Working Paper 94-8

A report prepared at the
International Executive Forum
Managing International Research and Development
of the Carnegie Bosch Institute

Pittsbrugh, November 1993

Managing International Research and Development

Ýby

Klaus Bolenz, Robert Bosch GmbH, Germany
Siegried Dais, Robert Bosch GmbH, Germany
Eugene Greenstein, Ford Motor Company, U.S.A.
Gezinus Hidding, Anderson Consulting, U.S.A.
Hyo-Bin Im, Daewoo Corporation, Korea
Howard Kell, Ford Motor Company, U.S.A.
Bernd Kiel, Carnegie Bosch Institute, U.S.A.
Bruce McKern, Carnegie Bosch Institute, U.S.A.
Francesco Natalini, Merloni Elettrodomestici spa, Italy
Werner Pollmann, Daimler-Benz AG, Germany
Eleanor Westney, Massachussets Institute orf Technology, U.S.A.
Wolfgang Ziebart, BMW AG, Germany

The Institute's one-week International Exchange Forums address issues in the strategic and operational management of global corporations. A distinctive feature of the Forums is the preparation of an "Issue Report" based on the experience of participating executives and the insights of academic researchers.

Chapter 1

The purpose of this chapter is to assess three alternative models for the internationalization of R&D. One of these models, which we view as particularly strong, builds on the premise that internationalization does not require geographic decentralization. This chapter provides guidelines by which a company might decide which model would be the most appropriate for its own context.

As markets and competition have become increasingly international, firms have faced the question of developing an optimal pattern for organizing worldwide R&D activities, and the best policies for ensuring the effectiveness of their R&D network.

Firms' responses to the challenge of internationalization have varied considerably, and also varied within firms according to business units or product groups. Some firms have adopted a "global" strategy, in which a globally homogeneous product is managed by a centralized organization with concentrated manufacturing centers. At the other extreme are firms which have employed a "multinational" or "multi-domestic" strategy, in which responsibility for most functions is decentralized and the asset base is dispersed. In most cases, some functions, such as treasury and foreign exchange risk management, are centralized, while others, such as sales and service, are usually decentralized. Recently, some observers have advocated a "transnational" approach to international strategy, in which the firm is urged to reconcile simultaneously the pressures for global efficiency with the need to be nationally responsive. The R&D organization plays a critical role in each of these strategic alternatives. This function, historically strongly centralized in the home country, has faced growing pressures to internationalize. Yet no clearly optimum solution for the internationalization of R&D has emerged. Indeed, some of the most favored strategies of the 1980's have proven to be far more expensive to implement and far less effective in operation than expected.

The purpose of this chapter is to assess alternative models for the internationalization of R&D and to provide guidelines by which a company might decide which model would be the most appropriate for its own context. The most common models of R&D internationalization involve establishing R&D centers outside the home country, and two of the models we present follow this pattern. However, it is possible to internationalize without geographic decentralization. One of the models proposed is "Centralized internationalization": that is, the development of systems and procedures to raise the sensitivity of the centralized R&D function to markets, competitors, and sources of technology outside the home base.

Approach

Strategies for internationalizing the R&D function should be evaluated by the extent to which they increase the value that R&D contributes to the firm. This chapter, therefore, begins by presenting five criteria by which the success of the R&D function can be evaluated by the company. It then presents four models: the first is the base case of centralized R&D, and the other three are alternative strategies for internationalization. The strengths and weaknesses of each model are assessed in terms of how each contributes to the key success factors in R&D. The discussion of each model also presents guidelines for its successful implementation. The Appendix describes in more detail the methodology used to assess each model, to provide a way of framing and expediting analysis of the range of internationalization options.

Key success criteria for the R&D function

The R&D function contributes value to the firm to the extent that it meets each of the following:

Its results meet the customer's needs at the time the product is introduced (not at the time the R&D project begins).

Project are completed at minimum cost and withing minimum time in R&D, relative to competitors.

Solutions are percieved as "state of the art" by customers.

Solutions are "elegant and powerful": that is, they have inherent qualities that result in minimum manufacturing costs and maximum reliability in the field (a high reliability/cost ratio).

The R&D organization expands its competencies over time rather than simply exploiting them: that is, it is stretching its capabilities to achieve the four goals outlined above.

The weights or values of these five criteria are expected to differ across companies and even across different businesses in a single company. (The weighting criteria used here is described in the Appendix).

The first two criteria (meeting customer needs and minimum cost and time) are most important for the short-term success of the firm; the last is essential for its long-term competitiveness.

Internationalization strategies for R&D

R&D has been the last function of the corporation to go abroad, in large part because the benefits of centralizing R&D have long been seen to be high. Therefore, geographic centralization is presented as the base case with which to compare three internationalization strategies. Each strategy is evaluated in terms of each of the criteria for a successful R&D function; the comments on the strengths and weaknesses of each method are based on this evaluation. The second chapter then deals with the question of what kinds of R&D should be centralized or decentralized.

A) The base case: Centralization of R&D

In this case, the R&D function is concentrated in one country, usually the home country, and its engineers develop products for all the markets served by the company. Several of the worlds auto companies, e.g., Volkswagen, Volvo and Toyota, followed this model until very recently.

Strengths:

This model results in minimum R&D cost and time;

Proximity allows engineers to share information easily;

Common management systems and shared culture make cooperation among engineers easier than in the model where communication takes place across national boundaries;

Proximity and shared organizational systems produce flexibility; direction can be changed with minimum coordination costs.

Weaknesses:

This model is the least satisfactory for meeting the needs of customers outside the home country;

Centralization increases the danger of insensitivity to "distant" signals from foreign markets;

The organization tends to be resistant to change because basic paradigms go unchallenged.

Key Success factors:

For a company to make this strategy work in an internationally competitive industry, it must be able to establish the world-wide standards in the industry -- "one world, one solution". Perhaps the closest example of this approach in the last few years is Microsoft.

This strategy also depends on low differentiation across national markets in the product niche or segment served by the company, i.e., a "global" product solution.

B) Centralized internationalization

Here, the RD function remains concetrated in one country, bu the company makes significant and ongoing investments in raisinng the international awareness and sensitivity of its engineers, especially their knowledge of "state of the art" technolgy elsewhere and their responsiveness to customer needs in markets outside the home country. This can be achieved in a variety of ways, such as:

Sending engineers abroad to work with suppliers and major customers,

Bringing key people from the local market to interact frequently with engineers;

Hiring engineers from other countries, especially those with experience in other companies, into the central R&D organization;

Hiring home country engineers with foreign language skills and experience abroad.

An example of a company that successfully implemented this strategy is Nissan in Europe. In order to develop a new car for the European market, Nissan selected a core project team and established its base in the European headquarters (primarily a planning, marketing, and sales organization). The teams's efforts were supplimented by those of hundreds of Nissan envineers who were sent to Europe on short visits to drive competing cars on European roads and interact with European car experts and potential customers. Nissan made very good use of the cadre of its engineers who had previously worked on a joint project with Volkswagen and who, therefore, already had some exposure to European car technology and design. The result was a car designed completely by Japanese engineers that was the first major Nissan success in Europe. Although Nissan made this approach to internationalization work very well from a marketing and technical point of view, it is now moving to set up R&D centers in the United States and in Europe, in large part to improve its image as a "local insider" and to improve its local content ratios in anticipation of further restrictions on the sales of autos produced by the Japanese "transplants" in North America and Europe.

Strengths:

This model best satisfies the short-term criteria of meeting customer needs outside the home coutnry and minimizing R&D cost and time.

This model is also best for long-term goals of leadership in the business. It is best at producing solutions that are percieved as state-of-the-art by all customers, because it draws globally on technical inputs and is built on understanding a wide range of customers' perceptions. As a result, it is also best at challenging and stretching engineers' capabilities.

It sustains the centralization advantages of the first model but addresses the major weaknesses, at a lower cost that the models that involve investing in offshore R&D center.

Weaknesses:

This approach is difficult to implement effectively, because it requires the greatest individual-level change in behavior, modes of working and attitudes.

There is a danger that the company will under-invest in the systematic internationalization of its engineers. (It may be embraced as the "cheap and fast" way to internationalize.)

In industries where local content regulations are significant, this model may enable competitors to erect barriers to the company's products. Furthermore, in industries where being seen as a local player is an important competitive advantage, the concentration of activities in a single foreign center may have "image" costs.

Key success factors:

Major, ongoing investments in internationalizing the R&D organization are needed, but these are still considerably lower than the investments in building offshore R&D centers.

This model requires systems to expose engineers repeatedly and effectively to stimuli from foreign markets and centers of technology.

A critical mass of engineers with international sensitivity and experience and open-minded project managers are needed to act as "change agents".

A major cultural evolution must take place within the organization.

Top-level R&D managers must make a visible commitment to this strategy. They must "walk the talk"; this is not a strategy that can be implemented just by exhortation.

C) Centrally directed decentralization

In this model, which is perhaps the one that most people envision when they think about the internationalization of R&D, a number of R&D centers have been set up outside the home country. THey are coordinated and directed by the central R&D organization. One example is the Eastman Kodak Corporation, which has R&D centers in England, France and (until recently) Japan that are dependent for some of their funding and much of their strategic direction on the corporate and business unit R&D organizations in Rochester, New York. Another example is NEC, which has several laboratories in the United States that are funded by and closely linked to its R&D organization in Japan.

Strengths:

This model is strong in meeting the needs of foreign customers in a timely fashion.

It performs well on producting solutions that the range of customers percieve as state-of-the-art, because the diversity of inputs increases the likelihood that solutions will have considered a broad range of inputs and expectations.

Relatively little change in the existing R&D organization is required.

It can draw on particular strengths in the dispersed units.

Weaknesses:

This model performs poorly on the criterion of minimum cost and time, because of high coordination costs and delays caused by geographic dispersion.

It can result in duplication of effort and poor coordination if the offshore centers develop (or maintain) high levels of autonomy.

Key success factors:

Acceptance by local centers of the centre's supervisory role is essential yet not easily achieved, especially when the local centers are the result of acquisitions or have evolved into autonomous locally-oriented organization.

Size is a critical variable: each offshore R&D facility must be large enough for critical mass without having the scale for independent product development.

Some overlap in the system is required so that the home country organization retains enough expertise in all technical areas to exercise supervision credibly and to coordinate effectively.

Face-to-face communication through annual meetins of all center managers plus regular contact between formal meetings is essential to build the trust required for coordination.

The "Not-Invented-Here" syndrome at the center is a constant danger. The home country's engineers must cultivate a "big ear".

D) Integrated Network

This model consists of a network of interdependent cetners with a flexible coordination system. While some centers are designated as leaders in particular technologies, their leadership is based on a combination of their own capabilities and the resources of their immediate environment, and subject to change and challenge. The home country R&D organization becomes one of several centers in the network rather than its core.

Companies that adopt this pattern usually do so because they already have in place a number of highly capable R&D centers outside the home country, either as a result of the evolution of strong country subsidiaries or of market-targeted mergers and acquisitions. Therefore, a company that has a dispersed set of autonomous locally-oriented R&D centers faces the task of either closing them, imposing stronger central coordination and control, or integrating them into an international network.

The "integrated network" model is much more appealing to the subsidiaries than either of the other two alternatives. However, a company that does not have a dispersed network already in place must seriously consider the costs and the long time horizon involved in creating one.

One prominent example of this model is Nestlé, which operates 16 research centers or "RECOs" with 21 facilities in 10 countries. Only seven of these were directly established by Nestlé the others were acquired in its strategy of expanding its markets through acquisitions. A corporate staff of fewer than 20 persons coordinates (but does not control) the network, and is currently establishing a pattern of "centers of excellence" so that each center specializes in a distinctive technology area. The corporate staff helps identify areas of synergy and prospects for cooperation across the centers, and is closely involved in the company's strategic management.

Strengths:

This model is best at meeting a wide range of local customer needs.

It provides a critical mass of "change agents" within the R&D network.

Weaknesses:

This model is worst at meeting R&D cost and time, because of the effort needed to achieve coordination.

DIstance and internal competition are likely in the absence of a clearly identified "conductor" (leader).

This model requires the greatest organization change: a move away from control structures to coordination structures, and the change in status of the home country R&D organization from the controlling center to one center among others.

Key success factors:

This modelrequires a cultural change toward a common "mindset" of shared values and compatible systems and procedures, along the lines of the "transnational" system.

A common language is essential

Efficient communications mechanisms are required, including frequent meetings and routinized information flows.

Also essential are effective conflict resolution mechanisms, whose legitimacy is accepted throughout the network.

In the absence of a permanently leading center, the competition for leadership needs to be controlled by the corporate headquarters.

Conclusion

There is no single strategy that is the optimum for all companies, or even most companies. The choice of a preferred model for any single company will depend on the weight that a company gives to the various success criteria in R&D, the number and kind of R&D centers the company already has offshore, and the capabilities for internationalization of its current R&D organization and management. None of the three internationalization strategies can be implemented without considerable investment of money and time.

However, for a company that is just beginning to consider internationalization strategies for R&D, we believe that the "centralized internationalizaiton" model, if properly implemented, offers the best prospects for both short-term and long-term company leadership. It combines the long-recognized strengths of a centralized R&D function with the sensitivity to international customers, offshore sources of technolgy, and foreign competitors that is essential for building sustainablecompetitive advantage in today's business environment.

Chapter 2

This chapter deals with the preferred strategic and organization policies for optimal performance of research and development in an international context. Of primary concern is optimizing the effectiveness of the worldwide R&D organization.

Of primary concern in this chapter is optimizing the effectiveness of the worldwide R&D organization. SInce it is not obvious that R&D structures should parallel the firm's choice of structure for manufacturing or marketing and sales, three important topics demand consideration:

Whether R&D functions should always be centralized;

If not, on what basis should some R&D activities be decentralized;

under either scenario, what policies are most likely to make the firm's R&D system function effectively.

Chapter 1 addressed the first two questions. However, what decision a firm makes regarding the mix of centralized and decentralized activities, the optimal operation of the system must be an overriding concern. Consequently, this chapter focuses mainly on the third topic -- the operational policies and practices needed to optimize the functioniong of R&D.

Assumed business context and R&D structure

For the purposes of keeping the discussion practical, the assumed business contexts is a multinational firm with operations in countries with open markets, producing state-of-the art products. The figure below presents a schematic overview of such a firm. It assumes a strong central R&D organization [Design Center I], production sites in multiple countries, and decentralized development activities [DC II and DC III] in several countries. (For the purposes of discussion, the product is assumed to be an automobile alternator). It is important to notice that Design Center III is located in a country where there is a major customer, but that the plant in country 1 neither has a Design Center nearby nor a major customer. This situation seems to present a realistic (though not necessarily optimal) situation. The activies listed for the plant in country 1 and Design Centers I and II are given as an illustrative example. We believe that the operational policies described will apply to any of the structures a firm may use, although their relative importance will vary.

Centralization versus decentralization: Basic research versus design and development

For a business context such as depicted in the figure on page 11, R&D should not be centralized for all core competencies. (This assumes the current and future core competencies are known, given the firm's strategy.) Thus, the question becomes how much to decentralize.

Firstly, basic research should be distinguished from design and development. Basic research can be defined as "not directly tied to a specific product", "Needing a long time (more than a year or longer)", "not tied to production", or "not tied to customers". Development can be defined as "the application of proven basic technologies to products and process", "implicitly tied to a specific product", "able to be done in a short time (few months)", "tied to production", or "tied to customers".

Basic research for each core technology should be centralized, although the central location does not have to be world headquarters or the home office location (see first chapter). The main reasons for centralizing basic research are cost (through economies of scale), synergies and critical mass with specialists in one place, and control.

Design and development, on the other hand, should be decentralized preferably, because of the value of competition among development groups, the need to tap into a variety of ideas, and the need to be close to markets and production sites. These advantages of decentralization have to be balanced, however, against the dis-economies of scale in investment in development technology (e.g., testing) and utilization of such technology. Another major challenge in decentralization, the facilitation of communications and organizational learning, is addressed later in this chapter.

For clarity, the subsequent discussion is based on a simplified model of a distributed R&D network, focussing on the objective of making the R&D function effective. A number of issues are grouped into six categories: People, culture, core competencies, organization, design process, and customer. In each of these six categories, the two most important issues are identified, together with some recommendations.

Performance measures for R&D operations

Given the objective of optimizing worldwide R&D performance, it is essential to state criteria for measuring that performance. These criteria are consistent with those proposed in Chapter 1 for deciding the R&D structure, but are expressed here in more operational terms.

The following criteria are useful to measure R&D operational performance:

Time:

Design cycle time, as influenced, for example, by the reuse of base designs or common parts, or the time to release;

Response time,e.g. to customer questions;

Failure rate, as indicated by, for example, the number of customer complaints, the number of changes after release of the product, or the number of field complains;

Quality:

Product quality, as judged by the fit to specifications, customer satisfaction;

Process quality, as in first pass success (first time right);

Total cost, as measured, for example, by the cost of the project relative to target;

Level of innovation.

Furthermore, R&D is more likely to be optimal when it:

Interacts directly with the customer;

Maintains and extends core competencies.

Optimizing the performance of distributed R&D

The following issues and recommendations, addressed in six categories, are applicable to managing the coordination among the central design center and remote design centers as well as the coordination between the design center(s) and the plant(s).

People

The most critical issue in this category is to define appropriate measures of performance for R&D groups (see above). Other major issues are ensuring that the decentralized groups stick to the firm's strategy, and rewarding (through compensation, recognition, etc.) the entire project team, recognizing that R&D is a team effort.

Culture

The most important issues relative to culture appear to be personnel policies on job rotation, development of personnel, promotion of personnel and internal communications. Another key issue is to guarantee a creative environment.

Job rotation between international locations should occur on a regular basis, with assignments lasting from 3 to 5 years. Every 5 years or so, the rotation should be from the central design center to a decentralized center or vice versa. Remote assignments should be staffed with respected people with demonstrated expertise.

In addition, a firm should:

Send central staff overseas to learn local culture and language, with language training being available for all relevant employees;

Give part of the initial training for overseas R&D personnel in the central location;

Encourage and enable permanent careers for foreigners in the central location;

In each location, employ a mix of local and overseas nationals;

Put personnel returning from a remote assignment in a position that builds on their newly gained experience and expertise;

Address reentry problems as part of long-term career planning.

Across countries, norms of performance should be established, adjusted for salaries, exchange rates, education, working hours, and so on. Evaluation reports from the remote group should be factored into the merit rating or promotion of the individuals. Promotions should be favored for people with experience in more than one center. In case of competing projects, any promotion of staff should take place only after some time, to avoid the feeling of having failed on the "losing" project. A remote assignment should be viewed as a special assignment that commands prestige.

To foster international communication, each location should maintain ties to its remote personnel. Furthermore, the company should:

Organize seminars, not necessarily related to specific projects or targets, on a regular basis with participants from all locations;

Hold organization-wide, cross-cultural events;

Designate 5 to 10% of the budget for non-targeted activities;

Involve personnel from central and overseas locations in training;

Encourage and fund frequent trips to other locations to work together (as opposed to just exchanging information);

Rotate employees through several or all locations for approximately three months at each place;

Give outgoing expatriates training in communication skills in the new foreign culture and also train the returning expatriates.

To ensure that intercultural communication issues are not getting in the way and to develop trust, new teams should work together for some time before the start of a project. Project teams should meet periodically, possibly using a team work facilitator. They should review process issues and focus on what is working well, what is not working well, and how to improve communication within the team.

To guarantee a creative environment:

Allow R&D staff some "free" time;

Let them "play" with ideas;

Have some funds for innovative new ideas;

Invite "crazy" ideas and have "crazy-idea" meetings;

Reward R&D engineers and their managers for innovation;

Create "fellows" (a la IBM) and give them time and money to pursue their ideas;

Organize internal exhibits of new ideas or products not yet introduced;

Reward improvement recommendations and patents, even when they are not implemented;

Recognize the team as a team for its work.

Core competencies

The most important issue in this category is maintaining and extending core competencies and supporting them through technology and information. First, of course, the firm needs to make sure the "vision" or "strategic intent", expressed in terms of basic customer needs, is known to all personnel. ON that basis, the core competencies can be identified and a critical mass of expertise in critical technologies established and maintained.

To maintain and extend core competencies, the company should:

Maintain contact with the scientific community;

Keep the work environment attractive for key people;

Make sure that new R&D projects are reviewed as to whether they add to the core competence ("competence audit");

Involve experienced key people in all designs and conduct "exit design review";

Apply lessons learned -- successes and failures -- in next designs;

Give awards for excellent work;

Let every person become known for his or her expertise;

Encourage presentations and attendance at technical meetings and regular presentations to other groups.

In order to support maintenance and extension of core competencies by technology and information, it is essential to:

Provide easy access to documentaiton findings;

Create an internationally shared data base and common world-wide CAE/CASE tools accessible on-line by as many engineers as possible;

Create interfaces for such data bases that are very user-friendly, especially to accommodate language barriers;

Provide automatic translation for important expressions in all relevant languages;

Establish and maintain standard data bases for component parts and past designs;

Establish a common configuration management system to control changes in versions and releases of designs.

Organization

A key issue here is technology transfer, as technology transfer from central to remote units does not happen by itself, but must be managed. The second key issue concerns the development of the system as a whole, as opposed to parts of the system. Responsibility for the whole system should be established clearly.

To encourage technology transfer, a company should:

Define a schedule for worldwide conferences;

Assign development personnel to projects and centers for a sufficiently long time;

Use a matrix management structure;

Set up teams that include foreign functional managers;

Have common data bases accessible worldwide and assign responsibility to appropriate groups for their contents.

To define responsibility for the whole system that is to be designed;

Define one person as the leader for the whole system;

Develop people who are both good operational leaders and good organizational managers;

Clearly identify centres of excellence;

Recognize success with special benefits and awards;

Establish a budget for informal meetins;

Design process

Two key issues relative to the design process are defining tasks and responsibilities within the project team, and ensuring close cooperation between development, marketing, and manufacturing.

In order to define tasks and responsibilities within the project team, the firm should make one program manager responsible for the entire program and define clear responsibilities for each member of the team, including the support organization. In addition, it should:

Institute lifetime ownership of and responsibility for designs;

Check often for overlap and duplication of tasks and responsibilities across various teams and locations;

Create check lists.

To foster close cooperation between development, marketing, and manufacturing, the firm should establish joint project teams with people from development, marketing, manufacturing, etc. reporting to the program manager. Additionally, it is important to:

Get concurrence on projects from related departments as each phase is completed;

Base rewards on performance of the team as a whole;

Have exit reviews after each phase to evaluate the quality of the process;

Require marketing and manufacturing to share in the R&D costs by charging deveopment cost directly, recuperating development cost on a per-piece basis from slaes, or by instituting a mix of central payments for specific projects.

Other key issues to be considered include how manufacturing provides input to design and how to accomodate failures.

Customer

Critical issues here are to make sure to talk not only to the next department in the chain, but also to the ultimate customer, and to define clear responsibility for the interface with that customer. Therefore, a form should:

Establish a hot-line for customers that does not necessarily go through sales;

Perform customer satisfaction audits and surveys;

Have customers sign-off on the conceptual design and the final specification;

Make sure R&D and manufacturing see the product in use.

With respect to the responsibility for the interface with the customer, the program manager should be responsible for all aspects of the project from design through manufacturing launch. Sales and marketing should be responsible for resolving all business-related probles, e.g., price, quantities. In addition, it is important to:

Let engineers of the company and customers work directly together;

Assign an engineer to the "customer team" for day-to-day work;

Have every R&D staff member visit customers regularly (e.g., every 6-12 months);

Invite customers into the R&D centre(s);

Once a year, look strictly into how well R&D fits the ultimate customer's needs.

While these recommendations are certainly not comprehensive, they appear to be applicable to the effective management of a distributed R&D system of any configuration.

Appendix

A) Success criteria

In our discussion of the key success criteria in R&D, one criterion was identified that determines whether the R&D function is widely perceived to be successful. It is also important in shaping the personal and professional satisfaction of the R&D engineer:

Solutions are percieved as leading the field and as setting the standards of quality and performance that others follow.

A firm may be very successful with an R&D function that excels at rapid followership rather than technical leadership, but its R&D function will not be widely respected. (A long-standing example is the contrast between Sony, whose R&D function is widely perceived as a technical leader, and Matsushita, a rapid follower widely respected as a firm but seen as having an R&D funciton considerably inferior to Sony's.) In the evaluation of internationalization strategies throughout the first chapter, this factor is, in effect, folded into the third, "The solutions should be percieved by the customer as state-of-the-art". However, this factor does deserve independent consideration in some companies, and, in general, firms should consider the fit between the strategy for internationalizing R&D and this aspect of their functional strategies as well as its fit with their business strategies.

B) Assessing international strategies

THe approach used in Chapter 1 to assess different strategies for internationalizing the R&D function could, we believe, provide a helpful tool for companies in their efforts to analyze the issue. The following steps are involved:

Agrteeing on key success criteria for the R&D function and on alternative models of R&D organization (here, the base case of centralized R&D and three models of internationally distributed structures).

For each success factor and for each model, individually assigning a rating in terms of how well the features of that model contribute to good performance on that criterion.

Normalizing these ratings with a weighting factor for each success criterion to reflect its relative importance to the company.

Adding the ratings to provide a measure of overall value for each R&D strategy.