Another of the book’s signal characteristics is that it generally discusses openness (particular with respect to vertical integration) in laudatory terms (pp 111, 195, 292–93, 305–06). Indeed, the book was supposed to be named Open until Knopf published Andre Agassi’s autobiography under that title the previous year.183 At some moments, however, Wu seems more ambivalent. Open media tends to fragment rather than unite a nation (pp 214–15). Moguls step in to centralize an industry when people are unhappy with the quality of content or the reliability of a service; indeed, the arrival of the centralizing mogul “heralds a golden age in the life of the new technology” (p 10). In addition, “the closing is driven by a hunger for quality and scale—the desire to improve, even perfect the medium and realize its full potential, which is limited by openness, for all its virtues” (p 78). Wu credits the new (closed) radio industry for “creating a broad listenership for quality programming” (p 84). Centralizing innovation improves coordination and reduces waste (pp 110–11, 306). Wu singles out for special praise the locked-down vision of computing embodied in the Apple Macintosh, iPod, iPad, and iPhone, which delivered unrivaled functionality and a seamless user experience (pp 278, 292). Indeed, the most casual review of the contemporary economy reveals a wide range of practices that completely span the spectrum.
In short, the book candidly recognizes that embracing openness necessarily involves a tradeoff. In so doing, it naturally invites inquiry into whether conditions exist that might tip the balance the other way. Fortunately, the scholarly literature on standardization and modularity has advanced analytical frameworks that can provide some traction on this question. Consider first the theory of optimal standardization.184 Standardization is largely a function of the heterogeneity of consumer preferences. If everyone wants the same thing, one can satisfy all of their preferences with a single network optimized in the way that everyone wants. As what people want becomes more diverse, one would naturally expect the services that the network is offering to become more diverse in response.185 If that is the case, shifts away from the way the current network is designed may represent nothing more than the network’s attempt to evolve to meet consumer demand. From this perspective, experiments with new standards should be regarded not necessarily as anticompetitive, but as a potential sign of a healthy environment for innovation.186 Indeed, if demand is sufficiently heterogeneous, firms generally avoid me-too strategies, preferring to pursue approaches that distinguish them from their competitors. When this is the case, industries are more likely to reflect a mix of business strategies rather than uniform tactics. Thus, the fact that one firm has adopted an open architecture may make it more likely that at least some of its competitors will adopt a closed one.187
Further insights emerge from modularity theory,188 which suggests that modular architectures are not appropriate if the tasks constituting the overall process exhibit too many interdependencies.189 In addition, changes in the technological environment may cause a previously modular architecture to break down.190 A growing number of theorists have recognized that the Internet’s layered architecture establishes a form of modularity that allows each actor to optimize its own behavior locally. The problem is that when these individually rational decisions combine and interact with one another, they can lead to aggregate behavior that is suboptimal.191
Rather than extolling the virtues of open architectures in the abstract, these frameworks provide a basis for a better understanding of both the benefits of open architectures and their limitations. Although not as simple and straightforward as the view of openness embodied in the Cycle, the additional nuance adds power to the analysis and makes it more likely that any decisions will reflect good policy.