ebook consumers. Amazon, the most popular ebook distributor, has over 300,000 titles and nearly all of the titles on the New York Times Best Seller list. Individuals buying from these distributors (and especially those who have purchased a dedicated e-reading device) have chosen their desired format before choosing the content.

Given this change in consumer purchasing behavior, the argument of print cannibalization rests on an even more difficult assumption -- that any given book title has limited substitutes. Those who subscribe to this idea believe that as long as their content is not available in an ebook format, consumers will continue to purchase the print book instead of purchasing a similar title in their preferred format. However, a publishing company cannot dictate the format when consumers cannot find an author’s content. While a limited number of authors will be able to overcome this (at least to some extent) through reputation-based marketing, dictating the format may alienate an author’s readers. This alienation poses a significant risk to the long-term success of a publishing company.

There is no doubt that the introduction of a new format brings about some level of cannibalization. However, publishing companies and authors must realize that there is a difference between avoidable and unavoidable cannibalization.

This post is a response to a blog written by the folks over at The Idea Logical consulting company. The blog lays out Mike Shatzkin’s predictions for the ebook market, dividing its future into four stages: vision, establishment, transition, and the new marketplace. In the transition stage, which will occur in one to three years, Shatzkin predicts:

Large horizontal aggregators (Amazon, B&N, and the full-line bookstores that build their offerings from wholesalers) will struggle to hold onto a large and loyal customer base as the vertical web increasingly takes hold. Almost all publishers will be among the zillions of sites offering direct downloads to consumers, many through explicit verticals that sell the books of their competitors (as Macmillan’s tor.com sci-fi site, presciently, is doing today.)

Much of the logic underlying this prediction can be found in another of Shatzkin’s posts, in which he discusses horizontal versus vertical markets. In this post, Shatzkin states that on the Internet, content is no longer scarce because any individual can write something and put it on a website. The issue for publishing companies, therefore, is no longer content creation but marketing. And, according to Shatzkin, Internet marketing can only be done cost effectively inside a vertical market because the Internet is organized by niche. Thus, vertical wins; horizontal loses.

The first part of this argument deals with the relevancy of publishing companies and is outside the realm of this response. The second piece of the argument, however, deserves a closer look. After all, there are a significant number of non-vertical, non-niche platforms of the web -- e.g. Facebook, Twitter, Amazon, Craigslist, etc. And, these platforms are highly successful in terms of users, (and with the exception of Twitter, all are successful financially).

While the benefits of vertical marketing are clear, Shatzkin assumes that you cannot market to specific niches using the existing, horizontal infrastructure. Google, among others, has proven this wrong. The context in which each user accesses these horizontal platforms and the user’s ability to search through content provides a pseudo-vertical environment in which companies can market cost effectively. As others have noted, advertisers can reach customers across the entire horizontal network or can focus on any combination of highly specific vertical networks contained within the overall platform. The horizontal web platform combines content breadth with marketing focus, resulting in a superior customer experience.

This customer experience, in fact, is what Shatzkin seems to ignore in his four-stage post. The Internet has changed quite a bit -- but it has not changed the consumer’s desire to minimize search and transaction costs. These are the non-monetary costs associated with using a service and purchasing a product.

Horizontal aggregators, like Amazon, Barnes & Noble, etc., minimize these costs for users by giving them a single place to search, browse, and buy. With a horizontal aggregator, users do not need to visit multiple websites with different content in order to make a final purchasing decision (as individuals do not exist in a single, exclusive vertical) and they do not need to create and maintain separate accounts at multiple sites. Disaggregating content would create real costs for users and would correspondingly lead to a decrease in overall demand. These horizontal networks have value, and that value is tied to the number of users and the amount of content offered.

In the case of Amazon, this value manifests itself most clearly in the relationships it is able to create and maintain with consumers. Horizontal networks can provide superior relationships because the availability of more products provides additional data points with which companies can better understand the needs of an individual consumer.

The introduction and subsequent growth of ebooks has changed much in the publishing business. However, these changes have been and will continue to be guided by consumer behavior and network economics.