Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Personal Computing Jason Dedrick University of California at Irvine Kenneth L. Kraemer University of California at Irvine INTRODUCTION August 2006 marked the 25th anniversary of the release of the original IBM personal computer (PC), the product that defined the standards around which a vast new industry formed. Unlike the vertically integrated mainframe industry, the PC industry consisted of a global network of independent suppliers of sys- tems, components, peripherals, and software (Grove, 1999; Dedrick and Kraemer, 1998). The key factor shaping the industryâs structure was the design of the IBM PC as a modular, open system with standard interfaces, which allowed many newcomers to enter the market by specializing in one industry segment and de- veloping innovations that could be integrated into any IBM-compatible system. It also permitted producers of parts, components, and systems to achieve global economies of scale as most of the world adopted the IBM standard. In time, desktop PCs were joined by portable laptop/notebook PCs and PC servers as the industry innovated on this common standard. Today, the core personal computing industry includes not only traditional desktop and laptop PCs and PC servers but also smart handheld devices such as personal digital assistants (PDAs) and smart phones. This core industry is sup- ported by a large number of component suppliers, manufacturing services and logistics providers, distributors, retailers, service specialists, and others. These companies also support other segments of the electronics industry, and so are counted here not as part of the PC industry but as part of its overall production and innovation network. This network not only supports innovation in the core industry segments but also provides the necessary infrastructure for innovations 19
20 INNOVATION IN GLOBAL INDUSTRIES in newer product categories such as ultramobile PCs, MP3 players (e.g., the iPod), and smart phones. Worldwide revenues for the core PC industry totaled $235 billion in 2005: $191 billion in desktop and portable PCs, $28 billion in PC servers, and $16 bil- lion in smart handheld devices (IDC, 2006a). In addition, PC software accounts for about half of the packaged software industry, whose 2006 sales were $225 billion, and PC use also drives sales of information technology (IT) services and of other hardware such as storage, peripherals, and networking equipment (IDC, 2006c). The PC has undergone considerable innovation and change since it was first introduced. The traditional PC is no longer expected to be the sole locus of inno- vation in the future, but simply one of many devices âorbiting the userâ (Econo- mist, 2006). Communications devices (phones, PDAs) have acquired computing capabilities and people now send e-mail with a BlackBerry or download music on a mobile phone. Digital photos can be transferred from a camera to a PC and uploaded to a website, transferred directly to a printer, or shot and e-mailed with a mobile phone. And although the traditional desktop and laptop PC is becoming less central to all computing activities, over 225 million PCs were sold in 2006 and the PC is often the first place to find innovations that may migrate later to other devices. As important as product innovation has been, equally important is the steady price declines in recent years, which have brought PCs within the reach of more of the worldâs population. Emerging markets such as China and India are growing much faster than the more mature developed markets, and PC makers have begun to focus on innovation that addresses the needs of those markets at low prices. Globalization of production has been credited for making computer hardware 10 to 30 percent cheaper than it would be otherwise (Mann, 2003). The availability of ever cheaper, smaller, and more powerful hardware has continued to expand the market and has stimulated ongoing innovation in hardware, software, and services. Although globalization has been a major factor in the growth and innovation of the PC industry, it raises issues for U.S. companies, government and other institutions, and workers. U.S. PC makers are struggling to eke out a profit in an environment of falling prices and intense international competition. Government policy issues include tax incentives, antitrust, immigration, and market access. Universities must ensure that they are training people with the skills that industry needs, and workers must invest their own time and money to acquire those skills even as more highly skilled knowledge work is moved offshore. The impacts of globalization have been debated extensively. An optimistic view is that U.S. firms are outsourcing and offshoring lower-end manufacturing and routine engineering work, freeing resources to focus on more dynamic in- novation that will sustain profitability and create new jobs in the United States.
PERSONAL COMPUTING 21 A more pessimistic view is that innovation will follow manufacturing offshore, leaving U.S. firms uncompetitive and draining the United States of the innovation that drives growth and employment (Kotkin and Friedman, 2004). While macro-level data can be useful in analyzing the impacts of globaliza- tion, trends and impacts can be easier to spot at the industry level, especially when looking at more dynamic industries where change is happening faster. Personal computing is one such industry. Therefore, this chapter examines the globalization of innovation in the PC industry, its causes, its impacts, and its strat- egy and policy implications. The focus is mainly on innovation-related activities in U.S.-branded PC companies set in their global context; it is not an analysis of PC companies in other economies such as Japan, Taiwan, or China, although it brings them in as part of the global supply chain and the competitive context. This chapter is a fact-based analysis grounded in over 200 personal inter- views with industry executives in the United States and Asia, data from the In- ternational Data Corporation (IDC), Taiwanâs Market Intelligence Center, Reed Electronics Research and other sources, published empirical research, and our study of the industry for over 20 years. We find that the global division of innovation-related activities can be char- acterized as follows: component-level research and development (R&D), concept design, and product planning are performed mostly in the United States and Japan; applied R&D and development of new platforms mostly take place in Taiwan; and product development for mature products and a majority of produc- tion and sustaining engineering are performed in China. U.S. PC firms have benefited from this international division of labor, which has supported rapid innovation and quicker integration of new technologies into their products. The growing demand for smaller, more mobile products plays to U.S. firmsâ strengths in product architecture and early-stage development. Their bigger problem is earning profits from innovation in an industry dominated by Microsoft and Intel, who capture very high profit margins thanks to their control of key standards. From the perspective of U.S. knowledge workers, the situation is more mixed. The shift in production away from the United States has pulled many new product development jobs to Asia, whereas design and early-stage development work has remained largely in the United States. Still, the new jobs created by the industryâs growth are largely outside of the United States. Finally, consumers in the United States have been clear beneficiaries of the very low cost structure that globalization has produced in PCs as average selling prices have been reduced continually. Following this Introduction, the structure of this chapter is as follows. The section âInnovation in the Industryâ analyzes the nature of innovation in PCs and how production and innovation are organized across the value network. âChang- ing International Structure of Demand and Supplyâ describes international trends in PC demand and production. The fourth section, âGlobalization of Innovation,â
22 INNOVATION IN GLOBAL INDUSTRIES reviews the global structure of innovation in the PC industry and the factors driv- ing globalization. âImplications of Globalization of Innovationâ considers the implications of the foregoing trends for firm strategy and U.S. national policy. INNOVATION IN THE INDUSTRY The PC industry has introduced many innovations in its 25-year history. Product innovation includes the creation of new product categories such as notebook PCs and PDAs, as well as the creation of new product platforms such as multimedia PCs and wireless âmobilityâ notebooks. The scope and outcome of product innovation in PCs is shaped by the presence of global architectural standards set originally by IBM and now largely controlled by Microsoft and Intel. Common interface standards enable innovators to reach a global market with standard product lines; thus, economies of scale can be achieved to support investments in product development and manufacturing capacity. This is different from other industries, such as mobile phones or video games, in which multiple incompatible standards exist. An example of the benefits of standardization is the acceptance of 802.11 as a common standard, which spurred the introduction of wireless networking as a standard feature on notebook PCs. On the other hand, standardization battles can constrain innovation because PC makers are reluctant to incorporate technologies before a standard is set, as is the case with second- generation DVD technology. When PC makers do innovate, they face hard choices in trying to capture profits from their innovations. One alternative is to incorporate the innovation only in their own products to differentiate their PCs from those of competitors, but there is a question of whether they can convince customers to pay for the differentiation and also whether customers will want to adopt a nonstandard technology. Another is to license the technology broadly, which might bring in license fees and even establish the technology as an industry standard, but which will eliminate product differentiation. One current example is Hewlett-Packardâs (HPâs) Personal Media Drive (PMD), a portable hard drive that slides into a spe- cial slot in HP Media Center PCs. HP incorporated the special slot into some of its own products, while letting customers connect the PMD to competitorsâ PCs using a slower USB connection, thus differentiating HPâs PCs. By contrast, HP has licensed its LightScribe technology, for labeling DVDs and CDs, to other PC makers. In either case, it can be difficult to translate innovation into profits sufficient to justify the R&D effort. Despite these challenges, which may discourage more fundamental product innovation, PC makers are pushed to incremental innovation by component makers (such as for semiconductors, storage, or power supply) who introduce frequent changes in their products (faster speed, greater capacity, smaller form factor, longer life) in efforts to gain greater market share within their industry sector. They also are pushed by consumers who want the latest technologies. PC
PERSONAL COMPUTING 23 makers feel they have to adopt these often-incremental changes rather than risk being left behind by a competitor that does adopt. As a result, PC makers have tended to concentrate on operational efficiency, marketing, and distribution rather than trying to use product differentiation as a source of sustainable competitive advantage (Porter, 1996). Product innovation at the system level tends to be incremental and emphasizes developing slightly different products for narrowly defined market niches, such as PC gamers who demand high performance or business travelers who desire ultralight notebooks, rather than more distinctively innovative products. Instead, most product innova- tion occurs upstream in components and software, which are then incorporated by PC makers. Consistent with the emphasis on efficiency and distribution, the industry has introduced business process innovations such as outsourced manufacturing, using the Internet as a direct sales channel, vendor-managed inventory, third-party lo- gistics, and build-to-order (BTO) production. At the plant level, some firms have replaced assembly lines with small production cells to facilitate BTO production and have adopted process improvements such as reducing the number of steps and improving quality in final assembly. They also have employed a range of in- formation technologies such as shop floor management systems, bar coding, and automated software downloads to improve manufacturing performance (Kraemer et al., 2000). However, while early adoption of these innovations benefited some companies, particularly Dell Inc., competing PC makers have since adopted these and other process innovations and closed the gap on key measures such as inven- tory turnover and time to market for new products (Dedrick and Kraemer, 2005). Today, most companies use a mix of build-to-forecast and BTO processes that is optimal for their targeted markets. The result is greater efficiency in the industry as a whole, but the biggest benefits have not gone to the PC makers. They have mostly gone to consumers in the form of lower prices, and to Microsoft and Intel, as software and microprocessors account for an ever greater share of the total cost of a PC. To understand innovation in the industry, it is important to look at the struc- ture of the innovation network, the innovation processes, the key personal com- puting products, and interdependencies among innovation processes, products, and the structure of the network. â An exception is Apple, which emphasizes attractive design and close integration of hardware and proprietary software in its products. While this has been very successful in its iPod line, Appleâs mar- ket share in PCs is under 4 percent worldwide, so it is unclear that its innovative PCs have done more than satisfy a small core of Mac users who are willing to pay a premium for its products. By adopting Intel processors for all of its products, Apple has abandoned its proprietary hardware platform in favor of global economies of scale and greater compatibility with Windows PCs. â Even these two face challenges: Intel from AMD and Microsoft from Linux in one product cat- egory (servers).
24 INNOVATION IN GLOBAL INDUSTRIES The Innovation Network The PC industryâs innovation network consists of component makers, con- tract manufacturers (CMs) and original design manufacturers (ODMs), branded PC firms, distributors, and resellers (Figure 1). The industry can be characterized as horizontally specialized, with the branded firms as the âsystem integratorsâ doing design and outsourcing development and production to CMs or ODMs. There are less than a dozen globally competitive PC makers and many smaller local assemblers, supported by another dozen ma- jor CMs and ODMs. There are several major suppliers of most key components (e.g., motherboards, hard drives, displays, optical drives, memory, and batteries). Farther upstream in the supply chain, there are several thousand suppliers of less expensive parts and components, most of which are small- and medium-sized firms. Distribution is mostly decentralized and local, although there are a few large distributors who operate internationally such as Ingram Micro, Tech Data, and Arrow Electronics. Our main focus in this chapter is on the branded PC vendors and ODMs who collaborate to bring new products to market using components from upstream suppliers. Most R&D is done upstream in the industryâby the suppliers of micropro- cessors, software, peripherals, and components. This innovation is global in the sense that there are major component makers in the United States (microproces- sors, graphics, memory, hard drives, networking, software), Japan (liquid crystal displays [LCDs], memory, hard drives, batteries), Korea (LCDs, memory), and Taiwan (LCDs, memory, optical drives, power supply, various peripherals). However, although some companies have set up R&D labs around the world, most R&D is still done in the home country. Some PC makers such as HP, Toshiba, Sony, and Samsung also make components and peripherals, but these are generally done in separate business units who sell to competing PC makers as well as their internal PC units. The pace of this upstream innovation is a major factor shaping innovation by branded PC vendors who innovate through âsystems integration.â The PC vendors identify new product markets and design systems that incorporate new technologies to serve those markets. For instance, PC makers identified mobile PC users who want network access without having to plug into a phone line or local area network. This capability was made possible when wireless networking technologies such as WiFi were introduced by component makers. It was then up to PC makers to incorporate the technology into their products. More impor- â The terms contract manufacturer and original design manufacturer are used commonly, but not always consistently, in the electronics industry. Contract manufacturers provide a range of manufacturing services, including subassembly, final assembly, logistics, and even customer service. Original design manufacturer is a term coined in Taiwan when its contract manufacturers began to offer product design and engineering as well as manufacturing of notebooks, motherboards, and other products.
PERSONAL COMPUTING 25 Product innovation Operations Customer relations Direct PC vendors Indirect Component Component suppliers suppliers Resellers Customers Contract manufacturers/ODMs Distributors Activities R& D Design Engineering Manufacturing Assembly Distribution Sales, service Components, subassemblies, box-builds Complete systems FIGURE 1â The PC industry innovation network. SOURCE: Adapted from Curry and Kenney (1999). PC industry-1.eps tant, they had to introduce a new technology at a time when the infrastructure to support wireless networking was nearly nonexistent, hoping that this would create the impetus for firms and consumers to invest in wireless networks. Apple initially jumped in by incorporating 802.11 wireless technology in all of its note- books, and was soon followed by other PC makers. Soon, wireless networks were available in offices, homes, schools, airports, and coffee shops around the world. Appleâs early decision was very risky, as there were few networks available, but taking the risk helped to create the market for them. The creation of new markets by PC makers, in turn, can shape the direction of upstream innovation in components. For wireless notebooks, PC vendors had to decide which networking standard(s) to incorporate as well as find components with low power consumption, longer battery life, and light weight. Available components seldom meet all these needs, so the lead PC vendors each developed their own product roadmaps, which signal to the component suppliers where the firm is headed, the target markets and expected volumes, and the price and per- formance of components needed to succeed. By doing so, they provided advance knowledge to the upstream suppliers who could respond in terms of feasibility, aggregate demand across PC vendors, plan for the coming changes, and inform their own suppliers. These PC maker roadmaps, which are different from those
26 INNOVATION IN GLOBAL INDUSTRIES provided by Intel and Microsoft to the PC makers, are essential to knowledge integration along the supply chain. Innovation Processes Product innovation in the industry occurs through two broad processesâ R&D and new product development. R&D is an ongoing activity that generates knowledge that can be applied to multiple products. New product development is a multistage process of design, development, and production that creates physical products for target markets. Although conceptually distinct, there is often a close interaction between the two in practice. New product development integrates knowledge developed by R&D, and R&D is often called on to solve a specific problem in product development. Given that most R&D is done upstream by the component suppliers, the process of knowledge integration occurs between the supplier and the PC maker. The focus is on knowledge needed to integrate a standard component, but occasionally it involves customization or even more intensive joint development. This is especially the case when an entirely new product is being created, such as the wireless notebook that requires integration of communication technologies, or in the case of a new product category such as the Apple iPod. Products and Innovation Activities Although new form factors are emerging, desktops and notebooks remain the leading products in the industry, with important differences between them that affect innovation activities. For desktops, product innovation mainly cen- ters on conventional systems integrationâincorporating new parts, components, and software into a system and ensuring that they work together. The system is largely standardized with respect to components, parts, and interfaces. So in- novation involves the selection of components to be included for different target markets (e.g., home, office, game, âvalueâ or âpowerâ user). Most use a standard full tower or midtower chassis with industrial design applied mainly to the bezel (face) to reflect a certain brand image. A few newer models aimed at consumersâ living rooms have moved away from the âbeige boxâ to smaller and more stylish designs with unique chassis and industrial designs. PC vendors generally keep concept design and product planning in-house for close control over brand image, user interface, features, cost, and quality. Outsourcing of physical development has occurred in a series of steps since the mid-1990sâfirst motherboard design, then mechanical design, system test, and finally software build and validation. â Adetailed discussion of these phases and the activities within each is provided by Dedrick and Kraemer (2006b).
PERSONAL COMPUTING 27 Intel facilitated this trend by providing support and reference designs to ODMs who develop motherboards and full systems. For notebooks, innovation involves high-level system integration with com- plex mechanical, electrical, and software challenges. Design of such a small form factor presents special challenges with respect to heat dissipation, electromag- netic interference, and power consumption, while the need for portability requires greater ruggedness. Although components such as disk drives and flat panels are mostly standardized, notebooks involve many custom parts. For example, to fit the modular components within the notebook chassis, the motherboard and bat- tery pack may have to be customized for each notebook model. The chassis and other mechanical parts require custom tooling. PC vendors usually keep notebook design in-house but coordinate physical development jointly with the ODM because there is a strong interdependency between the physical product development and manufacturing. It is critical that product development take manufacturability into account from the beginning; otherwise a product may be developed that cannot be produced at the neces- sary volume, cost, or quality. Most notebook PCs are designed to be built in a particular assembly plant with specific manufacturing process requirements. As a result, product development and final assembly are almost always handled by one company. In some cases, this means the PC maker keeps both in-house. In most cases it means outsourcing both development and manufacturing of each model to a single ODM. Thus, the interdependencies of PC form factors and new product develop- ment (NPD) activities have led to different organizational arrangements for desktops and notebooks (Figure 2). Because desktops are less complex and more standardized, a complete product specification can be handed off for develop- ment and production to ODMs, or a fully developed product can be turned over to a CM for manufacturing. However, because of their greater complexity and customization, notebooks tend to be designed and developed jointly by the PC vendors and ODMs. R&D NEW PRODUCT DEVELOPMENT Design Development Production MS/Intel, PC Notebook ODMs Components vendors MS/Intel, PC Desktop Components vendors CMs/ODMs FIGURE 2â Organization of innovation for desktops and notebooks. PC industry-2.eps
28 INNOVATION IN GLOBAL INDUSTRIES As a result of the interdependencies in notebook PC development, leading PC makers HP and Dell have set up design centers in Taiwan to work closely with ODMs, whereas others frequently send staff from the United States. The ODMs may divide product development and manufacturing between Taiwan and China but keep very close interaction between the two locations. For desktops, it is easier to separate development and manufacturing geographically as well as across firm boundaries. CHANGING INTERNATIONAL STRUCTURE OF DEMAND AND SUPPLY Trends in Demand PC demand has been shifting steadily for over a decade toward smaller, more integrated, and more communications-oriented products. The global demand for PCs is changing in terms of form factor, commercial versus consumer markets, and regional consumption. Portable devices (laptops and notebooks) are the fast- est growing form factor, totaling 32 percent of unit demand in 2005 compared to just 10 percent in 1990 (Figure 3), and are expected to exceed desktops in the next 5 years (IDC, 2006b). Other portable devices such as smart phones have seen rapid growth as well. This means that there will be more demand for complex 100 90 80 Desktops 70 Portables Percentage 60 50 40 30 20 10 0 9 0 91 992 993 994 995 996 997 998 999 000 001 002 003 004 005 19 19 1 1 1 1 1 1 1 1 2 2 2 2 2 2 FIGURE 3â Global demand for desktops and portables, 1990-2005 (percent of units sold). SOURCE: Juliussen (2006). PC industry-3.eps
PERSONAL COMPUTING 29 innovation in concept, design, and engineering in the future and that coordination among these stages will have to become closer. Continued price and performance gains in key components as well as the shift of production to lower-cost locations have driven prices lower, expanding overall demand for PCs. One impact is in consumer markets, whose share of the total market increased from 28 to 38 percent between 1994 and 2005 (Figure 4). Another impact is in emerging country markets where economic growth is providing the income to afford these ever-cheaper PCs. Although North and South America are still the biggest market in the world, followed by Europe, the Middle East, and Africa (EMEA), the Asia-Pacific region is the fastest-growing market (Figure 5). The United States is the single largest market, with 61 million units shipped in 2005, but fast-growing China has surpassed Japan as the second biggest market. 80 70 60 50 Percentage 40 30 20 WW Consumer WW Commercial 10 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 FIGURE 4â Global PC consumption by commercial/consumer markets (percent of units sold). SOURCE: IDC (2006d). PC industry-4.eps
30 INNOVATION IN GLOBAL INDUSTRIES 50 45 40 35 30 Percentage 25 20 The Americas EMEA 15 Asia-Pacific 10 5 0 00 04 05 02 03 01 0 96 98 99 94 95 92 93 97 91 9 20 20 20 20 20 20 19 19 19 19 19 19 19 19 19 19 FIGURE 5â Global PC consumption by region, 1990-2005 (percent of units sold). SOURCE: Juliussen (2006). PC industry-5.eps Geographic Location of Production With desktop PCs, final assembly by the branded vendors historically was located close to end-user demand because of logistics (they are too heavy to ship affordably by air) and greater customization for national or regional markets. Major PC vendors such as IBM, Compaq, HP, Apple, and Gateway initially had their own production facilities in each world region, but they later outsourced production to CMs such as SCI, Flextronics, Solectron, Mitac, and Foxconn (the registered trade name of Hon Hai Precision Industry Co.), starting in the late 1990s. Dell kept final assembly in-house, but it outsourced base unit production, including chassis with cables, connectors, drive bays, fans, and power supplies. Japanese and Asian vendors generally kept production in-house. As the branded PC vendors moved offshore and then outsourced, there was a shift in the location of production from the Americas and EMEA to the Asia-Pacific region (Figure 6). Initially, production was spread throughout East Asia in Japan, Malaysia, Singapore, Taiwan, and Korea. Production of desktop base units and various components and subassemblies by Taiwanese companies shifted to the Pearl River Delta in Southern China, but final assembly was usually done regionally: in the United States and Mexico for the Americas, in Ireland
PERSONAL COMPUTING 31 250,000 The Americas 200,000 EMEA Asia Pacific 150,000 US$ Millions 100,000 50,000 0 03 01 5 5 9 9 3 7 97 91 8 8 8 9 9 9 20 20 19 19 19 19 19 19 19 19 FIGURE 6â Computer hardware production by region, 1985-2004. 2004 data are a fore- cast. The graph includes parts PCsubassemblies such as base units that are specifically and industry-6.eps produced for use in computer equipment. SOURCE: Reed Electronics Research (2005). and Scotland for EMEA and Malaysia, and in Taiwan and China for the Asia- Pacific region. Some U.S. companies outsourced notebook production to Japanese, Tai- wanese, and Korean manufacturers but eventually shifted mostly to Taiwanese ODMs. In 2001, the Taiwanese government changed investment limitations for Taiwanese firms and the notebook industry moved en masse to the Yangtze River Delta near Shanghai. Japanese firms such as Toshiba moved their own notebook production to the region to take advantage of the supply base, but they also out- sourced much of their production. Chinese firms such as Lenovo used these same supply bases for their own production and outsourced some as well. â These locations are now changing once again. For example, Dell is moving final assembly and suppliers to Poland for EMEA; both Dell and HP are encouraging their CMs to move to India for the Asia region; and Dell is setting up final assembly in India. âSome notebook ODMs and suppliers moved to the area as early as 1998 so there was already a supply base when most of the industry moved. For example, Asustek had 300 employees in China in 1999 and 45,000 by 2005 (Einhorn, 2005). âThis was the case with the IBM PC Company and Lenovo both before and after their integration.
32 INNOVATION IN GLOBAL INDUSTRIES By 2005 China was the single largest producer of PCs and computer equip- ment in the world. Although the production facilities were located in China, they were mostly owned and managed by Taiwanese firms, such as HonHai/Fox- conn and Mitac for desktops, and Quanta, Compal, Wistron, and Inventec for notebooks. The supply chain was also composed largely of Taiwanese firms. Foxconn has a huge facility in Shenzhen that employs over 100,000 workers and produces base units and complete systems for nearly every branded PC vendor, while also assembling products such as game consoles and iPods and making components such as cables, connectors, chassis, and motherboards. Taiwanese ODMs produced 85 percent of all notebooks in the world in 2005 (Table 1), mostly in the Shanghai/Suzhou region of China. In the past, the location of final assembly was driven by the need for prox- imity to demand in the United States and Europe but now appears to be driven by growing demand in Asia as well as by the growing capability of firms to ex- ploit lower costs for labor, land, and facilities, the availability of cost-effective skilled labor, and government incentives in China. For instance, low-cost sea shipment of standard (not BTO) desktop PCs from China to the United States, supported by more sophisticated demand forecasting and planning tools, allows PC makers to build a 3-week shipment time into the new product introduction cycle. Notebooks can be economically shipped by air, so even BTO production can be centralized in Asia. Also, with most of the supply chain in Asia, it can be cheaper to assemble there and minimize shipment time for components because the supply base is concentrated there. GLOBALIZATION OF INNOVATION The location of NPD activities by the branded PC firms is driven by the product and process interdependencies discussed earlier, the capabilities and relative costs of different locations, and relational factors that tend to âpullâ in- novation outside the PC vendor or offshore. The relative capabilities and costs of U.S. firms and those in other countries have resulted in a new global division of labor: higher-value architectural design and business management, along with associated âdynamicâ and analytical engineering work, is done in the United States, whereas the development and manufacturing of the physical product, along with the more routine, âtransactionalâ product and process engineering, is done in Taiwan and increasingly in China. The result is that both component and system innovation is increasingly global, but U.S. firms continue to play leading roles in both. â After IBM sold its PC Division to Lenovo, only Dell (among the U.S. PC companies) had its own final assembly plant in China. Dellâs largest assembly site in Asia is still in Penang, Malaysia. â Dell is the only U.S. PC maker who still assembles desktop PCs in the United States; most final assembly of notebooks is centralized in Malaysia. The subassemblies come from the Pearl River Delta (desktops) and the Yangtze River Delta (notebooks) in China. Dell also does final assembly in China and other major markets.
PERSONAL COMPUTING 33 Capabilities and Cost The design of desktops and notebooks involves understanding markets and customer demand, as well as technology trends, anticipating how customer de- mand and technology trends are converging, and coordinating mixed teams of marketing people and technologists. It requires people with skills and experience in high-level architectural design, with the associated dynamic engineering skills, industrial design, and business and product management.10 In terms of proxim- ity, it is important to be located in leading markets where new technologies are developed and adopted first. Development for desktops or notebooks involves more routine, transactional product and process engineering. Therefore, it requires people with mechanical, electrical, and software engineering skills and technical project management experience. In addition, notebook development requires specialized skills in ther- mal and electromagnetic interference, shock and vibration, power management, materials, radiofrequency, and software. These require a combination of formal training and experience working in a particular engineering specialty, as well as working on the specific product type. Such knowledge and skill levels vary significantly in different locations due to at least three factors: (1) historical industrial development leading to creation of specialized skills, (2) output of educational systems, and (3) the nature of de- mand, including market scale and the extent to which the local or regional market may be described as cutting edge, with demanding and innovative customers. In the United States, there are business skills such as market intelligence and product management that are hard to find elsewhere. There are also leading industrial design firms that specialize in small electronic products such as note- books and cell phones, and strong software and high-level engineering skills. These skills are taught in universities, invested in by leading domestic firms in the industry, and honed through proximity to leading-edge users. In Japan, there are industrial designers that are very good at designing for the Japanese market, but who also have experience designing for global markets. Japanese engineering teams have deep skills in design and development, with specialties such as miniaturization that have developed to meet Japanese demand for small, lightweight products. Japan also is very strong in process engineering and manufacturing operations, thanks to its historical and continued emphasis on manufacturing. In Taiwan, mechanical and electrical engineers are available with strong 10â Gereffi and Wadhwa (2006) distinguish between dynamic and transactional engineers, a classi- fication that we find useful in characterizing the engineering workforces in different countries based on our interviews. Dynamic engineers are capable of abstract thinking and high-level problem solving using scientific knowledge and are able to work in teams and work across international borders. These engineers have at least 4-year degrees in engineering and are leaders in innovation. Transactional engineers have engineering fundamentals but not the skill to apply this knowledge to larger problems. They usually have less than 4-year degrees and are responsible for rote engineering tasks.
34 INNOVATION IN GLOBAL INDUSTRIES practical experience as well as formal training. Taiwanâs historical specialization in the PC industry, and with notebooks in particular, has created a pool of engi- neers with a great depth of knowledge of these products. Taiwan also has strong process and manufacturing skills. These have developed over time as Taiwanese firms have taken on greater responsibilities in PC development and manufactur- ing. Taiwan mostly lacks marketing skills and industrial design skills that would allow it to take over the concept and product planning stages, because of its focus on original equipment manufacturer/ODM production rather than development of branded products. China has many well-trained mechanical and electrical engineers, but most lack the hands-on skills that come with experience. Industrial design is weak, and marketing and business skills are very underdeveloped. A large number of engi- neers are produced each year, but quality varies greatly by university. According to one interviewee, Chinaâs engineers âwork perfectly at doing what they have been told, but cannot think about what needs to be done; they lack both creativity and motivation. They are good at legacy systems, but not new things; they canât handle âwhat ifâ situations.â In comparing cost across countries, the average salary for electronics engi- neers in all industries in the United States is about $80,000, compared to $60,000 in Japan, $20,000 in Taiwan, and under $10,000 in China (Dedrick and Kraemer, 2006b). Obviously there are cost advantages to moving engineering to China, but differences in productivity related to education and experience can negate the direct cost differences. Also, it is reported that engineering salaries are rising quickly in China, especially in industry clusters such as the Shanghai/Suzhou area, as multinationals and Taiwanese firms compete with domestic companies for talent. The willingness of multinationals to pay higher salaries gives them ac- cess to more experienced engineers and graduates of top universities, but turnover rates are high. Based on a survey of Taiwanese PC and electronics firms, Lu and Liu (2004) found that the main reason these companies were moving R&D (primarily de- velopment) to China was the availability of well-educated and cost-effective local engineers. This finding is supported by our own interviews with Taiwanese companies. As Taiwanâs supply of engineers has failed to keep up with demand, the attraction of a large pool of engineers with both linguistic and geographical proximity has been strong. This has enabled Taiwanese engineers to concentrate on more advanced development activities while lower-value activities such as board layout and software testing have moved to China. The New Global Division of Labor This confluence of product and process interdependencies with changing ca- pabilities and costs in different locations has led to a new global division of labor (Figure 7). In 1990, the entire NPD process was located in the United States (and
PERSONAL COMPUTING 35 Design Development Mfg. Product Design Proto- Pilot Mass Sust. Concept planning review type Prod. Prod. support 1990 United States Japan United States 2000 Japan Taiwan 2006 United States Japan Taiwan China FIGURE 7â New global division of labor in the PC industry. PC industry-7.eps Japan) in large vertically integrated companies like IBM, HP, Digital Equipment Corporation, and Toshiba, or PC specialists like Apple, Compaq, and Dell, which handled virtually all elements of system-level design and integration. By 2000, only design remained in the United States, while development and manufacturing of notebooks was outsourced mainly to Taiwan and manufacturing of desktops outsourced to major world regions. Japanese PC firms still kept NPD in-house, at least for higher-value products. In 2006, the U.S. position was unchanged. However, PC vendors like HP and Dell had set up design centers in Taiwan to manage NPD for some products (usually more mature product lines). Locating design in Taiwan allows closer coordination with CMs and ODMs and potentially speeds up NPD, allowing better quality control and problem resolution. They also use these design centers to transfer knowledge to the ODMs and to train locally hired hardware and soft- ware engineers to take on more project management and advanced development activities. This division of labor is similar for notebooks and desktops, although some U.S. companies keep desktop development in the United States and then outsource manufacturing to Asia. However, desktop development is being shifted to Taiwanese ODMs in many cases. The next critical development was the rapid shift of production to mainland China. Encouraged by U.S. PC vendors, Taiwanese manufacturers had moved the production of desktops and many components and subassemblies to the Pearl
36 INNOVATION IN GLOBAL INDUSTRIES River Delta near Hong Kong in the 1990s. Even more dramatic was the shift of notebook production to the Shanghai/Suzhou area after 2000. Many Taiwanese suppliers to the notebook industry had moved to China before 2001. When the Taiwanese government lifted its restrictions on notebook production in China, the ODMs and the rest of their local suppliers moved nearly all of their production to the mainland (Dedrick and Kraemer, 2006a). In response to U.S. PC makers outsourcing production to Taiwanese ODMs in China, the Japanese PC makers also shifted significant production to China, both through their own subsidiaries and through outsourcing to the Taiwanese ODMs. This further illustrates the compelling economics of the production bases in China as Japanese firms have previously tended to keep production in-house, either in Japan or in Southeast Asia. Chinaâs Expanding Role as a Locus of Innovation As a result of âproduction pullâ as well as the large pool of lower-cost en- gineering skills, there is an ongoing shift of product development activities from Taiwan to China. During our interviews with notebook makers in Taiwan and China, one major ODM told us that they did all of their board layout and most packaging design in China, while doing mechanical engineering and software engineering in Taiwan. They were in the process of training people in their elec- tronic engineering methods in China in order to move more development there. As one manager said, âChina is a gold mine of human resources, but if you donât get in and train them you wonât be able to take advantage of it.â It is expected that more of the NPD process and the associated engineer- ing tests will be conducted in China by many notebook makers (Dedrick and Kraemer, 2006a). These will be relocated from Taiwan and, in some cases, Japan. The shift of product development to China is distinguished not only by which activities have moved or are moving, but also by the type of products that are being developed. Some ODMs are moving product updates to China. However, the development of completely new products and platforms is still done by the ODMs in Taiwan, or by PC makers such as Lenovo (for Thinkpad notebooks) and Toshiba in Japan. More recent interviews with Taiwanese companies suggest that they are hesitant to move these activities to China. This is due in part to the high turnover rate of engineers in China, which makes it hard to create cohesive development teams and also raises the risk of intellectual property loss. Also, unless intellectual property protections are strengthened, China is not likely to become a center for advanced component-level R&D (e.g., in microprocessors, LCDs, or wireless technologies). A near-term division of labor for product development is likely to be as follows: component-level R&D, concept design, and product planning in the United States and Japan; applied R&D and development of new platforms in Taiwan; and product development for mature products, and nearly all production
PERSONAL COMPUTING 37 and sustaining engineering,11 in China. It is difficult to estimate how long this division of labor will last. A recent study of Taiwanese manufacturers (Li, 2006) shows that the rapid growth of low-margin outsourcing business from foreign multinational corporations has provided Taiwanese firms with the resources and motivation to invest more in R&D to develop greater technology expertise and capture more high-value design work. As the ODMsâ expertise grows, multina- tional corporations have greater incentive to outsource more design activities to further lower costs. Li also shows that Taiwanese firms are attempting to capture value from their innovation efforts by filing for more patents. So the shift from Taiwan to China may be slowing but the shift from the United States to Taiwan could continue. In addition, Taiwanese manufacturers such as Acer, Asus, BenQ, D-Link, and Lite-on have developed their own brand-name PCs, motherboards, monitors, networking equipment, smart phones, and other products. Acer and Asus brands have captured 14.1 percent of the world market for notebooks (Digitimes, 2006), whereas D-Link has become the top seller of wireless routers for the consumer market. As these companies enhance their R&D, design, and marketing capabili- ties, U.S. companies may find Taiwan to be a source of competition as well as cooperation. As China gains experience, it is still possible that the ODMs will shift more of the development process and newer products there, but, unless it becomes a key final market for PCs, it is not likely to capture the market-driven functions of concept design and product planning. As of now, Chinaâs PC market is still only about one-third the size of the U.S. market and does not have leading-edge users who are defining what features and standards are developed for the global market. However, as Chinaâs PC market continues to grow, and its users become more demanding, it may become the leading market at least for the Asia-Pacific region, and definition and planning of products suitable for the region may be done there. Finally, while Chinese brands remain minor players in the global PC industry for the most part, this may change. Chinese companies such as Lenovo, Huawei, and Haier are already leading brands at home and are expanding to in- ternational markets for PCs, network equipment, and other electronics products. 11â Sustaining engineering is the second of two phases in production; the first is mass production. Mass production involves the physical manufacturing of a product in large volumes. It requires manufacturing engineers to manage and plan the production process and test facilities and quality engineers to continually improve product and process quality. Over time, these engineers come to know the product extremely well and are best positioned to provide sustaining engineering support that was previously provided by the original product development teams. Sustaining engineering deals with changes that occur because of new chips, failing or end-of-life components, or improved components. Each change must be evaluated in terms of its implications for system performance and assembly, and incorporated into the production process. The sustaining engineers also provide the highest level of technical support when problems occur during use during a productâs 2- to 3-year warranty period.
38 INNOVATION IN GLOBAL INDUSTRIES Lenovoâs acquisition of IBMâs PC business has put it directly in competition with HP and Dell around the world, while Huawei uses its relationship with 3Com to access technology and markets and compete with Cisco and others. These com- panies can use the supply base of Taiwanese and foreign companies in China to match the multinationals on cost, develop products that fit the local market, and then target other emerging markets where innovations developed for the Chinese market are likely to be attractive. Measurement of the Globalization of Innovation Measuring the globalization of innovation is more difficult than measuring globalization of manufacturing, which can be captured in national production, trade, and foreign investment accounts. Innovation might be indirectly mea- sured by R&D spending and employees, patents, and new product introductions. While some public data on these measures are available, often the data are not sufficiently disaggregated at the firm level so that they can be tied to a product line such as PCs. This is especially true of multidivision firms such as HP, Fu- jitsu, Toshiba, Hitachi, Samsung, and Sony. Also, firm-level data do not show the extent to which R&D or other innovative activity is carried out in the home country or other locations. Given these difficulties, an alternative approach is to measure the innovation effort by the CMs and ODMs who are doing much of the manufacturing in the industry. The share of global notebook shipments produced by Taiwanese ODMs rose from 40 percent in 1998 to 85 percent in 2005 (Table 1). Since manufactur- ing and development are usually outsourced together, this suggests that the share of offshore product development activity has increased proportionately. This trend is supported by data showing that R&D spending by Taiwanese ODMs and CMs increased significantly from 2000 to 2005 (Table 2), as did the proportion of employees with Ph.D. and masterâs degrees in these firms. However, most of this R&D spending is on the development side rather than the research side. Also, reiterating a point made earlier that most innovation is done by up- stream component makers, the R&D spending by the ODMs and CMs, as well as nearly all of the PC makers, is minor in comparison to that of upstream suppliers. For example, Table 3 shows that in 2005 some of the lead PC makers12 spent 1.4 percent of revenues on R&D on average (weighted), the leading ODMs and CMs spent 1.3 percent, and the upstream suppliers, which is where innovation occurs in the PC industry, spent an average of 11.8 percent, or nearly nine times greater than the PC makers, ODMs, and CMs. 12â Wecould not get public estimates of R&D investment for the PC divisions of large multidivision companies such as HP, Fujitsu, Toshiba, Sony, and NEC, so they are excluded from the table.
PERSONAL COMPUTING 39 TABLE 1â Taiwanese Notebook Industry Share of Global Shipments, 1998-2005 1998 1999 2000 2001 2002 2003 2004 2005 Shipments 6,088 9,703 12,708 14,161 18,380 25,238 33,340 50,500 volume (thousands)a Global market 15,610 19,816 24,437 25,747 30,033 37,857 46,110 59,411 by volume (thousands) Taiwanâs share of 40% 49% 52% 55% 61% 66% 72% 85% global market volume aShipments by Taiwan-based firms, regardless of location of production. SOURCES: For 1998-2004, MIC (2005); for 2005, Digitimes (2006). Industry-Level Drivers of Globalization of Innovation The globalization of innovation in the PC industry has been driven primar- ily by economic factors and secondarily by relational factors that involve inter- dependencies of activities, as well as social networks that often influence the choice of suppliers or location. Examples of relational factors include the close interdependence between development and manufacturing of notebook PCs, and the âguanxiâ social networks that link Taiwanese firms and managers. TABLE 2â R&D Investment by Taiwanese ODMs and CMs (million U.S. dollars) Company Name 2000 2001 2002 2003 2004 2005 Quanta 27.13 38.36 54.55 74.31 92.56 102.36 Compal 24.77 44.69 62.11 70.21 78.78 Wistron 61.12 55.06 68.94 72.49 Asustek Computer 31.97 40.57 53.14 65.87 97.38 128.57 Mitac 24.37 24.70 25.28 32.66 36.90 46.62 Inventec 30.75 25.14 27.38 39.42 48.56 Arima 13.42 12.74 14.85 15.00 19.60 16.71 ECS 3.58 7.20 21.03 14.98 12.74 11.00 First International Computer (FIC) 28.21 10.91 46.72 44.58 Clevo 8.71 8.10 8.97 9.28 10.28 10.05 Twinhead 7.24 5.31 1.10 0.31 0.43 0.47 Uniwill 7.27 8.20 9.89 11.15 11.55 12.48 Foxconn (HonHai) 32.43 58.14 64.45 66.69 128.78 132.86 Subtotals 239.85 239.37 433.17 491.42 549.37 660.95 NOTE: Blank cells occur where data was not available in annual reports or elsewhere. SOURCE: Annual reports of the companies.
40 INNOVATION IN GLOBAL INDUSTRIES TABLE 3â R&D Investment as Percent of Firm Revenues, 2005 R&D as % Taiwan ODMs R&D as % Component R&D as % PC Makers of Revenue & CMs of Revenue Suppliers of Revenue Dell 0.9 Quanta 1.1 Microsoft 15.5 Apple 3.8 Compal 1.4 Intel 13.3 Gateway n.a. Wistron 1.6 AMD 19.6 Lenovo 1.7 Asustek 1.7 ATI Technology 14.7 Acer 0.1 Mitac 2.0 Seagate (HDD) 8.5 Inventec 1.4 Western Digital 6.6 (HDD) Arimaa 2.8 Maxtor (HDD) 7.5 ECSa 1.6 Chunghwa 3.4 (Displays) FICa n.a. Tatung (Displays) 2.6 Clevoa 4.2 AU Optronics 2.2 (Displays) Twinheada 0.2 Molex (Cables/ 5.2 connectors) Uniwilla 1.6 Delta (Power 4.8 supply) HonHai 1.0 Creative (Sound 6.7 cards) Total firm revenues $92,535 $76,191 $128,773 (millions) R&D (% of 1.4 1.3 11.8 revenues) for selected firms (weighted) NOTE: Large multidivision PC makers like HP, Toshiba, Sony, Fujitsu, and NEC are omitted because R&D investment is not available by division. aValue calculated from data in company annual reports. SOURCE: Electronic Business Top 300 (2006), unless otherwise indicated. Regarding economic factors, the manufacturing of desktops was primarily pushed offshore to major world regions to reduce production cost, and second- arily for proximity to markets. Manufacturing was then outsourced to CMs as most PC makers looked to further cut costs and concentrate on product design, branding, sales, and marketing. These CMs are currently moving to new loca- tions within each region (Eastern Europe for EMEA, Mexico for North America, and China for Asia-Pacific)âonce again to reduce costs. As noted earlier, for
PERSONAL COMPUTING 41 standard build-to-stock desktops, production is increasingly done in China for the U.S. market, because low-cost shipping by sea is viable when fast order turnaround is not necessary. Cost was also the key factor for notebooks, where both development and manufacturing were outsourced or offshored almost from the beginningâfirst to Japan, then to Taiwan, and currently to China. Japanâs capabilities with develop- ment and manufacturing of small form factors provided an initial pull, but lower costs, development of strong indigenous engineering capabilities, and the fact that Taiwanese firms were considered less likely to compete directly with U.S. firms resulted in U.S. PC vendors shifting to Taiwan. In turn, Taiwan has moved manufacturing to China for lower-cost labor, and manufacturing is now pulling some development activities to China as well. Taiwan is trying to expand its role in R&D, design, and other high-value activities, and PC vendors have facilitated this through continued outsourcing and by setting up design centers in Taiwan. Regarding relational factors in the PC industry, it appears that once produc- tion moves to a low-cost location, it will pull some higher-level activities to it. Reinforcing our findings about production pulling knowledge work, Lu and Liu (2004) found that the second major location factor for R&D (after access to low-cost engineers) is proximity to the manufacturing site. This is particularly true for notebook PCs given the importance of design-for-manufacturability. For example, production engineering and sustaining engineering clearly benefit from proximity to manufacturing, because production problems can be addressed im- mediately on the factory floor and engineering changes in existing products can be tested in production models from the assembly line. It also makes sense to move pilot production to China rather than to maintain an assembly line in Tai- wan just for this purpose. Then the question arises whether to move the expensive test equipment from Taiwan to China. If so, then there is more reason to relocate the design review and prototype processes as well. Beyond proximity considerations in manufacturing, there is a relational âpullâ from the ODMs. They often bundle development with manufacturing in order to win contracts. But once the ODM has a contract, the relationship creates incentives for the PC maker to work with the same ODM for future upgrades and enhancements to the product. In addition, there is a great deal of tacit knowledge created in the development process that is known only by the ODM, which cre- ates a further pull. Finally, the close linkage of development activities to manu- facturing and the feedback to design from manufacturing has created linkages that favor continuing the ODM relationships. The concentration of product development and manufacturing in Taiwan and China has reduced cost and accelerated new product innovation, driving down average unit prices, and helping to expand markets. For example, the worldwide average unit price for a PC and monitor has declined markedly over the past 15 years (Figure 8), with desktops and notebooks selling at an average of under $1,100 and $1,400, respectively, in the United States in 2005, and many models
42 INNOVATION IN GLOBAL INDUSTRIES 4.5 4 3.5 3 US$ (in thousands) 2.5 2 Desktops 1.5 Portables 1 0.5 0 00 05 04 03 02 01 0 96 94 98 99 92 95 93 97 91 9 20 20 20 20 20 20 19 19 19 19 19 19 19 19 19 19 FIGURE 8â Average unit price, desktops and notebooks, 1990-2005. SOURCE: Juliussen (2006). PC industry-8.eps available for well under $1,000. Of course, when adjusted for quality improve- ments, the price decline is much more dramatic. Moreover, the price differences between the United States and other regions have declined so that there is now effectively one world price. Beyond cost reduction, the globalization of innovation also has been driven by a desire to develop a better understanding of the needs of big emerging mar- kets such as China, India, and Brazil to enable the right versioning of existing products. Some PC vendors and ODMs (as well as other suppliers like AMD, Intel, and Microsoft) are seeking new markets in less-developed economies by developing new PCs with much lower price points while also tailoring the technologies to the more extreme environments of these countries. These new product concepts include the One-Laptop-Per-Child design, Intelâs Classmate PC, and Asusâs eeePC. While previous efforts to develop very-low-cost PCs for developing countries have failed, PC makers and others continue to experiment with new designs. IMPLICATIONS OF GLOBALIZATION OF INNOVATION The globalization of innovation has led to a new global division of labor as described earlier. This new international structure of the PC industry has implica-
PERSONAL COMPUTING 43 tions for firm competitiveness and strategy, location of innovation, employment, and U.S. policy. Implications for U.S. Firm Competitiveness Overall, the changes in the industry appear not to have hurt the competitive- ness of U.S. firms. U.S. companies dominate key components such as micropro- cessors, graphics and other chips, and hard drives, and PC vendors Dell, HP, and Apple hold nearly 40 percent of the world market for PCs. U.S. firms are still unquestioned leaders in operating systems and packaged applications. On the other hand, Asian firms are leaders in displays, memory, power supplies, batter- ies, motherboards, optical drives, and other components and peripherals. Asia has some leading PC brands such as Lenovo, Toshiba, Acer,13 and Sony, and Taiwanâs CMs and ODMs increasingly compete with U.S. contract manufactur- ers for outsourced development and manufacturing. On another measure of firm competitiveness, the largest share of industry profits flows to U.S. companies, particularly Microsoft and Intel, but also to Apple, Dell, HP, and to component makers such as Nvidia, TI, and Broadcom. The profitability of most Japanese and Asian companies is generally lower. Implications for Firm Strategy For branded PC vendors, the international innovation network described earlier enables faster product cycles with quicker integration of new technologies because the Taiwanese companies are good at fast turnaround and there is a good supply of cost-effective engineers in Taiwan and China to handle more models, changes, and upgrades. It has increased consumer choice, helped grow the mar- ket, and for a long time was advantageous for Dell because its direct model gave it an advantage in getting those products to the business customer. But now that most firms are efficient in minimizing inventory and getting new products into the market, the fast product cycles could be seen as an expensive race to the bottom that no PC vendor or component supplier really wins (except Intel and Microsoft).14 Some PC vendors complain that component innovation is too fast, 13â Acer, which has been a successful Taiwanese branded company, purchased Gateway Computer and Packard Bell in October 2007. 14â As desktop PCs in particular have become commoditized, business model innovations such as direct sales, BTO, and just-in-time inventory have provided temporary advantage in the industry. They provided an initial advantage to Dell and Gateway, who were the first to adopt direct sales, but Gateway stumbled badly and Dellâs efficiency advantage has been reduced as other PC vendors have gone to direct BTO sales. The Dell model also has proved less successful in overseas markets where direct sales are less popular than in the United States. The most important impact of past business model innovation has been a general improvement in the efficiency of the industry as a whole, as most vendors have adopted these practices.
44 INNOVATION IN GLOBAL INDUSTRIES and they feel pressured to introduce too many products for too small markets. For example, one major PC vendor introduces around 1,000 different consumer desktop SKUs (stock-keeping units) in one year globally (Dedrick and Kraemer, 2006b). A question raised by more than one company that we have interviewed is whether the cost of managing so many products might outweigh the benefits of being able to offer products that more closely match the needs of customers. Beyond desktop and notebook PCs, the growing demand for new products that are smaller, are more mobile, and integrate new functions is bringing new innovation and new players into the personal computing industry. Hit products such as RIMâs BlackBerry and Palmâs Treo have been developed by firms with no traditional PC business, while Appleâs iPod was developed on an entirely dif- ferent platform from the Macintosh computer line. Such radical or architectural product innovation (Henderson and Clark, 1990; Utterback, 1990) has important differences from the incremental model of development as illustrated in Table 4. The scale and scope of global collaboration is often greater for radical innova- tion, as existing technologies are adapted to new uses and new technologies are developed. As a result, there is greater need for joint development with partners, while key technologies (particularly software) are developed internally and the entire process is shaped by strong central vision, integration, and control. An example of the nature of radical innovation is the iPod, which was developed by Apple in collaboration with many external partners in multiple geographic locations. Apple used its internal capabilities to create a closely integrated hardware and software design, while relying on outside partners for both standard and custom components, and for manufacturing. For instance, Apple used a reference design and worked jointly with PortalPlayer to develop the microchip that controlled the iPodâs basic functionality. It worked with oth- ers for additional chips (e.g., United Kingdomâs Wolfson Microelectronics for the digital-to-analog sound chip; New York-based Linear Technology for power management chips; California-based Broadcom for a video decoder chip); with Toshiba for the 1.8-inch hard drive; and with Taiwanâs Inventec for manufactur- ing (Murtha et al., forthcoming). Apple designed the system architecture that affected critical features such as sound quality and power consumption and developed the distinctive industrial design of the iPod; it developed most of the iPod and iTunes software in-house or adapted othersâ software. Apple tightly managed the whole process, coordinating closely with outside partners so that it could design the iPod, and its manufac- turer and suppliers could concurrently prepare the tooling and supply chain for large-volume manufacturing, and bring it to market in 8 months. As put by the iPodâs lead engineer, âToday, there is too much complexity in products for one person or organization to understand. You need a team of internal and external resources working with you to conceive, design, and implement new productsâ (Murtha et al., forthcoming). The resulting design process is much different from
PERSONAL COMPUTING 45 TABLE 4â Features of Incremental and Radical Innovation Design Development Production Radical â et system architecture, S â ollaborate with many C Outsourced innovation sometimes building on partners in multiple to CM or (iPod, iPhone, external reference design geographies ODM Treo) â trong central vision and S â ollaborate with partners of C industrial design partners â ightly control all aspects T â et partners to adapt G of NPD existing technologies to â evelop key software D proprietary architecture internally â ntegrate hardware, software, I even services (e.g., iTunes, iTMS) â esign or license D complementary assets (SW, content) and distribution system â ollaborate closely with a C few key partners for core components Incremental â nnovate on Wintel I â ollaborate with one C Outsourced innovation architecture established ODM in one to ODM (desktops, â ontrol product planning, C geography notebooks) brand image, marketing, â utsource detailed physical O concept design internally design, test, and software â nternal or outsourced I built within standard industrial design architecture â W and SW are modular H â everage existing L complementary resources and distribution that in PCs, with more internal development and much closer interaction with key component suppliers. Finally, for the iPod to be successful in the market, Apple created a new business model that integrated hardware, software, and online content delivery. It developed iTunes software to collect and manage content on a PC or Mac and easily transfer that content to the iPod. It also developed the online iTunes Mu- sic Store and tightly integrated that with the iTunes application. Apple licensed content from all the major music labels and subsequently from the audio book, movie, and television industries, and established pricing and digital rights models that were attractive to consumers. The result was a U.S. market share of over 70 percent in both the personal music player and the music download markets.
46 INNOVATION IN GLOBAL INDUSTRIES Given that such design innovation has the potential for creating differen- tiation in products and gaining competitive advantage, the strategies of at least some branded PC firms are likely to focus more on creating new product plat- forms. However, examples such as the iPod, Treo, and BlackBerry suggest that radical innovation requires a different process of new product development. As illustrated by our earlier discussion of these innovations, elements of the process include leveraging a firmâs unique internal capabilities with those of external partners; working closely with external partners in multiple geographies; engag- ing in a global search for technologies that can be adapted and integrated into new products; maintaining tight architectural and managerial control over the process; and possibly introducing new business models to provide complementary content and services. This kind of process is far removed from the incremental innovation within a well-established product architecture and the mature market of the Wintel PC world. As a result, it has been more diversified companies such as Samsung and Sony, wireless specialists such as Nokia, as well as many startups that are trying to innovate with new product platforms that mix communications, entertainment, and computing capabilities in smaller form factors. In these cases, firms have worked with outside partners to exploit external sources of knowledge while keeping their own innovative activities mostly in-house and close to their home base. Increasingly, hardware-software integration is becoming important as a means of tailoring products to different market requirements such as communications standards, power consumption, language, and customer tastes. Such integration also helps to reduce product costs by enabling standard physical platforms to be produced in large volumes for global sales. More important, it enables greater product differentiation for ever-finer market segments by customizing through changes in software, rather than through costly physical changes in hardware. Location of Innovation Innovation at the national level is closely tied to the presence of both tech- nically skilled and entrepreneurial individuals, the quality of infrastructure, and the presence of advanced users who drive firms to innovate. Rapid diffusion of Internet infrastructure in the United States led to ongoing innovation in hardware (e.g., routers, switches), software (e.g., browsers, search engines), and services (e.g., online retailing, banking, stock trading, travel services). The United States has seen strong user-driven innovation (Von Hippel, 1998) such as IT-enabled business process redesign and e-commerce in the corporate world and user- created content in the consumer world. From Cisco and Amazon, to Dell and Wal-Mart to Google and MySpace, innovation on the web has largely occurred in the United States. By contrast, the relatively slow adoption of broadband and advanced mobile
PERSONAL COMPUTING 47 technologies in the United States has left the country falling behind in new areas of innovation. For instance, South Korea is a leader in online computer gaming, thanks in part to its widespread deployment of cheap broadband Internet service. Japanâs iMode system for mobile Internet was years ahead of similar services in the United States. High rates of wireless adoption have benefited firms from South Korea, Japan, and Northern Europe, while Chinaâs large mobile phone market has attracted firms such as Motorola, Nokia, and Siemens to do product development there. In short, the lack of innovation in industries that are providers of complementary assets (which in turn may reflect the outmoded infrastructure underpinning the large and otherwise highly sophisticated U.S. domestic market) is a major factor hampering innovation in the PC industry. If the United States is to retain its position as a leading market for computing innovation, it cannot af- ford to remain behind in providing high-quality, low-cost infrastructure to support user-led innovation and drive demand for new personal computing products. Our field interviews indicate that design innovation, especially concept de- sign and product planning, is likely to remain concentrated in the United States for the major U.S. firms in the personal computing industry. However, there will be increasing use of offshore R&D and design centers in locations that have specialized and cost-effective talent, that lead in particular technical innovations, or that represent important markets in terms of growth potential, special market opportunities (fewer regulatory requirements, government incentives), or chal- lenges (need for cheaper or environmentally friendly PCs), or that may influence technical standards (as China is trying to do in a number of technologies). Private interviews with industry executives indicate that the primary motivation for such offshore outposts is cost reduction, through hiring less costly engineers, program- mers, and managers to perform activities previously performed in-house in the United States or in a foreign subsidiary. In time, secondary benefits may also arise as these locations gain capabilities or as local markets develop. Other product development activities tend to be pulled by production, begin- ning with manufacturing process engineering, then moving up to prototyping and testing and eventually electrical, mechanical, and software engineering. These are in the process of shifting to China from Taiwan and Japan, although R&D, design, and development of the newest generation of products is still likely to be concentrated in the home countries of the manufacturers (Dedrick and Kraemer, 2006a). Impacts on Jobs and Employment With respect to U.S. workers, much of the potential shift of jobs offshore has already taken place with the offshoring and outsourcing of production from 1990 to 2005. There has also been a shift in innovation-related jobs after 2000, as production has pulled development and some design activities to Asia (Dedrick and Kraemer, 2006a). Further movement of jobs offshore is likely to occur in
48 INNOVATION IN GLOBAL INDUSTRIES the future to meet competitive pressure for continuous cost reduction. The jobs will be in engineering, software, industrial design, engineering management, and project management at all levels. As one PC industry executive told us in interviews, he has to âpushâ more physical design and project management jobs overseas in order to keep concept design jobs at home. The number of jobs directly moved offshore is not large and occurs incre- mentally. However, another indicator of the impact of offshoring is the number of new jobs that are created offshore rather than in the United States to support the industryâs continued growth and proliferation of products. One indicator of this impact is the growth of knowledge jobs in the notebook industry in Taiwan as these firms take on more design and development activities for the United States and other firms. Interviews and company data on the top ODMs in the notebook industry indicate that they hired thousands of new R&D personnel and product engineers in Taiwan between 2000 and 2005, while also hiring thousands more for product and process engineering, testing, and production in China. For example, Quanta, which is the largest notebook ODM, has increased the number of R&D engineers from 750 in 2001 to around 7,000 in 2005 (company annual reports). As software becomes an increasingly important part of new PC products, there will be a proportionately greater increase in software jobs being moved offshore. In one company we interviewed, 50 percent of the 1,000 employees are engineers and 80 percent of these are software engineers. These jobs are currently in the United States, but the firm is experimenting with offshore teams. While there is broad awareness of the shift of jobs to India and elsewhere by software and IT services companies, there is less awareness of the number of software jobs within the computer hardware industryâjobs that are likewise vulnerable to offshoring. For the United States, the fact that growth and innovation in the industry are not creating new knowledge jobs (engineering, software, design) in the United States but are creating them in Taiwan and China appears to be a negative. But the number of U.S. engineering jobs in the broader computer industry is fairly stable at about 60,000 between 2002 and 2005 (Dedrick and Kraemer, 2006b), and without globalization there may not be as much growth and innovation. The risks of globalization for the United States are that individuals, firms, or related industries will lose technological advantage and the ability to innovate. A Korn/ Ferry International report posed the issue for industry executives as follows: North American industrial executives must choose between two fundamental responses to their current competitive environment. One approach is to simply accept that their companies need to focus exclusively on marketing, finance and the design and development functions, while offloading their manufacturing needs and technologies to more accommodating locations, usually overseas. While this strategy can generate short-term profits, it almost inevitably guaran- tees that a company will lose control of its design and production capabilities.
PERSONAL COMPUTING 49 Eventually, if history is a reliable guide, even home office and corporate func- tions will cease to exist. (Kotkin and Friedman, 2004) However, earlier industry innovations as well as recent innovations like the iPod, the Treo, and the Microsoft Xbox were developed mostly in the United States, even though some component innovations came from offshore suppliers and all the manufacturing was done offshore. Moreover, there is little evidence thus far that these firms have âlost controlâ of the designs or technology for these products. Such innovation is less likely to move offshore and should continue to support engineering and other knowledge jobs in the United States, as long as the United States retains the capabilities needed for such innovation. Implications for Policy: Sustaining U.S. Innovation Leadership Although U.S. PC vendors still lead innovation in the industry, they are mov- ing more innovation activities offshore both through setting up design centers and through outsourcing design and development activities to ODMs. The U.S. suppliers of key components such as microprocessors, storage, and software are also setting up R&D and design centers offshore, sometimes in locations with specialized skills such as Israel or Japan, and sometimes in big emerging markets with low-cost engineering talent such as India and China. The engineering, software development, and management skills associated with these activities are key to the innovation capabilities of the United States and therefore consideration needs to be given to developing people with these skills if such innovation is to remain in the United States (Committee on the Engineer of 2020, National Academy of Engineering, 2005). Our interviews with executives indicate there is a growing need across the PC industry for engineers who are specifically trained to work at the interface between hardware engineering, com- munications, and computer science. The executives also indicate that many U.S. engineering schools produce specialists in a single engineering discipline, but few schools produce people who can work at the interfaces of these disciplines. There is a need, for example, for hardware engineers who can work with communica- tions standards, and software engineers who can produce embedded software that enables customization of products for markets. When universities fail to develop such talent, firms may rely on on-the-job training, look offshore for experienced people with the needed skills, or develop the skills offshore through on-the-job training of low-cost specialists. It is also likely that U.S. firms need to make greater efforts to hire rookies and develop them. Several of the companies we interviewed prefer to hire fairly experienced engineers rather than beginners and report no problems in doing so in Silicon Valley or elsewhere. They simply hire people away from other com- panies, or bring in engineers from foreign countries under immigration policy. However, one highly innovative company we interviewed hired engineers as
50 INNOVATION IN GLOBAL INDUSTRIES interns from the best engineering schools in the United States (e.g., Cornell, MIT, UC Berkeley, Carnegie-Mellon) and, if they worked out, made commit- ments to hire them even before they graduated. Starting as interns, they worked as part of project teams with operational roles and real challenges to overcome. Such on-the-job training can help sustain a career ladder for new engineers as firms offshore more lower-level jobs that would normally be filled by entry-level engineers. An executive for the firm argued that this process benefits the firm as well, by giving it access to the best talent available and the chance to incorporate that talent into product development teams and learn how the company works before the engineers develop bad habits elsewhere. From a policy perspective, the U.S. government can encourage cross- disciplinary education and more university-industry cooperation through its funding choices, and by documenting and publicizing the need for such changes. While universities are responsive to employer needs, there can be significant iner- tia in academic departments and university bureaucracies, and external resources and pressure can encourage greater responsiveness and flexibility. All of the firms we interviewed indicated a need for more H-1B visas, or for reform of the visa process. One issue involves procedures for keeping people who have been educated in the United States and perhaps interned with the firm. Another involves recruiting from abroad for skills for which the U.S. supply of talent is limited, but for which other countries are noted for having people with the needed skills. For example, it appears that the supply of engineers in analog fields in the United States such as radiofrequency is limited, whereas there is a good supply in some European countries. A reported problem with the current immigration process is that the nature of U.S. supply of talent is not considered. From an immigration standpoint, an engineer is an engineer regardless of educa- tion level (bachelor, masterâs, Ph.D.) and there is no way to identify and respond to shortages of very specific skills or levels (e.g., bachelor vs. Ph.D.). In addition to such human resource issues, another key concern is sustain- ing the demand for innovation. PC demand, and associated innovation, has been driven in the past decade largely by the Internet and networking in general. With the United States leading in Internet adoption, the PC industry was quick to adopt networking technologies such as Ethernet and wireless networking, and new products such as the BlackBerry and Treo were developed in the United States. However, the United States has fallen behind a number of countries in both wireless and broadband adoption and is not the lead market for products and services such as mobile phones and online gaming. As a result, innovations in new personal computing devices such as smart phones, video game consoles, and other network devices are likely to target foreign markets initially, making it more likely that innovation will occur in those markets rather than in the United States. While specific policy issues with regard to telecommunications, Internet regulation, content, and pricing are beyond the scope of this chapter, those deci-
PERSONAL COMPUTING 51 sions should be made with an awareness of their potential impact on U.S. inno- vation in industries such as personal computing. Innovation in PCs can require cooperation by providers of complementary assets, such as content or communi- cation infrastructure. Government policies on telecommunications can influence the speed of diffusion of infrastructure like broadband, 3G, or municipal WiFi networks. Similarly, government policies on copyright can influence the terms under which content can be distributed. While these policy issues are usually debated in terms of impacts on competition, intellectual property rights, or even consumer choice, policy makers also should consider their impact on innovation in high-technology industries. ACKNOWLEDGMENTS The research on which this chapter is based has been supported by grants to the Personal Computing Industry Center of the Paul Merage School of Business at UC Irvine from the Alfred P. Sloan Foundation, the U.S. National Science Foundation, the National Academy of Sciences, and the California Institute for Telecommunications and Information Technology. Any opinions, findings, and conclusions or recommendations expressed in these materials are those of the author(s) and do not necessarily reflect the views of these sponsors. The authors acknowledge the very helpful comments of David Mowery, Jeffrey Macher, and anonymous reviewers on drafts of this chapter. REFERENCES Committee on the Engineer of 2020, National Academy of Engineering. (2005). Educating the En- gineer of 2020: Adapting Engineering Education to the New Century. Washington, D.C.: The National Academies Press. Curry, J., and M. Kenney. (1999). Beating the clock: Corporate responses to rapid change in the PC industry. California Management Review 42(1):8-36. Dedrick, J., and K. L. Kraemer. (1998). Asiaâs Computer Challenge: Threat or Opportunity for the United States and the World? New York: Oxford University Press. Dedrick, J., and K. L. Kraemer. (2005). The impacts of IT on firm and industry structure: The personal computer industry. California Management Review 47(3):122-142. Dedrick, J., and K. L. Kraemer. (2006a). Is production pulling knowledge work to China: A study of the notebook computer industry. Computer 39(7):36-42. Dedrick, J., and K. L. Kraemer. (2006b). Impacts of Globalization and Offshoring on Engineering Employment in the Personal Computing Industry. Report for the National Academy of Engineer- ing. Irvine, CA: CRITO. Digitimes. (2006). ICT Reportâ1Q 2006: Taiwanâs Notebooks. Taipei: Digitimes Research. Economist. (2006a). The dream of the personal computer: The PCâs 25th birthday. 380(8488):17. Economist. (2006b). Getting personal: The PCâs 25th birthday. (July 27):57-58. Einhorn, B. (2005). Why Taiwan matters. Business Week. May 16. Electronic Business Top 300. (2006). Electronic Business August:3-14.
52 INNOVATION IN GLOBAL INDUSTRIES Gereffi, G., and V. Wadhwa. (2006). Framing the Engineering Outsourcing Debate: Placing the United States on a Level Playing Field with China and India. Duke University, Master of Engineering Management Program. Grove, A. (1999). Only the Paranoid Survive. New York: Time Warner. Henderson, R. M., and K. B. Clark. (1990). Architectural innovation: The reconfiguration of exist- ing product technologies and the failure of established firms. Administrative Science Quarterly 35(1):9-30. IDC (International Data Corporation). (2006a). Worldwide Black Book, Q2 2006. Framingham, MA: IDC. IDC. (2006b). Worldwide PC Market: 4Q05 and 2005 Review. Framingham, MA: IDC. IDC. (2006c). Worldwide Software 2006-2010 Forecast Summary. Framingham, Mass.: IDC. IDC. (2006d). IDC Worldwide Quarterly PC Tracker. December 2006. Juliussen, E. (2006). Worldwide PC Market: May, 2006 Version. Arlington Heights, IL: eTForecasts. Kotkin, J., and D. Friedman. (2004). Executive leadership in the industrial economy. Korn Ferry International. Available at: http://www.kornferry.com/Library/ViewGallery.asp?CID=865& LanguageID=1&RegionID=23. Kraemer, K. L., J. Dedrick, and S. Yamashiro. (2000). Dell Computer: Refining and extending the business model with information technology. Information Society 16:5-21. Li, W. C.-Y. (2006). Global Sourcing in Innovation: Theory and Evidence from the Information Technology Hardware Industry. Los Angeles: UCLA. Working paper. 46 pp. Lu, L. Y. Y., and J. S. Liu. (2004). R&D in China: An empirical study of Taiwanese IT companies. R&D Management 34(4):453-465. Mann, C. (2003). Globalization of IT services and white collar jobs: The next wave of productivity growth. International Economics Policy Briefs PB03-11:1-13. MIC (Market Intelligence Center). (2005). The Greater Chinese Notebook PC Industry, 2004 and Beyond. Taipei: Asia IP Report Murtha, T., F. Giarratani, and T. Sturgeon, eds. (Forthcoming). Massive Coordination: Creating and Capturing Value in Global Knowledge Networks. Porter, M. (1996). What is strategy? Harvard Business Review (November-December):1-22. Reed Electronics Research. (2005). Yearbook of World Electronics Data. Vols. 1-4. Utterback, J. M. (1990). Radical innovation and corporate regeneration. Journal of Research Technol- ogy Management 37(4):10-21. Von Hippel, E. (1998). Economics of product development by users: The impact of âstickyâ local information. Management Science 44(5):629-644.