The Practice of Global Product Development (With Updates by Steven D. Eppinger)

J U LY 2 0 0 9

Steven D. Eppinger and Anil R. Chitkara

R E P R I N T N U M B E R 5 0 4 3 7

Special article courtesy of:

The Practice of Global Product Development

G lobalization pressures have begun to have a major impact on the practice of product development across a wide range of industries. A new paradigm has emerged whereby companies are utiliz-ing skilled engineering teams dispersed around the world to develop products in a collaborative manner. Best practice in product development (PD) is now rapidly migrating from local, cross-functional collaboration to a mode of global collaboration. Global product de-velopment (GPD) therefore represents a major transformation for business, and it applies to a broad range of industries.

The objective of this article is to present frameworks that can help companies ad-dress various strategic and tactical issues when considering adoption of GPD. The concepts have been developed mainly through detailed discussions with manag-ers at more than 100 companies in 15 countries in North America, Europe and Asia. Some data are from a recently com-pleted study on GPD that product development company PTC has conducted with BusinessWeek Research Services, in-terviewing and surveying more than 1,100 engineering managers worldwide. (See “About the Research,” p. 2.)

In our discussions with managers,

many have found the ideas, frameworks and perspectives presented in this article to be helpful in addressing the transforma-tion to global product development and its implementation today. There is no blue-print, but senior managers can more effectively plan for global product devel-opment and take fuller advantage of its promise by examining the various strate-gies, staged approaches and key success factors described herein and adapting those insights to their own unique set of circumstances.

Defining Global Product Development

Several best practices in product develop-ment evolved through the 1980s and 1990s. By 2000, it had become widely accepted that highly effective product development in-cluded co-location of cross-functional teams to facilitate close collaboration among engi-neering, marketing, manufacturing and supply-chain functions. Co-located PD teams could concurrently execute the range of activi-ties involved, from understanding market and customer needs, through conceptual and de-tailed design, testing, analysis, prototyping, manufacturing engineering and postsales technical product support/engineering. This concurrent engineering practice resulted in

JULY 2009 MIT SLOAN MANAGEMENT REVIEW 1WWW.SLOANREVIEW.MIT.EDU

Many manufacturers have established prod-uct development activities in different countries around the world. Yet their senior managers often strug-gle to tie those decentralized organiza-tions into a cohesive, unified operation that can efficiently drive growth and innovation. New empirical frame-works may help unlock practices with which managers can deploy well-coordinated global product development strategies.

BY STEVEN D. EPPINGER

AND ANIL R. CHITKARA

Editor’s Note: MIT Sloan Management Review first published “The New Practice of Global Product Development” in our Summer 2006 edition. In the article, authors Steven D. Eppinger and Anil R. Chitkara examined state-of-the-art and emerging best practices in global product de-velopment (GPD). GPD continues to evolve, and in 2009, we asked Eppinger, the General Motors Leaders for Manufacturing Professor of Management Science at the MIT Sloan School of Man-agement, to update the article with new insights about GPD. His additions are published here alongside the 2006 article, annotating the original text.

better product designs, faster time to market and lower-cost production. PD activities were generally located in corporate research and development centers, which maintained linkages to manufacturing sites and sales offices around the world.

By contrast, the emerging best practice in PD today utilizes a highly distributed, networked development process facilitated by a fully digital PD system. Global product development combines certain centralized functions with some engineering and related PD functions distributed to other sites or regions of the world. This practice may involve outsourced engi-neering work along with captive offshore engineering facilities. The benefits of GPD are beginning to be-

come clear. They include greater engineering efficiency (through utilization of lower-cost resources), access to technical expertise that is distributed internationally, design of products for more global markets and more flexible PD resource allocation (through use of outsourced staff). (See “Comparing the New Practice of Global Product Development With the Conven-tional Approach,” p. 4.)

The academic literature on global R&D offers a rich variety of perspectives, using terms such as virtual teams, distributed development, international R&D and other variants. Oliver Gassmann and Maximilian von Zedtwitz1 present several alternative models for organization of global teams and a list of factors affecting choice of locations for GPD operations, along with a literature review on manag-ing GPD teams. Jose Santos, Yves L. Doz and Peter Williamson2 argue that a truly global innovation process uses global

R&D sites both to gain access to foreign technology expertise and to access foreign market knowledge.

However, much of the academic discussion of GPD has been about what it is and why it should be done. There has been less focus on frameworks that managers can use to decide how to implement their decisions to adopt GPD as a corporate practice. Cur-rent views of globally dispersed PD are muddied by anachronistic assumptions about labor rates, vague-ness about the value of intellectual property, outdated ideas about a company’s core competen-cies and more. As executives make GPD a more strategic priority, the implications of not getting it right may have a significant impact on their busi-nesses. It is important, therefore, to reframe the GPD discussion in light of today’s market dynamics, en-abling tools and underlying infrastructure. We extend popular definitions of GPD by characterizing it as a single, coordinated product development op-eration that includes distributed teams in more than one country utilizing a fully digital and connected, collaborative product development process. This may include third parties that provide engineering or design capacity, or it may be an entirely captive, company-owned operation.

Why Is the GPD Transformation Happening Now?

Over the past five years, many industries have seen a rapid shift to global product development. In a 2003 Deloitte Research study3 of North American and Western European manufacturers, 48% of the companies surveyed had set up engineering opera-tions outside of their home region. In fact, 22% of the North American manufacturers already had lo-cated engineering functions in China, as did 14% of the Western European manufacturers polled. In just

2 MIT SLOAN MANAGEMENT REVIEW JULY 2009

UPDATE EPPINGER: The key message in the article is still valid and highly relevant today. Global

product development (GPD) is not the only way to do business, but it is one of the organiza-tional choices

companies can make. It is a relatively new way to do business that has been en-abled over the last 15 to 20 years by some important trends: the digitization of prod-uct development, the growth of engineering talent in lower-cost, high-growth economies, and increasing globalization of business through extended business ecosystems. In recent years, companies’ drives for internal efficiency along with expansion into new global markets have accelerated the trend toward adoption of GPD.

ABOUT THE RESEARCHThe authors set out to chart the new practice of global product development (GPD) in the manufacturing sector with the goal of presenting

frameworks that would help senior managers more effectively address strategic and tactical GPD issues. The frameworks have been devel-oped chiefly through detailed discussions with managers at more than 100 companies in 15 countries in North America, Europe and Asia. Data inputs also come from a study that PTC conducted recently with BusinessWeek Research Services.

The initial qualitative phase of the PTC research, conducted in the first quarter of 2006, involved telephone interviews with 30 executives from large manufacturing companies that currently practice GPD. Twenty of the interviewees were U.S. executives, five were European and five Asian. Each interview lasted 30 to 40 minutes. Findings from the interviews were used to design an online survey featuring more than 40 questions about GPD. In March 2006, 1,157 online surveys were completed by product development–oriented executives and engineering/de-sign professionals from large manufacturing companies. Respondents from the United States comprised 65% of the sample, with 17% from Europe and 18% from Asia.

WWW.SLOANREVIEW.MIT.EDU

JULY 2009 MIT SLOAN MANAGEMENT REVIEW 3

the last two months of 2005, Microsoft, Cisco and Intel each announced major investments in product development operations in India totaling $3.8 bil-lion, according to the companies’ press releases.

As noted in the literature, global R&D net-works have been utilized for many years. Two relatively recent factors, however, are now making feasible a truly integrated, yet distributed PD pro-cess. First, product design processes today are fully digital and completely networked. Computer de-sign tools are the norm; high-bandwidth networks are ubiquitous. As author and New York Times col-umnist Thomas Friedman writes, such a digital business process enables its distribution across the globe in today’s “flat world.”4 Second, many more businesses now have experience with global col-laboration. Throughout the 1980s and 1990s, many U.S., Western European and Japanese man-ufacturing companies located production operations in regions where labor cost much less than at home. Such organizations now have deep experience with globally distributed operations and suppliers. Many multinationals have also grown by acquisition of regional companies whose operations have since been integrated. Most man-ufacturers today know how to collaborate across global supply chains, and this experience applies directly to collaboration in GPD.

Companies are building their GPD capabilities today for any of four reasons:

Lower Cost Many companies strive to reduce PD operating costs by redistributing activities to take advantage of labor arbitrage or to access more af-fordable capabilities. There is a huge pool of engineering talent in low-cost regions such as China, the Czech Republic, India and Vietnam — and in medium-cost nations including South Korea, Hun-gary, Poland and Taiwan. (We consider “low-cost” to be 10% to 20% of the equivalent engineer’s salary in the United States, and “medium-cost” to be 20% to 50% of U.S. wage rates.5)

Improved Process Many engineering managers can recall the key lesson learned from both the 1980s em-phasis on design for manufacturing (DFM) and the 1990s emphasis on time to market (TTM). This

lesson was that co-locating development teams — par-ticularly the design engineers with the manufacturing engineers — yielded both the cost benefit of DFM and the agility benefit of TTM. The prospect of moving de-sign engineering to global manufacturing locations can be attractive again today.

COORDINATION INEFFICIENCYEPPINGER: One concern that we did not clearly articulate in the article – but my work since has explored – is the coordi-

nation inefficiency that is sometimes created with global product development processes. In my subsequent research, I have found that it is ac-tually very hard to keep processes efficient as you distribute them. Imagine a highly focused,

co-located local process that you have made as lean and efficient as possible. When you disrupt that process and distribute it globally, it becomes less lean; in particular, it becomes less efficient to implement all the communications required. Although it is theoretically possible for people in all parts of the world to collaborate using the Internet and all the other communications technologies we have, that doesn’t mean it is easy. I believe that we have underestimated the coordination cost penalties involved in distributing pro-cesses globally.

However, the research I have done since this article was originally published also suggests a solution to the prob-lem of coordination inefficiency. It turns out that you achieve coordination efficiency not by simply distributing what looks like the easiest task for a foreign site to do – or by distributing globally those components that may be the most neatly decomposed or disconnected. Instead, the key is to deeply understand the structure of the product devel-opment process.

In particular, you need to understand which activity needs to connect with which other activities -- and who makes those connections. When you do that analysis, you find that not everybody has the same amount of interconnection with everybody else. People have a few connections that are strong and many connections that are weaker. And, if you keep that in mind, then you can create a global product de-velopment process that doesn’t require too much high-bandwidth communication to happen over great dis-tances. In fact, a well-designed global product development process could require most of the high-bandwidth commu-nication to happen within local sites -- with less communication required over the distances. If you maximize the communication within each site, you can get a lot of effi-ciency within local pockets, and then those local groups coordinate as necessary with other efficient local groups. In other words, look not only at the modularity of the product or system that you’re creating but also at the modularity of the product development process itself.

WWW.SLOANREVIEW.MIT.EDU

4 MIT SLOAN MANAGEMENT REVIEW JULY 2009 WWW.SLOANREVIEW.MIT.EDU

Global Growth Locating some PD activities in se-lected international locations can give companies access to critical information about markets in those regions. By using local engineers, companies make direct connections with potential new markets.

Technology Access Companies are using GPD to de-velop integrated PD processes that include engineers in regions where critical new technology has been devel-oped and where technical experts reside.

Although cost remains the primary reason that many companies initially consider GPD, it is tech-nology, process innovation or revenue growth that drives a GPD strategy. This move from cost to growth and innovation has been a major shift in stated GPD objectives over the past two to three years.

Who Is Leading the Way to GPD? GPD has been adopted across many industries and

in many regions around the world. Some industries and regions, however, seem to be embracing it more quickly than others. A recent PTC/BusinessWeek Re-search Services survey of 1,157 engineering managers at manufacturing organizations across the United States, Europe and Asia found that 70% of the compa-nies were either planning or executing GPD.6

Some Industries Are Rapidly Adopting GPD It is not surprising that software developers adopted GPD quickly. By the early 1990s, as the Internet began to enable global connectivity, leading com-panies were taking advantage of development operations in several countries. More recently, companies such as Microsoft, Accenture, Siemens, Intel, Hewlett-Packard and Toshiba have located

software development operations in Bangalore. We have also seen the rapid growth of outsourced soft-ware development and support with Indian suppliers such as Infosys, TCS, Wipro, Satyam, HCL Technologies and others. Software development now represents approximately one-third of India’s service exports.7

Electronics manufacturers led the way to pro-duction outsourcing, and most major manufacturers are now taking advantage of the density of electron-ics design expertise in the Far East. Their GPD efforts are not limited to lower-cost regions. East-man Kodak recently created its Digital Product Center in Japan to drive product engineering for its digital cameras, under the assumption that access to highly skilled electronics and optics engineers was worth the premium wages.

Traditional manufacturing industries are close behind. Blue-chip names including Alcoa, Gen-eral Electric, Schneider Electric, General Motors, Toyota and Siemens appear on new development centers in China, India, Thailand, Mexico, Russia and other countries that have an abundance of en-gineering talent.

Some Regions Are Adopting GPD More Quickly American companies are rapidly embracing GPD. U.S. businesses have a strong culture of global col-laboration and thrive on an entrepreneurial spirit that focuses resources on high-value opportunities. Their counterparts elsewhere face more stringent regulations with regard to changing the size and composition of their workforce. German manufac-turers are resisting the loss of “quality German engineering” for as long as possible. Japanese busi-

COMPARING THE NEW PRACTICE OF GLOBAL PRODUCT DEVELOPMENT WITH THE CONVENTIONAL APPROACHConventional product development includes co-location of cross-functional teams. By contrast, best practice in PD today

Conventional Product Development! Largely co-located teams! Uses engineering located in existing engi-

neering centers! Uses a combination of digital PD tools and

conventional paper-based processes for en-gineering

Global Product Development! Globally distributed teams! Takes advantage of engineering in multiple geo-

graphic locations, including low-, medium- and high-cost regions

! Uses an entirely digital PD process to facilitate distributed, collaborative engineering features a highly distributed, networked development process facilitated by a fully digital PD system.

JULY 2009 MIT SLOAN MANAGEMENT REVIEW 5WWW.SLOANREVIEW.MIT.EDU

ness culture includes strong bonds of loyalty between employers and their workers, and thus a reluctance to move engineering work to lower-cost regions. However, substantial pressures for global growth and leaner operations dictate that the prac-tice of GPD is growing in every high-cost country.

China is perhaps the world’s fastest-growing manufacturing region, and many multinational corporations are locating offshore R&D facilities in China, often alongside their own production sites. At the same time, China-based businesses aim to become more global. While most still do all their engineering (and production) in China, some are taking advantage of highly experienced engineers in the West to develop new technologies and to help connect with international markets. Haier — one of China’s largest manufacturers of electronics and home appliances, with exports of more than $1 bil-lion — now has an R&D center in New York state as well as a manufacturing plant in South Carolina, both paying wages on an order of magnitude greater than in China. These operations have al-lowed Haier to connect with American consumers and retailers, understand American lifestyles and design products such as refrigerators, washers and wine chillers that Americans are now buying.

India is known as a source of experienced out-sourcing partners for engineering work. In recent years, approximately one million IT, software, business process and engineering services jobs have been created in India, ranging from call cen-ter and CAD drawing work to patent research and tooling design.8 Having experienced success with outsourcing such jobs, many Western businesses are now establishing their own offshore R&D fa-cilities in India.

Some of the most impressive companies in me-dium-cost regions have adopted a hybrid approach to GPD. In South Korea, for example, an effective GPD strategy takes advantage of low-cost engineer-ing (co-located with production in China) for certain functions; uses limited amounts of higher-cost engineering in the United States or Europe for access to the latest technologies and key markets; and still keeps much of the engineering process at home in Korea. Hyundai Motors is an excellent ex-ample. The company does most of its engineering work in Korea, but its GPD strategy calls for tap-

ping engineering skills worldwide. The automaker has established operations in the United States for engine calibration and testing (Michigan), vehicle styling (Southern California) and high-tempera-ture testing (California desert). Hyundai also operates research centers in Frankfurt (for diesel engine technology) and in Tokyo.

Other medium-cost regions are attractive due to their proximity to and strong ties with high-cost countries. For example, Eastern Europe is an in-creasingly appealing location for Western European companies to set up operations — either captive or third party. Alcatel, the French engineering con-glomerate, inaugurated its R&D center in Bucharest in mid-2005 to support its advanced rail control systems business in Romania, the Balkans and the Commonwealth of Independent States (the former USSR). Alcatel has had a presence in Romania since 1991, with nearly 1,000 people in production, in-stallation, maintenance and marketing.

The Essential Elements of GPDGlobalizing product development is an evolution

that typically takes place over a number of years. Many companies have at least some experience with global operations and distributed engineering pro-cesses before they decide to formally develop a GPD strategy. These experiences may derive from the ac-quisition of a business that has PD resources. They may also be the result of collaboration with manu-facturing engineers located near manufacturing facilities or with outsourced design partners used for specific projects.

Outsourcing Versus Offshoring As companies think more holistically about their product development operations and the distribution of various activi-ties, four fundamental modes of GPD emerge. (See “Defining Modes of GPD Based on Ownership and Location of Resources,” p. 7.)

Recent discussion about outsourcing and off-shoring — much of it politically charged — is relevant to GPD practice. “Outsourced” typically means the PD resources are owned by a third party, while “insourced” (or “captive”) means they belong to the manufacturer. Outsourced resources can be located on-site at the company, down the road at the third party’s offices or halfway around the

WWW.SLOANREVIEW.MIT.EDU6 MIT SLOAN MANAGEMENT REVIEW JULY 2009

world. The term “offshore” refers to the location of those resources — generally meaning lower-cost regions. Viewing the two concepts together allows us to envision four modes of global product devel-opment operation:

Centralized This mode is the traditional one, in which all product development resources are within the company and at onshore locations. Centralized operation can include different project teams in multiple countries, such as a U.S. team and a Ger-man team. All resources are owned by the company and located in the “headquarters” countries — that is, generally in high-cost regions.

Local Outsourcing This mode has been commonly used, often in conjunction with the centralized mode. Many large manufacturers use on-site con-tractors to support their product development activities. Local outsourcing is used for two pri-mary reasons: to gain access to specialized skills or to meet temporary requirements for capacity. For example, in the concept development phase, a number of ideation activities are the distinctive competencies of specialized outsource firms, such as IDEO, Design Continuum and Smart Design.

Captive Offshoring One of two relatively recent off-shore GPD modes, captive offshoring is useful when a company believes it should own a PD oper-ation in a region in which it has not done business before. This requires choosing a location, hiring a management team, securing a facility, establishing the operation as a legal entity, understanding the local regulatory and tax requirements, hiring and training staff, and putting in place the supporting finance, human resources and information tech-nology processes. Cummins and Textron have established centers in India; ABB and Delphi have set up centers in China; and Boeing and Airbus have established centers in Russia. These compa-nies are seeking to utilize some of the specialized skills, talent and experience in these respective

countries — India for complex, highly engineered products; China for its proximity to manufacturing and the large domestic market; and Russia for its long heritage in aerospace.

Companies that have established captive off-shore engineering centers report that a scale of approximately 200 people is needed to economi-cally justify the infrastructure and management investment. However even if financially justified, such an operation takes several years to build up to the point where the center is effectively incorpo-rated into the company’s culture and processes.

Global Outsourcing This type of arrangement — the other offshore GPD mode — gives many companies early experience with GPD without re-quiring the commitment to establish a captive center. Companies contract with a service provider to un-dertake basic engineering tasks, such as updating drawings, implementing engineering change orders or writing technical publications. This is typically done by staff augmentation on a time-and-materials basis, with an early focus on transferring knowledge and building a working relationship. By using out-sourced staff only at certain points in the process, the company keeps control of the PD process overall. However, this is an important step in moving toward a deeper integration of product development with the outsourced provider because it helps the pro-vider understand the company’s PD processes, methods and protocols, and it helps the company understand the technical capability, costs and timeli-ness of the outsourced provider.

Many companies eventually want outsourced providers to take on larger projects where owner-ship of the whole process shifts entirely to the outsourced provider. For example, the company may define the specifications for a subsystem. Then the outsourced provider does the high-level design, detailed design, prototyping, testing and redesign, ultimately delivering a completed design along with the necessary models, documentation and test re-sults. There may be design reviews throughout the

Companies that have established captive offshore engineering centers report that a scale of approximately 200 people is needed to economically justify the infrastructure and management.

WWW.SLOANREVIEW.MIT.EDU JULY 2009 MIT SLOAN MANAGEMENT REVIEW 7

process, but overall ownership lies with the out-sourced provider. Outsourced providers are increasingly focused on developing strategic rela-tionships with their customers by providing dedicated, secure centers; however, the employees and competencies developed within these centers generally remain with the providers.

Many organizations begin GPD with various trials of outsourcing and tend to move toward the captive offshore mode as a longer-term strategy. (See “Defining Modes of GPD Based on Ownership and Location of Resources.”) There are three rea-sons why an offshore center may become strategic, thus warranting a captive approach: (1) the work product contains intellectual property related to products or processes that provide valuable differ-entiation, (2) the skills and expertise that will be developed in the center relate to a core competence for the company or (3) the center will provide a basis for understanding local markets and design-ing products based on that understanding. For each reason, companies will want to retain control of the center and its human and capital resources, pro-cesses and systems.

The Staged Deployment of GPDWe have observed that best-practices leaders

typically deploy a GPD strategy in stages, allowing them to gain experience gradually by moving more and more development responsibility to new loca-tions. Companies often start by using an outsourced provider that augments an existing process in sim-ple ways and then takes on more substantial responsibilities. As a new location contributes more value, companies may open their own captive cen-ters to develop capabilities and retain competencies within their own PD operations. Recognizing, of course, that there are many hybrid variants, here are three basic scenarios of staged GPD deploy-ment . (See “Staged Approaches to GPD Deployment,” p. 8.)

Process Outsourcing The first approach involves outsourcing of PD process steps, which allows com-panies to work with outsourced service partners that may be distributed globally. Stage 1A starts by out-sourcing simple tasks to an engineering services provider. These tasks are easy to document and sepa-

rable from other activities, and they pose relatively low risk to the critical PD activities. Such tasks may include creation of detailed drawings from a CAD model or translation of technical publications into different languages. In Stage 2A, as the organization gains more experience with distributed PD, its man-agers outsource more complex and integrated tasks, such as CAE analysis or tooling design.

Component Outsourcing The second approach — also in the outsourcing category — is to decompose the product into components and modules. This path is generally followed when a company is out-sourcing development work to component suppliers. In Stage 1B, development of some simple components is assigned to chosen suppliers (which likely would also produce those components). In Stage 2B, the design of integrated components is outsourced. In Stage 3B, suppliers develop com-plete product modules, such as the exhaust system for an internal combustion engine or the circuit board for a mobile phone.

Captive Design Center The third approach sees de-velopment of a captive global design center — a strategic decision to invest in global engineering capability. Many organizations take this route after learning to utilize globally distributed resources for large parts of their PD processes. Because the tran-sition from outsourcing to a new captive center is difficult, the center is likely to begin working at Stage 1C or 2C, in parallel with outsourcing, to de-

DEFINING MODES OF GPD BASED ON OWNERSHIP AND LOCATION OF RESOURCESFour fundamental modes of GPD emerge as companies think more holistically about their product development operations and the distribution of various PD activities.

Offshore

Onshore

OutsourceInsourceOwnership

of Resources

CaptiveOffshore

Centralized(Local)

LocalOutsource

Location ofResources

GlobalOutsource

WWW.SLOANREVIEW.MIT.EDU8 MIT SLOAN MANAGEMENT REVIEW JULY 2009

velop these important capabilities first. In Stage 3C, the center executes complete subsystems, such as the design of a control system for an electrome-chanical product. In Stage 4, the center may take on complete design responsibility for derivative prod-ucts and/or ongoing engineering support for existing products. Typically, companies try to move to Stage 4 as quickly as possible to generate the stra-tegic benefits. Finally, in Stage 5, the center is fully able to develop new global products, product plat-forms and next-generation innovations.

A mature GPD structure combines all three ap-proaches and uses a balance of captive and outsourced resources, distributed globally. At this point the company may have one or more captive offshore R&D centers and will augment these facili-ties with selected outsourced engineering services. The service providers may work within the captive centers as part of a team or in their own facilities. Some companies will develop global technical cen-ters of excellence that nurture specific core competencies, develop new technologies and create product innovations. The mature GPD structure may evolve over time as new competencies and geographic regions become important, and older products and technologies may move to a mainte-nance mode, better supported by third parties.

Research on the dynamics of process improve-

ment shows that PD productivity almost always falls off when dramatic change is imposed. Repen-ning explains such “worse-before-better” system dynamics in the context of PD process improve-ment efforts.9 Offering strong evidence in the GPD context is a 2004 Columbia University Center on Globalization and Sustainable Development study10 of companies with experience in global outsourcing. The study found that companies new to outsourcing (less than a year) are unsatisfied with the experience, while those with more experi-ence are satisf ied with their outsourcing arrangements. The GPD learning curve appears to require at least a year of experience with each suc-cessive stage of process change before the global process runs smoothly.

Key Success Factors for GPD Deployment

Several key strategic decisions should be made before deploying GPD. These include defining a product strategy that delineates new technologies and target markets, identifying core competencies and intellectual property strategies and choosing locations for design centers and the GPD mode for each (for example, captive offshore, global out-sourced). The transition to a GPD operation brings its fair share of risks that must be managed. GPD

STAGED APPROACHES TO GPD DEPLOYMENTBest-practices leaders typically deploy a GPD strategy in stages, allowing them to gain experience gradually by moving more and more development responsibility to new locations.

Integratedtasks

Process Outsourcing

Simpletasks

Stage

1A

Stage

2A

Integratedcomponents

Component Outsourcing

Completemodules

Simplecomponents

Stage

2B

Stage

1B

Stage

3B

Integrated tasksand/or components

Captive Design Center

Complete modulesor subsystems

Derivativeproducts

New globalproducts

Simple tasksand/or components

Stage

2C

Stage

1C

Stage

3C

Stage

4

Stage

5

WWW.SLOANREVIEW.MIT.EDU JULY 2009 MIT SLOAN MANAGEMENT REVIEW 9

may impair PD productivity, put intellectual prop-erty and core competencies at risk or cause organizational disruptions as roles and processes change. Below are 10 key success factors that can help companies overcome some of the challenges presented when deploying GPD.

1. Management Priority GPD may require signifi-cant changes to the organization, processes and culture. Therefore, the executive team must have strong, visible commitment to the success of GPD. This commitment includes investment of the re-sources necessary to endure the worse-before-better performance of the PD organization. Some compa-nies appoint a corporate GPD executive to drive the GPD strategy across the various businesses, set up global design centers (captive or outsourced), facil-itate the distribution of work to the global centers, and ensure that tools, training, infrastructure and processes are in place.

2. Process Modularity To enable PD activities to be carried out in different locations, there must be a methodology to segregate the work packages for global distribution. For example, where a remote cen-ter will be handling tasks in a process that continues to be owned by the “central” PD location, a modular process is needed. The process must be broken down into clear steps, the steps distributed to different loca-tions and the process reconfigured to allow for the necessary handoffs, reviews and approvals.

3. Product Modularity Modular product architec-ture is very helpful for GPD in which development of complete subsystems or components is to be carried out by teams in different locations. Clearly defined interfaces between modules facilitate their separate development and eventual integration into the prod-uct. Without such modularity, more intense collaboration across design interfaces is necessary.

4. Core Competence Development of core compe-tencies is critical to a company’s sustainable advantage, so it is vital to control these competen-cies. “Don’t outsource the core business” is a fundamental principle of GPD. Accompanying any outsourcing decision is the danger that the com-pany will lose these elements of PD expertise as the

provider gains the knowledge. The company may become captive to the supplier for the work, and the supplier may grow to become a competitor. Conversely, a company can gain access to needed competencies by tapping the expertise of out-sourced providers.

5. Intellectual Property As critical product data, designs and technologies are shared more widely outside the company, protection of IP becomes more difficult. Defining products and processes in a modular structure not only can help with the dis-tribution of activities, but also can help protect IP.

6. Data Quality The availability, accessibility and au-ditability of data become key challenges when many locations contribute to the PD process, often using different tools and databases. Teams may be working on different aspects of the product with similar “source data.” As these data change during the pro-cess, all users of the data must be aware of the changes and the implications for their work. One system or database must be used as the “source of truth.”

THE OUTSOURCING TRAP EPPINGER: This is what my colleague Charles H. Fine at the MIT Sloan School

of Management calls “the outsourcing trap.” Essentially, the outsourc-ing trap is a dynamic that starts with outsourcing for good reasons. Your company may need some additional capacity, so you find a partner either locally or globally to provide it. Alternatively, you

may need some special capability in your development process and so outsource to a supplier that has that capability.

The problem can be that, over time, your business becomes more and more dependent on that supplier. Eventually, you may find you cannot do without that supplier. And, if the activity the supplier is doing reflects a critical technical capability that may in fact be a core capability of your business, all of a sudden you realize that you have outsourced the core. Usually, the process has been a slow drift over time, not guided by the right strategy. I have seen many compa-nies discover they were doing GPD only because GPD, in effect, snuck up on them – as a supplier or suppliers gradually took on more and more responsibili-ties, including aspects of product design. I have talked with people at many companies who have said to me, in effect, “We didn’t realize we were doing GPD, but I guess we are. But it looks like we’re not doing it very well.”

To avoid the outsourcing trap, you need to understand the core technical capa-bilities and technical needs of your business -- not just today, but in the future. Be sure to invest in building those internal capabilities and don’t outsource those; otherwise, you’ll lose those capabilities internally. When outsourcing, ask yourself: Which components and processes are really core to the business? Can we maintain the core? Choose very carefully which activities you outsource and to which suppliers. Before outsourcing, ask: In 15 years, where would we be if we don’t keep that capability in-house?

WWW.SLOANREVIEW.MIT.EDU10 MIT SLOAN MANAGEMENT REVIEW JULY 2009

7. Infrastructure Globally distributed PD teams often have tools, technologies, systems and pro-cesses that are specific to individual locations. GPD requires a more unified approach to infrastructure and systems to ensure that the appropriate informa-tion can be readily accessible regardless of location or bandwidth constraints. As design centers open in locations such as China, India and Eastern Europe, the network and power reliability may not be suffi-cient to support a fully operational center. Intermittent power and network outages may se-verely impact the center’s productivity.

8. Governance and Project Management Manag-ing a globally distributed PD operation is a complex activity, and it becomes more complex with every outsourced arrangement in a new time zone. So it is critical to have a strong capability to coordinate and monitor the entire program in terms of milestones, technical work quality and cost. Detailed project planning determines which decisions are made at what levels and locations, and how to coordinate across the operation to ensure alignment and proper execution.

9. Collaborative Culture The transition to GPD must incorporate new ways to collaborate among teams and individu-als across time zones, languages, cultures and companies. A key enabler is a consistent set of processes and stan-dards. As PD processes span locations, the practices and methods must be con-sistent and the standards for each work element explicit in order to reduce the variability of PD task outputs. Many companies have had success in creating a collaborative GPD culture by trans-ferring a manager from a central location to a new, remote design center for up to two years to educate the local team on the PD processes and to act as a liaison with the home office. The re-mote center may also send engineers to the home office for similar reasons. The trust necessary for effective GPD can only be developed over time, but exem-plar companies do not skimp on bidirectional travel between centers at

multiple levels in the organization. They view travel not just as a startup activity, but as an ongoing ne-cessity to build and sustain relationships.

10. Organization Change Management Many com-panies that have set up a new GPD operation find that some of the most challenging changes are those involving individuals’ roles, behaviors and the new skills required of them. Careful planning, training and education should go toward the individuals who play critical roles in making the new GPD model operational.

The GPD Mandate and Its RisksGPD is rapidly becoming the next-generation

practice of product development. Companies are adopting the approach quickly, spurred by interna-tional competition, new market opportunities, digital and networked PD process connectivity and the availability of low-cost skilled engineering workers worldwide. Although GPD is just now starting to win mainstream attention, companies are rapidly adopting new PD strategies, testing var-ious modes and moving to more global operations.

THE CHALLENGES OF SYSTEM ARCHITECTURE EPPINGER: The third concern that I would bring up relates to system engineering.

Through my research, I have developed a strong appreciation for the tremen-dous difficulty of integrating complex systems. When a complex system is developed, we define its system architecture, including the decomposition of the system into different elements – subsystems, modules, and compo-

nents -- and performance metrics for each one. After each of these components is developed, then comes the key challenge – to integrate them

all together. System integration is tremendously difficult, particularly when the work has been distributed globally and across companies. Such system integra-tion is another aspect of coordination cost in GPD.

The solution is: Have good, careful planning of the system architecture. Some of the companies we studied in our recent research do this quite well, by spend-ing a lot of time on the system architecture up-front. They also reserve a lot of time for the system integration on the back end. The actual component and subsystem development may be relatively fast.

In a 2007 working paper called “A System Architecture Approach to Global Prod-uct Development,” Anshuman Tripathy and I wrote cases studies about five companies’ GPD processes. One of the case studies that we researched was Intel Corporation’s microprocessor development for the multicore microprocessors in laptop computers. In the product development process for those microproces-sors, Intel spends about half of the time on the system architecture phase -- before even developing all the details of the components. And they do this architecture work at a single site, up-front. Then, the component development work may hap-pen across seven different Intel sites. Finally, after the components are developed, the system integration work comes back to the initial site.

WWW.SLOANREVIEW.MIT.EDU JULY 2009 MIT SLOAN MANAGEMENT REVIEW 11

This is a major structural change that will continue to develop over the next decade.

The political concerns surrounding globaliza-tion of PD are significant. Many engineering jobs and related management and support roles are being lost in the United States, Europe and Japan as these jobs move to lower-cost regions. Contentious questions abound. How will we keep engineers and managers employed in these high-cost countries? What is the future growth outlook? How can these nations fight back? Should they fight back? The glo-balization of business often runs head-on into national economic agendas.

Alan Blinder, Princeton economics professor and former vice chairman of the Board of Gover-nors of the Federal Reserve argues in a recent article11 that we have seen only the beginning of the trend of American service-industry jobs migrating to lower-cost regions. When this takes off, he argues, it may amount to a new industrial revolution. The accompanying economic and social adjustments will be substantial, and managers considering launching or expanding GPD initiatives must be aware of them and their impact on GPD decisions.

ACKNOWLEDGMENTS

We would like to thank the many colleagues, MIT research sponsors and PTC customers who have generously shared their GPD experiences and knowledge with us. We could not have de-veloped the observations in this article without their support and partnership. We are also grateful to several colleagues and anony-mous reviewers who provided insightful comments and helpful suggestions.

REFERENCES

1. M. von Zedtwitz and O. Gassmann, “Market Versus Technology Drive in R&D Internationalization: Four Different Patterns of Managing Research and Development,” Research Policy 31, no. 4 (2002): 569-588. O. Gassmann and M. von Zedtwitz, “Trends and Determinants of Managing Virtual R&D Teams,” R&D Management 33, no. 3 (2003): 243-262.

2. J. Santos, Y. Doz and P. Williamson, “Is Your Innovation Process Global?” MIT Sloan Management Review 45, no. 4 (summer 2004): 31-37.

3. “Mastering Complexity in Global Manufacturing: Powering Profits and Growth Through Value Chain Synchronization,” Deloitte Research, 2003.

4. T.L. Friedman, “The World Is Flat: A Brief History of the Twenty-first Cen-tury” (New York: Farrar, Straus and Giroux, 2005).

5. These specific country categorizations are based on salary survey re-

sults obtained from Mercer Human Resource Consulting for a representative type of engineer. (Systems Engineer – Intermediate, de-fined as “under general supervision, provides technical support to sales force during sales negotiation. Configures hardware, software, and de-sign application requirements of products offered to customers to meet their requirements. Resolves complex technical issues with guidance from senior engineers. Frequently reports to a Systems Engineering Manager. Typically requires a Bachelor’s Degree and two to four years of experience.”)

6. “Global Product Development — Moving From Strategy to Execution,” PTC and BusinessWeek Research Services, 2006.

7. “Services Export Study,” Federation of Indian Chambers of Commerce and Industry, July 7, 2005.

8. “Strategic Review 2006, The IT Industry in India,” NASSCOM, 2006.

9. N. Repenning, “Understanding Fire Fighting in New Product Develop-ment,” Journal of Product Innovation Management 18, no. 5 (2001): 285-300.

10. N. Bajpai, J. Sachs, R. Arora and H. Khurana, “Global Services Sourcing: Issues of Cost and Quality,” CGSD working paper no. 16, The Earth Institute at Columbia University, Center on Globalization and Sustainable Develop-ment, New York, June 2004.

11. A.S. Blinder, “Offshoring: The Next Industrial Revolution?” Foreign Af-fairs 85, no. 2 (March-April 2006): 113-128.

NEW TRENDS GOING FORWARD EPPINGER: Two current trends in GPD come to mind. One is what we would

call “reverse offshoring.” Reverse offshoring involves businesses that are based in medium-cost or low-cost locations but are putting some jobs in the U.S. and other relatively expensive locations. In the 2006 article, we gave the example of Haier Group Co. Ltd., a company based in Qingdao, China that makes electronics and

home appliances; Haier has production in South Carolina as well as offices in New York and a design center in California. Another example is Hyun-dai Motor Company, which is based in Seoul, Korea but has some operations in the U.S. for vehicle styling, engine calibration and testing.

Now we are also seeing another kind of reverse offshoring, in which busi-nesses based in the West had moved some engineering and other white-collar jobs to countries such as India and China. However, as wages in India and China for certain kinds of engineering specialties are increasing, they are now realizing that there are many underemployed engineers here in the West whose capabilities aren’t being fully utilized. As a result, we are seeing some repatriation of jobs and some reversing of that initial offshoring trend.

The second trend represents, perhaps, a maturing of global product devel-opment. We wrote in the original article about what we called Stage 5 of global product development centers -- how eventually global captive design centers would grow increasingly sophisticated in their work and begin developing new global products. Today, that is happening. Mature captive offshore global development centers have created innovations, sometimes for those emerg-ing markets, and some of those innovations are coming to Western markets in the form of lower-priced products. For example, GE Healthcare, a division of General Electric Co., in 2009 obtained clearance from the U.S. Food and Drug Administration to sell a portable electrocardiogram (ECG) device in the U.S. that had originally been developed in China, for the Chinese market.

Title: Managing Complex Product Development Projects Apply: Apply online Dates: July 15-16, 2009 Duration: A Two-Day Program for Senior Technical Managers

Program also o!ered on these dates: November 11-12, 2009 March 22-23, 2010 July 14-15, 2010 November 10-11, 2010

Location: Cambridge, Massachusetts Cost: $2,600 (covers tuition, program materials, and breakfasts and lunches) Brochure: Download the brochure Podcast: Prof. Steve Eppinger discusses innovations and trends in product development Schedule: Sample Program Agenda Executive Technology, Operations, and Value Chain Management Certi!cate Track:

MIT Sloan Executive Education

DescriptionManaging complex product development projects is a massive integration e!ort at many levels. Product and production plans must be integrated into components, components into subsystems, subsystems into systems and systems into quality products.

"e traditional project management does not provide the kind of detail required today to both accelerate product development and improve product quality in the 21st century. Managing Complex Product Development Projects presents a revolutionary design structure matrix (DSM) that MIT researchers use to determine which tasks within each phase of a complex project should or should not be performed concurrently. "e DSM method is already applied in a number of corporations.

MIT researchers developed the DSM modeling approach to learn how to solve problems facing large-scale projects. A#er $eld-testing DSM in dozens of organizations and industries around the world, they found that it successfully streamlined the development of a wide array of projects including:

complex automotive components systems and subsystems aerospace con$guration design concept development and program roll-out electronics and semi-conductor development equipment and machine tool development plant engineering construction projects

"rough lectures, exercises, interactive discussions, and teamwork, participants in the program learn how to use DSM to map complex design procedures into simple arrays.

Most important, they learn how to solve four key problems that confound complex product development project management: iteration, overlapping tasks, architecture, decomposition and integration.

In Managing Complex Product Development Projects, participants learn to:

better document existing procedures reduce complexity share data with con$dence facilitate project %ow expose constraints and con%icts design iteration strategically Project Management Institute (PMI)Participants in this program are eligible for 11 PDU credits from PMI. Provider ID: 2420. Course number: PD001.

"e ParticipantsManaging Complex Product Development Projects is designed for senior managers involved in complex product development and project management as well as those responsible for speeding up the process of improving design procedures and designing and developing better products. "ese include:

vice presidents of engineering, manufacturing, and technology directors of project management product and business development engineering and R&D program managers chief project engineers product design and process development engineers technology strategists project leaders

FacultySteven D. Eppinger, General Motors LFM Professor of Management Science and Engineering Systems, is co-director of MIT's Center for Innovation in Product Development and co-director of the Leaders for Manufacturing and System Design and Management masters programs. His research creates new approaches to improve complex product development processes, work that has been applied primarily in the automotive, electronics, aerospace, and equipment industries. Eppinger lectures regularly for international corporations and in executive education programs and has con- sulted for or conducted research with more than $#y organizations. He has worked as a manufacturing engineer, product designer, and consultant in both prototype and production operations.

Executive education program led by Professor Steven D. Eppinger

PDFs ! Reprints ! Permission to Copy ! Back IssuesArticles published in MIT Sloan Management Review arecopyrighted by the Massachusetts Institute of Technology unless otherwise specified at the end of an article.

Electronic copies of MIT Sloan Management Review articles as well as traditional reprints and back issues canbe purchased on our Web site: sloanreview.mit.edu or you may order through our Business Service Center (9 a.m.-5 p.m. ET) at the phone numbers listed below.

To reproduce or transmit one or more MIT Sloan Management Review articles by electronic or mechanical means (including photocopying or archiving in any information storage or retrieval system) requires written permission. To request permission, use our Web site (sloanreview.mit.edu), call or e-mail:

Toll-free in U.S. and Canada: 877-727-7170International: 617-253-7170Fax: 617-258-9739e-mail: smrpermissions@mit.edu

Posting of full-text SMR articles on publicly accessibleInternet sites is prohibited. To obtain permission to postarticles on secure and/or password-protected intranet sites, e-mail your request to smrpermissions@mit.edu.

Hyperlinking to SMR content: SMR posts abstracts ofarticles and selected free content at www.sloanreview.mit.edu. Hyperlinking to article abstracts or free content does notrequire written permission.

MIT Sloan Management Review77 Massachusetts Ave., E60-100Cambridge, MA 02139-4307e-mail: smrorders@mit.edu