2005/06 Research & Review

A General Perspective on Multi-Sided Software Platforms and the Japanese Computer and Consumer Electronics Industries

Andrei HAGIU
Fellow, RIETI


An increasing number of high-technology industries in the modern economy are based on computer-like systems: computers, personal digital assistants (PDAs), video game consoles, digital televisions sets, smart mobile phones, digital content carriers such as i-mode and iPod and many other consumer electronics products. Every component of computer systems plays a critical role in these businesses. The characteristics of the software platform, however, can be the key to the viability of a computer system, facilitating the creation of software applications and increasing the value to consumers of the integrated system. Moreover, the degree to which software platforms are integrated with other components affects many aspects of the organization and evolution of computer systems.

Computer systems are often organized as "multi-sided platform markets." A business is based on a multi-sided platform if it serves two or more distinct types of customers who depend on each other in some important way, and whose joint participation makes the platform valuable to each. Windows is an example of a three-sided platform. The platform provides services to application developers, hardware manufacturers, and end-users. Each group needs the others to participate if the platform is to generate market value. The emerging economics literature on multi-sided platform markets provides interesting observations into pricing and business strategies, offering insights into the workings of industries based on computer systems.

It is worth noting, though, that some markets organized around multi-sided platforms could be viable as traditional one-sided markets. Firms can choose to integrate vertically into the supply of a component rather than relying on the market. For example, Apple's iPod could, in principle, be operated as a multi-sided market platform, but Apple has decided to build its business model around a fully integrated single-sided product. Apple produces both the hardware and software, and operates its own content-provision service for which it buys the content in competition with others. Some of its competitors may, however, find that it pays to take a multi-sided approach, creating portals through which consumers buy music or other content from third-parties.

In a recent paper and soon-to-be-completed book with David Evans and Richard Schmalensee, we have analyzed several important computer-based industries, focusing on the economic and business role played by software platforms, as well as on the organization of these industries around such platforms. Here, I would like to provide a brief overview of the main insights of this work and discuss what I believe are the most important lessons from the perspective of the Japanese computer and consumer electronics industries. The key point is that multi-sided software platforms have become central to computer-based industries and this concept can throw a great deal of light on what is behind the relative loss of competitiveness of Japanese electronics manufacturers in recent years. It can also help articulate industrial and information policy priorities for the corresponding METI divisions in order to regain and conserve economic comparative advantage in digital and electronics industries.

Two-Sided Software Platforms in Computer-Based Industries

This part sketches the key insights from my work with David Evans and Richard Schmalensee, based on case-study analysis of five computer-based industries: personal computers, videogames, personal digital assistants (PDAs), smart mobile phones and digital media devices. A number of forces are driving the proliferation of industries that rely on computer systems. Microprocessors have become cheaper, smaller, and more powerful, while memory devices have become much less expensive. Today, many products come with built-in computer systems complete with microprocessors and operating systems, which in some cases are more powerful than the PCs of less than a decade ago. Often, these computer systems are integrated with the product and are designed to perform a specific set of tasks.

In the markets in which we are interested, the computer system is the core of the product and supports a multi-sided market by becoming a focal point for the development of complementary products: software applications, hardware add-ons, game, content, etc. There are three main aspects we study in our work: the current industrial organization, the evolution which has led to the current market structure and platform business models.

1. Current industry organization around two-sided platforms
As pointed out in the introduction the industries we are focusing on are organized around platforms, all of which have software at their core, some of which are also integrated into hardware. Figure 1 provides a simple representation.

Figure 1: Software Platform Layers

Figure 1: Software Platform Layers

In the computer, videogame and PDA industries, the main platforms are currently either pure software platforms (Windows, Linux, Palm) or integrated hardware-software platforms (Playstation, Xbox, Apple's Macintosh). The smart mobile phone industry is different, in that nowadays there are two software platforms sitting on top of each other. One is an operating system such as Symbian (or Windows CE) and the other is a network operator service such as i-mode: both can support applications and content and the technological relations between these different parts are quite complex.

2. Industrial evolution: vertical disintegration
One interesting common pattern in these industries is a clear "vertical disintegration" trend over time, as illustrated in figure 2. In short, looking at the computer industry for example, it started off with suppliers such as IBM providing fully integrated and standalone systems including hardware and software. The unbundling of IBM led to the emergence of an independent software industry supplying programs for IBM machines. Later, as computers became smaller and the workstation and PC revolutions unfolded, operating systems software provided by independent vendors such as Digital Research, Apple, Microsoft, Novell and others became the key focal points (i.e. platforms) in the industry. It is noteworthy that some companies like Apple have decided not to vertically disintegrate all the way, by keeping operating system and hardware tightly integrated.

Figure 2: Vertical Disintegration in Computer-Based Industries

Figure 2: Vertical Disintegration in Computer-Based Industries

A similar evolution (although on a much shorter time scale) has taken place in the PDA and mobile phone industries: these devices have evolved from single purpose electronics products supplied by individual manufacturers into small-size computers based upon software platforms able to support a variety of applications and content from third-party suppliers. Vertical disintegration has also occurred in the videogame industry, as it has evolved from single-game systems such as Home Pong, through multiple-game systems provided by the same manufacturer (Fairchild's Channel F, Atari's VCS 2600) and ultimately to videogame consoles supported by hundreds of third-party game developers (Playstation, Xbox). Interestingly enough, in this industry the software and hardware platforms have always remained tightly integrated without exception.

3. Platform business models: pricing structures
The single most important business decision two-sided platforms have to make is the choice of a pricing structure, i.e. how much to charge each side relative to the other (users vs. application/game/content developers). There is great variety here across the platforms we have studied, due to a complex combination of economic characteristics. It is therefore hard for economists to do more than provide several non-mutually exclusive pricing principles that can guide the analysis on a case by case basis. For example, in a recent RIETI discussion paper1 proposed a theoretical framework for explaining the differences in platform pricing structures in computer-based industries based on the notion of consumer demand for product variety. For the purpose of this short article, I simply show these differences in table 1. A "<0" sign (">0" respectively) sign means that the platform is incurring losses (making positive profits respectively) on that particular side of the market. For instance, the starkest contrast if between the videogame consoles and computers. Computer software platforms such as Windows and Apple OS incur losses on the application developer side by charging no usage fees and very low access fees for software tools and developer support (oftentimes below cost) and also providing numerous useful services for free, while making most of their profits through licensing to end-users (fixed fees per copy charged to OEMs). To the contrary, all videogame console makers derive virtually all their profits through usage fees (royalties) charged to independent game developers (also through sales of in-house produced games), while selling the consoles to users at or below marginal cost. i-mode and Real Player have more balanced pricing structures, in that they derive positive profits from both end-users and content providers.

Table 1: Software Platform Pricing Structures

Table 1: Software Platform Pricing Structures

Main Lessons from Successful Platform Strategies and Implications for Japan's Computer and Electronics Companies

Given Japan's international leadership in consumer electronics, one would expect it to be home to many prominent and promising platforms for high-technology computer-based industries. Interestingly enough, however, it is very difficult to find more than two examples of such platforms today, namely, Sony's PlayStation and NTT DoCoMo's i-mode. The main reason is a long-standing weakness in Japan's high-technology sector: the absence of a strong software industry, aside from videogames. In a recent column2 explained why such a mature and technology-intensive economy as Japan's has been unable to produce an internationally competitive software industry. Here I am concerned with the implication of this weakness.

Despite their apparent lack of interest in software as a product, Japanese electronics manufacturers have been able to maintain their global leadership position by relying on the innovative design, performance, and reliability of highly integrated, stand-alone systems. To put it in the familiar modularity/openness framework developed by Professor Fujimoto, these are integrated and closed (proprietary) systems, as opposed to modular and open ones.

As extensively documented by Paprzycki (2004) and others, the competitiveness of Japanese electronics companies has eroded in the face of the developments which have been transforming high-technology industries since the beginning of the 1990s. The first and most obvious one has been intensified competition from low-cost manufacturers from Korea, Taiwan, Malaysia and China, which has driven down profit margins tremendously, even for the most established Japanese electronics makers such as Sony, Matsushita, Hitachi and Fujitsu.

The second and far more profound development has been triggered by the rise of networking and connectivity (the Internet and wireless networks), and the advent of digital convergence, i.e., the growing tendency of various electronics products to interoperate and exchange information with each other. The most important consequence of this trend is that product value in high-technology industries has been shifting away from integrated and stand-alone performance toward the ability to support a variety of digital content (software applications, music, movies, games, etc.) and share that content with a variety of other electronics products. This calls for skills in organizing horizontal, platform-based networks in production, quite different from the vertical and pyramidal networks mastered by Japanese electronics companies.

Providing access to digital content is the raison d' être of the multi-sided platforms we have enumerated above. The most successful and profitable ones have been created on the premise that it is beyond the ability of any single firm to provide the entire range of products and services that consumers need, but that they can capture value by focusing on the technological support upon which consumers and the providers of these various products interact. PlayStation and i-mode are excellent examples of this approach.

PlayStation was the first truly two-sided platform in the video game industry. When it was introduced in 1995, Nintendo's and Sega's consoles still relied on in-house production for more than 60% of their games, especially their respective "killers," Mario Brothers and Sonic the Hedgehog. Sony, by contrast, had no in-house game-development experience. PlayStation became successful because it was extremely attractive for developers: it was easy to program, was versatile, offered good tools, and was less costly to support because it was based on CD-ROMs and not on cartridges.

Similarly, i-mode's platform business model succeeded brilliantly because it offered an extremely attractive platform for both third-party content providers and mobile phone users: DoCoMo was the first mobile operator to choose the c-HTML language over WAP (wireless application protocol), which greatly reduced costs for providers that already had an Internet website. And it created an advanced and efficient billing system that charged users based on actual network capacity use rather than on network time. At the time i-mode was launched in Japan, the global mobile telecommunications industry was the scene of spectacular third-generation platform failures such as Vizzavi, a joint venture between Vodafone and Vivendi, which made the mistake of attempting to rely exclusively on in-house content provided by Vivendi Universal.

Both of these examples illustrate the fact that providing all products (content, applications, games, etc.) needed by consumers is neither necessary nor sufficient for platform success. Nor is it economically feasible in most cases. Rather, it is sufficient to focus on the highest value-added part of the platform, make it attractive to all market participants, and rely on the market to supply the other products.

Moreover, as the study by Evans, Hagiu and Schmalensee suggests, most of the value of computer-based platforms seems to reside in the software platform. Think of Microsoft and Apple. By controlling the operating system and encouraging competition among PC hardware manufacturers, Microsoft managed to establish itself at the center of the PC industry. Apple, on the other hand, although it had preceded both Microsoft and IBM in the PC industry, made what today appears as a key strategic misstep by insisting on keeping the software and hardware platform integrated, which prevented it from achieving the necessary scale to compete with IBM-compatible PCs. More recently, in the market for hand-held digital assistants, Palm started off in 1998 integrated like Apple, but by 2001 it realized that its most valuable product was the Palm operating system, which it began licensing to competing hand-held manufacturers such as Sony and Handspring, before eventually spinning it off in 2003 and creating Palm Source.

Digital convergence reinforces this value shift from hardware to software. This process brings together formerly distinct products -PDAs and smart mobile phones, DVD players and videogame consoles, computers and digital TV sets- and creates significant economic value for consumers by allowing them to access and transfer digital content seamlessly through a variety of devices. In a converged world, content and the software systems and protocols (operating systems, digital rights, and intellectual property management technologies) that control content flows will command an increasingly large share of economic value relative to the underlying hardware, which follows an unmistakable trend towards commoditization.

Strikingly, none of Japan's big names in electronics seems to be in a good position to provide a dominant platform for the converged digital economy. Business Week recently dedicated an entire issue to digital convergence and named five companies it thinks will play a major role because they own critical components: Microsoft, Intel, Comcast, Samsung, and IBM - none of them is Japanese. The same issue contains a ranking of what it views as the top 100 information technology companies: The highest-ranked Japanese firm comes in at number 39 and, not surprisingly, it is NTT DoCoMo - behind no less than 15 Korean, Taiwanese, and Chinese companies.

Perhaps the clearest example of Japanese electronics firms' difficulties to build competitive and innovative platforms around their hardware strength is Sony. It is quite clear that Sony is aware of the challenges we have just described: it is in the process of developing the Cell chip technology, which will help link together its many electronics products, from computers to mobile phones and TVs. However, up to now it has stumbled when it came to combining hardware and content (Playstation being the only exception). In the market for downloadable digital music, it trails Apple's iPod by 3 years and tens of millions of consumers. This seems all the more surprising as Sony had invented the "Walkman" and possesses an unmatched amount of content (Columbia, CBS Records and, as of September 14, MGM). By contrast, Apple was prominent in neither of these areas before the iPod; however it does have experience in one important area: the creation of successful software platforms. Indeed, the core of the iPod/iTunes system is Apple's "QuickTime" software, which allows consumers to purchase music through the iTunes website, encrypt it in MP3 format on their computers, and transfer it to their iPod digital music player, all the while managing digital rights.

This is not very different from the PlayStation platform strategy and it is somewhat paradoxical that Sony was not able to learn from its own success: provide an attractive platform and rely on the market for content (be it games or music), rather than attempting to acquire and self-provide it. In fact, Sony's digital technology development has been hindered rather than helped by its content acquisitions because of internal conflicts between Sony Music and its electronics division over digital rights management and piracy issues.

The awkward position of Japanese electronics manufacturers when it comes to establishing platforms is caused by their lack of experience with software platforms. In other words, there is a lack of "platform thinking" - not of technical capacity -, which prevents them from becoming competitive software players. The implication is a risk of seeing "their" industries become dominated by foreign firms that provide scalable and cross-device software platforms, something which has already happened in the PC industry.

Still, computer-based industries will not necessarily be dominated by software firms, as illustrated by the quickly growing market for smart mobile phones. The dominant operating system platform in this industry is the Symbian operating system, which currently supports more than 2,500 applications specifically designed for smart phones. Symbian is a joint venture created in 1998 by the major handset manufacturers led by Nokia and Ericsson, in order to prevent proprietary operating systems such as Windows CE from taking control of the industry.

Concluding remarks

In conclusion, I would like to point out that in this short article I have mostly discussed relatively simple platforms, offering one type of content (music for instance) on one device (portable media player). Currently, much more complex challenges are shaping throughout the industries mentioned, which involve creating and channeling content and services through a variety of media. More sophisticated concepts of multi-sided platforms are needed but they can definitely help come up with efficient and competitive solutions, both for business executives and policy-makers.

In particular, in my view there are two important markets in Japan to which multi-sided platform frameworks could be applied successfully: the contents industry understood in the broadest sense (i.e. including content from manga, television, Internet, games, etc.) and the enterprise computing market.

For the contents industry, the biggest challenge today is to provide new and viable business models for stimulating the creation and delivery of content through many different channels: from independent creators, movie and television studios, through the Internet or mobile networks, to consumers' digital TV sets or mobile phones or computers. Doing so requires creating ingenious and flexible platforms, able to support a variety of formats and standards and run on various devices.

As for the enterprise computing market, it used to be regarded as the last bastion of vertically integrated system providers, relying on one-sided rather than multi-sided strategies. This is drastically changing as illustrated by the current battle between two key players, Oracle and SAP, for providing the dominant software platform (not service) of the corporate market3. For SAP, the platform is NetWeaver, which is software that sits above a database, links together applications and allows third party vendors to create innovative software services on top of it. By contrast, Oracle sees the database as the crucial platform. Furthermore, IBM, the market leader, has also embraced a two-sided platform strategy with Websphere, which is designed not only to provide service directly to end-users but also to support services and applications from other vendors. In this context, large Japanese players such as Fujitsu, Hitachi and NEC are looking to shift away from purely integrated, and standalone offerings towards multi-sided platform strategies in order to remain competitive. It is a very challenging shift to manage, however I believe a multi-sided platform framework can go a long way in helping them undertake it4.


  1. "Two-Sided Platforms: Pricing and Social Efficiency," December 27th 2004.
  2. "Perspectives on the Japanese Software Industry: Fragmentation, Modularity and Upcoming Challenges," January 12th 2005.
  3. "Business Software: Soft War or Hard?", The Economist, March 23rd 2005.
  4. The author is currently conducting study sessions on this subject with several Fujitsu researchers.

June 22, 2005