Innovation in the Software Sector (Summary)

Information systems hold the key to productivity improvement in all industries, including financial services, communications and transportation. Moreover, embedded software has come to account for a significant share of the value of such products as cars and mobile phones. Software innovation plays a crucial role in innovation not only in the IT industry, but also in innovation throughout the economy and society in general. It is frequently noted that the innovation process in the software sector is different from that of other industries because of the speed of change in software technologies and the existence of strong network externalities. Today there is a need to achieve a fuller understanding of these characteristics of the software sector in order to devise effective policy measures.

In light of this need, a project entitled "Innovation in the Software Sector" was launched in March 2007 under the aegis of the OECD Committee on Industry, Innovation and Entrepreneurship (CIIE) for governments and industries to jointly discuss and analyze the factors shaping the developments in the sector from a multifarious perspective. In this conference, the interim findings of the project were presented and various efforts that have been made to enhance dependability in the software sector in Japan were introduced with the aim of promoting multidisciplinary discussion among experts.

The Tokyo conference began with opening speeches given by TAKAICHI Sanae, Senior Vice Minister of Japanese Ministry of Economy, Trade and Industry (METI) and AMANO Mari, Deputy Secretary General of the OECD. Ms. Takaichi welcomed the participants to Tokyo and congratulated METI, RIETI and the OECD, who organized the conference jointly. She also recalled the Japanese ministerial proposal to analyse innovation processes in the software industry, which eventually resulted in the launching of the OECD study. Mr. Amano stressed the importance of innovation and the complexity of the related processes, sketching the wider context in which the project on innovation in the software sector is set, in particular the work on the OECD Innovation Strategy. He also recalled the Japanese initiatives on "Innovation 25" and "New Economic Growth Strategy" that well demonstrate Japan's experience in initiatives for promotion of innovation.

Two distinguished researchers participated in the keynote speech and outlined certain critical points to keep in mind when considering the characteristics of the software sector and the process of innovation in this sector. KOKURYO Jiro, a professor at the Graduate School of Media and Governance at Keio University, first emphasized the importance of "emergent value creation" in software innovation. Next, Dr. Michael CUSUMANO of the MIT Sloan School of Management described the advantages of hybrid firms whose revenues from product and software sales reinforce each other. It is notable that both speakers presented ideas that could serve as new business models for the software sector.

According to Dr. Kokuryo, innovations in hardware technologies have yielded dramatic improvements in processor capabilities. However, software development methods remain basically bound by the coding process that must be completed by software programmers. Given the difficulty of improving productivity in software development, the key to higher productivity lies in the reuse of existing software.

The software sector faces high fixed costs for development. On the other hand, its marginal costs are low because once completed; software can be easily copied to electronic media. Due to the development of the Internet, delivery costs have been minimized in recent years. The importance of "emergent value creation" is increasing in software innovation.

Emergent value creation is based on the self-organization of interactions among a very large number of players, which leads to outcomes that were unexpected by any of the players. The essential question is how to efficiently link this process with innovation. Experiments in emergent value creation using the Internet are being conducted at the Shonan Campus of Keio University. The development and provision of platforms will have a very important role to play in business models for the future information industry. Such platforms can be developed by providing an Internet space that is always open and can be connected to anything, then allow individual players who are recipients of information in this space to develop value-generating content.

In the following address, Dr. Cusumano explained how software industry revenues from the sale of packages and licensing have been decreasing while income from software updates and services is increasing. One example is Siebel, a CRM (customer relationship management) software vendor. When first listed in 1995, product sales accounted for 95% of its revenues, but the share of services has been growing steadily ever since and eventually overtook product sales in 2002.

New business models are evolving for recent startups that are providing web-based software and services. Instead of earning income from software services, their revenues are generated by advertisements. This can be explained using the concept of innovation dynamics - product innovation leads to dominant design that results in process innovation as advocated by W.J. Abernathy. However, the more recent trend features the emergence of new business platforms. This can be interpreted as the precursor to the start of the next cycle.

An analysis using a database covering 300 U.S.-listed software firms shows that hybrid companies that have effectively combined product sales and software sales enjoy the highest profit margins. Such firms are not only good at software product innovation, but have also invested extensively in technology development and marketing for providing services. Software production firms are putting greater emphasis on services, and thereby encroaching upon the domain of system integrators and IT service firms. This has intensified competition in markets for IT services.

Software product firms have made progress in developing standardized platforms, but only a small number of platform leaders will be able to survive. Pure IT firms specializing in services are threatened by intense competition from India with its advantage of low personnel costs. Therefore, to survive, firms will have to take the hybrid approach of combining products and services. To succeed in the hybrid approach, it will be important to consider possibilities for "servitizing" products and "productizing" services.��

In the first session, presentations were made from the perspective of software development firms on the challenges of improving software productivity and promoting innovation. Mr. KUSHIKI Yoshiaki, Senior Fellow at Panasonic Corp., approached the subject from the perspective of the qualities needed in a software development leader, while Dr. Marshall PHELPS, Microsoft Corporate Vice President for IP Policy and Strategy, outlined a new approach to constructing an environment of coexistence between open-source software and Microsoft.

In the field of digital consumer electronics, explained Mr. Kushiki, the ratio of product development costs to total cost is increasing, as is the weight of embedded software costs in product development costs. Human resources in the development of embedded software have become very important.

Embedded software engineers must have expert knowledge of both software and hardware, including mechanical and design aspects. They must also have an awareness of the hardware aspects of the product while developing software to realize the required functions.

The qualities needed in a software development leader have changed over time. In the first stage of digitally controlled appliances, the leader's role was that of a project manager whose principal concern was meeting delivery deadlines. The second stage of digital audio-visual equipment started around 1996. The leader's role in this stage was that of a platform architect. An essential aspect of the leader's role was the selection of platforms for the creation of multiple models and continuous development of new products.

The third stage started around 2006 and can be characterized as the age of network appliances. The leader in this stage is a systems architect who is guided by global business strategies. In this new environment, the leader is required to develop business models that take into consideration multiple numbers of network-connected appliances. One of the key issues is arriving at a balance between economies of scale available through globalization and maintaining compliance with the environmental and power consumption standards of individual countries. In this context, the leader must be capable of developing proposals that respond to various new challenges, such as global business strategies and environmental problems.

Dr. Phelps discounted the common belief that Microsoft Windows and open-source software, such as Linux, are in conflict. According to Dr. Phelps, they actually coexist as heterogeneous software systems where users can select the functions they want from both Microsoft and Linux vendors, such as Red Hat and Novell, to develop the systems they want. Microsoft is supporting this form of user-led systems development through a cooperative project with Linux vendors. Under the name "Bridge Builder Project," it is being carried forward with the cooperation of Novell.

However, there are serious problems with the GPL (general public license) clause advocated by open source groups. Software development requires enormous R&D expenditures, and to recover these costs intellectual property rights must be protected. To maximize user benefits, it is necessary to maintain a certain level of protection for intellectual property rights and continue R&D activities while also offering heterogeneous software systems that incorporate the advantages of open systems.

In the second session, the focus shifted to the current situation and challenges in the software sector as seen from the perspective of users. Two presentations were made on this subject from the standpoints of academia and the business world. Dr. Marshall VAN ALSTYNE of Boston University and MIT reported on the interrelations between innovation, openness and platform control as an economist. Mr. KASAHARA Yutaka, Vice President of Software Business Promotion Unit at NEC Corporation, argued that dependability was the key to innovation in software and services. Dr. Jürgen MÖSSINGER, Vice President of Automotive Systems Integration at Bosch, emphasized that mastery of increasing complexity is a key requirement in the development of the automobile industry, and that the re-use of software promoted by AUTOSAR was one possible solution.

In the second part of the second session, Mr. HAMAGUCHI Tomokazu, Chairman of the Japan Information Technology Services Industry Association (JISA); Mr. KAMIYAMA Shigeru, Vice Chairman of JISA; and Mr. HIRONISHI Koichi, Corporate Senior Executive Vice President of Fujitsu Ltd.; each presented case studies concerning the realization and improvement of dependability, a key factor in considering the promotion of software innovation.

To open the session, Dr. Van Alstyne pointed out that software platforms increase innovation speed by optimizing openness and connectivity between software layers. Open relations among players are more efficient in achieving innovation than cooperation among specific firms because they (a) increase network externalities, (b) lower opportunity costs in software development, (c) raise the software output level, (d) accelerate the pace of technological innovation, and (e) increase the number of developers.

Technological uncertainty reduces the willingness to open up, but large-scale software houses will try to avoid this problem through their production scale and risk diversification. Competition among software developers reduces openness, while competition among platforms has the opposite effect and encourages openness. Software developers prefer commercial platforms to public standards.

In the following presentation, Mr. Kasahara described how the scale of embedded software in cell phones has increased with the growing complexity of products, and has reached the level of 5 million-10 million lines. Bugs affecting as few as a 100 lines can cause systems failure. In cell phones, nearly 44% of all post-shipment defects are software related. A recent systems failure in automatic ticket gates affected 2.6 million people. The realization of overall dependability is very important in large-scale systems comprised of numerous products and subsystems.

There have been three waves in software development. The first came in the 1970s and featured the development of more efficient offices. The second wave came in the 1990s and was represented by open architecture and the Internet. The third wave will be centered on the services that will be developed over the next decade or so. Signs of this third wave can be seen in wireless technologies, IP network technologies, services based on RFID (radio frequency identification) and other element technologies, and concepts based on cloud computing.

Service value is decided through interactions between service providers and receivers. Therefore, it is very important for metrics of service dependability to include human and social factors. It is necessary to establish technologies for developing services that match the levels of these metrics.

Dr. Mössinger next spoke about innovation in the automobile industry and how it is supported by electronics and software. In 2003, 70% of all innovation occurred in these two areas. This percentage is expected to go up in the future. In order to keep up with the complexity of product functions, software has become increasingly complex. How to cope with complex software poses an important challenge. Before we can master complexity, we must first understand complexity. One solution is to engage in software reuse through "sharing among companies." AUTOSAR, a partnership involving 150 automotive firms from around the world, plays a crucial role in this area. AUTOSAR is pursuing standardization in three areas: architecture, software development methodologies (function-based software development that is independent of hardware), and interconnectivity of application software.

A key challenge for the future is a shortage of engineers. Universities in Germany, the home of Bosch, are noted for training and educating excellent engineers. However, Bosch is also recruiting on the international level and is conducting in-house training of software engineers. To overcome the shortage in human resources, the company is also actively outsourcing development work offshore, primarily to India.

Mr. Hamaguchi and Mr. Kamiyama next talked about the urgent need to achieve high levels of reliability in information systems due to the enormous impact that systems failures can have on people's lives and corporate activities. One approach to this challenge is outlined in "An Investigation on the Software Metrics to Promote Best Practices for Reliability Improvement" undertaken by JISA. Their presentation reviewed actual cases of the use of these metrics and their results, and offered an outlook for future developments.

Mr. Hironishi discussed information systems reliability and pointed out that to ensure reliability it is not enough to approach the issues from only the perspective of vendors. The perspective of the actual users of the system must also be incorporated in the process, and users and vendors must work together to achieve reliability. He then went on to introduce specific methods for ensuring reliability. The "customers' view study group" aims to avoid gaps in awareness between users and vendors during the development process by examining review methods and how to describe external specifications in ways that are easy to understand for users. The "grades standards for nonfunctional requirements" aims to visualize customer requirements by focusing on nonfunctional requirements such as the speed at which functions are processed in the development of information systems, and failure levels at which functions should continue to be provided. Finally, "IT industry activities for developing common views with users" and "improving technological capabilities by actively adopting industry standards for internal practices" were mentioned as important issues in ensuring reliability.

Session 3 featured a panel discussion concerning policy implications. Drawing on the preceding presentations, the panelists covered a variety of topics and engaged in an active exchange of views that included questions from the floor.

RIETI Faculty Fellow MOTOHASHI Kazuyuki, Professor of the University of Tokyo, led off the panel discussion with his observation that the software industry has high fixed costs and near-zero marginal costs. As mentioned by Dr. Cusumano, the weight of services is increasing. Services are labor intensive and need to be analyzed based on economic principles that differ from what has been used for conventional software products. This analysis requires a highly complex economic model that can handle multilayer structures of mutually interacting OS, middleware, applications, and the presence of significant network externalities. Two types of policy instruments are available for the software industry; one consists of direct measures for promoting R&D and human resources development, whereas the other consists of indirect measures for promoting innovation by developing the necessary framework conditions.

In the area of direct policy measures, the development of human resources is very important. Analysis shows that there is a positive correlation between the percentage of applicants passing the Information Technology Engineers Examination and software productivity levels. However, such skill standards as the Information Technology Promotion Agency's ITSS and ETSS have not been adequately publicized. For this reason, we must increase public relations and promotional activities.

In the area of indirect policy measures, the intellectual property system is very important. Software became patentable in many countries during the 1990s, and there is no doubt that this has had a positive impact on innovation by independent software houses. On the other hand, it is difficult to determine the novelty and inventiveness of software. The resulting increase in the number of patent infringements claims and lawsuits cannot be ignored.

It is known that the Japanese and U.S. software industries differ significantly in structure (U.S. is package-centered while Japan is more focused on customized applications). In this project, efforts should also be made to derive country specific policy implications.

Dr. Arnaud LE HORS, Program Director, Standards & Emerging Markets, IBM, next talked about groups of players working in collaboration across geographical and organizational boundaries, and the increasingly important role they play in innovation in the IT industry. To promote this type of open innovation, it is critically important to achieve standardization and interoperability. Governments have a highly significant role to play in this process. However, it is important for governments to focus on developing an environment that is conducive to fair competition among private companies. In this context, it is problematic for the government to provide protection or support to specific companies or industries.

Mr. David CIERCO JIMENEZ de PARGA, General Director for Information Society Development in Spain's Ministry of Industry, Tourism and Commerce, closed out the third session with his presentation on policies for software innovation in Spain, where implementation of Plan Avanza is underway to promote information technologies and nationwide broadband connectivity. The IT industry is being promoted primarily through CENATIC. For instance, knowledge centers have been established throughout the country and cooperation between universities and local small and medium enterprises has been encouraged. To continue these programs for promoting software innovation, it is necessary to clearly identify national strategies and steadily implement them.

Questions concerning the following policy issues were received from the floor and resulted in active discussions:
What is the relationship between open innovation and patent systems?
What is the relationship between software patent management policies and innovation?
What impact has government software procurement had on software industry development and innovation?
Do environmental and safety regulations obstruct software innovation?

The discussions at this conference indicated that analysis of the software industry is still in the development stage in the academic community as well. A common view emerged from the conference that OECD and RIETI should continue to promote analysis of the software industry, and that progress in such analysis has an important role to play in resolving policy issues.

The views expressed in these proceedings do not necessarily reflect those of OECD and its member countries. The editorial department of RIETI is responsible for the compilation of these proceedings.