Page 136
7—
Organizational and Societal Infrastructure
Chapters 3 through 6 discuss a research agenda for information technology related to manufacturing. If this research agenda is fully pursued, the scope of what is achievable will be greatly expanded. However, an equally important challenge is understanding how manufacturing enterprises can actually make use of information technology; even today, much technology remains unused.
Put differently, the effective use of information technology (IT) in manufacturing depends on more than technology research. It also depends on success in (1) understanding the concerns of manufacturing decision makers; (2) motivating effective collaboration and technology transfer between industry and academia; (3) motivating individuals within manufacturing enterprises to implement information technology; (4) developing open standards and appropriate metrics of performance; and (5) ensuring through education and training that the skill base in manufacturing adapts to the new types of tools, techniques, and organizational structures made possible by information technology and to other "best practices" and "benchmarks" not necessarily associated with information technology. Also needed will be execution of global business strategies that protect high-value-added jobs for U.S. manufacturing facilities. In some instances, these essentially practical recommendations contain their own requirements for research, particularly social science research. The challenge for the nation is both to improve on existing capabilities (the focus of the technology research agenda described in Chapters 3 through 6) and to enhance the potential for manufacturing firms to actually use advanced information technology successfully (the focus of this chapter).
Page 137
Targeting The Decision Maker
Planned change is driven by decision makers. Of course, these decision makers may not necessarily reside in the corporate executive hierarchy; nearly all employees in a company have some decision-making authority. However, people at different levels of a hierarchy have different concerns; as a result, they look to new technologies to answer different questions, as suggested in Figure 7.1. Technology researchers who wish their innovations to be adopted must craft their research in a way that the benefits of the research are highlighted to match the concerns of the decision makers who will decide on their technologies as well as those who will use them. Focusing on the technology alone is rarely sufficient to persuade a decision maker to adopt a particular solution. Unless the use of information technology demonstrably provides substantial payoffs for the end user, the end user will not take the trouble to learn and adapt to the technology, and that will guarantee that the new technology will not be used.
A good example of technology fitting the needs of particular users is evident in the rapid diffusion of personal computers into the workplace. When personal computers became available for a few thousand dollars apiece, they could be deployed to address the problems of decision makers who controlled budgets on that scale without the need for action at higher levels of authority. However, the purpose of introducing personal computers was not to proliferate them or the technology they represent; their use is justified because they can solve real factory problems with a demonstrated dollar benefit.
Motivating Technology Transfer
And Academic-Industrial Interaction
No matter how good are the ideas and advances developed by academic research groups, they are useless in a manufacturing context unless transferred to industry. Effective transfer requires a stronger relationship between academia (or other sources of new technology) and industry.1 Although most of the important recent developments in design and manufacturing have come from industry, manufacturing firms and their information technology system vendors need the knowledge and expertise of the academic research community, whose role is to pursue longer-term research that may not have obvious immediate payoffs. This
1 For example, new design methods developed in academia are especially hard to transfer to industry because industry normally gets such methods from computer-aided design (CAD) vendors. So the transfer takes two steps (from developers to CAD vendors and then from CAD vendors to their customers). However, CAD companies are typically small and very limited in their resources, and so they do not take risks on new research ideas. Instead, they take their cues from their major customers. Consequently, a technology transfer strategy for design methods must enlist the support of potential users as well as equip the CAD vendors.
Page 138
FIGURE 7.1 Questions asked at different levels in the manufacturing hierarchy to which information technology can be expected to be responsive.
Page 139
is not a new problem, but it is particularly acute in the area of manufacturing technology—and resolving it is more important now than ever before, especially since organizations long known for their ability to facilitate the transfer of basic research to practical implementation (e.g., Bell Laboratories, Fairchild R&D, IBM's T.J. Watson Research Center laboratories) are being cut back, reoriented, or even eliminated.
The success of computer-aided design (CAD) programs for electronic design in the marketplace is instructive and provides a good model for university-industry relationships in other manufacturing domains. CAD programs for electronic design have been developed over the last 20 years, primarily in universities and industrial laboratories, often aided by governments grants (mostly from the National Science Foundation (NSF) and the Advanced Research Projects Agency (ARPA)). The transfer of these technologies from laboratories to industry is due to a robust and aggressive set of private-sector vendors, whose role has been largely to implement and package research results for general consumption. Electronic design analysis research is still going strong, primarily in raising the level of design abstraction and in improving the capabilities of existing algorithms. Development and deployment of field-programmable gate arrays are starting to make it possible to build early hardware prototypes, rather than relying on software simulations. Again, private-sector vendors (Xilinx, Actel, Altera for chips, Quickturn for systems) are marketing the basics needed for such prototyping.
Some of the seeds of the problem of transferring technology for manufacturing lie in the culture of universities. Unfortunately, manufacturing is dismissed in many academic circles as a shrinking portion of the gross national product symbolized by unexciting smokestack factories.2 Inducing academics to work on the inherently interdisciplinary and multidisciplinary problems that characterize manufacturing problems is difficult when tenure in universities often depends on evidence of the faculty member's ability to be successful as an independent researcher gaining peer recognition; teamwork and industry collaboration are not regarded as important contributions. Industry recognizes this fact by forming design teams and finding ways to reward all the participants. In academia, many forces are starting to motivate collaborative research, which is presenting new challenges and opportunities to administrators; these include, for example, interdisciplinary manufacturing research centers (CSTB, 1993).
Industry can help by articulating to academia the basic intellectual issues of manufacturing, not only to help guide research (e.g., by providing data and problems to academic researchers) but also to make those issues recognizable to people in traditional disciplines. Strong statements from industry would help to impress on manufacturing faculty (many of whom have limited expertise in information
2 For a perspective on the larger problem of nurturing academic work related to system development, see CSTB (1994b).
Page 140
technology) the need for industry involvement and the increasing importance of information technology to manufacturing. Finally, if industry were widely regarded as being open and accepting of new ideas, academic researchers might have more incentive to work on industry's problems.
A variety of mechanisms could foster better academic-industrial interaction relating to manufacturing. A fellowship or sabbatical program, for example, could place academics in factories and factory people in academia for periods of about a year. (Shorter periods have proven less satisfactory; periods of about a year allow a researcher to better understand the special qualities and inherent problems of the manufacturing environment, to explore its different operations and dimensions (e.g., by rotating through different units), and to contribute substantially to addressing a problem.3) Historically, this concept has been thwarted by tenure pressures that militate against long absences from the university by researchers during the period in which they are forming their research programs. To succeed now, mechanisms are needed to ensure that researchers are not penalized for this kind of investment of time and effort.
The committee is particularly attracted to the concept of a "teaching factory." The study of manufacturing requires a strong laboratory experience, in which difficult concepts and poorly understood interactions can be demonstrated and learning can be reinforced, and a teaching factory is designed to provide this laboratory. A teaching factory would gather together the necessary processing and control equipment (including tools, fixtures, machines, computers, and interfaces) and organize these elements so that a product or family of products could be produced. However, the teaching factory would also be organized to demonstrate and make manifest a variety of activities and principles, including the specifics of processing, control, system and product design, associated management, and their interactions. A teaching factory would serve as a testbed that could help researchers see if their ideas are valid. Setting aside portions of real factories for such experiments is usually too costly to be considered. Yet testbeds must be realistic enough to enable researchers not only to find out if their ideas work but also to learn the constraints of real manufacturing environments and discover opportunities and research problems.4 Research to develop innovative and economic means of creating teaching factories is necessary to prepare manufacturing
3 Industry could reciprocate by funding further education for its workers. One example of a useful exchange is the Massachusetts Institute of Technology's Leaders for Manufacturing program, which links engineering and business school faculty with industrial representatives and involves master theses developed by students "on location" in manufacturing environments.
4 The Metal Oxide Semiconductor Implementation Service (MOSIS) provides a facility to experiment with innovative chip designs; a researcher provides a chip design, and the MOSIS service returns (in a very short time) a chip fabricated to the design's specifications. The concept of a "mechanical MOSIS" service has been proposed in the past, and it continues to be attractive as a vehicle for rapid testing of interesting manufacturing ideas.
Page 141
specialists for the factory of the future as well as to develop deep understanding of specific processes.
One variation on the concept of a teaching factory—the extension of the manufacturing enterprise into academic institutions—would be made possible by a broad, enhanced national information infrastructure (NII) such as that envisioned as enabling greater inter-enterprise integration and the virtual factory concept (see "Inter-enterprise Integration" in Chapter 6). Efforts to develop a capability for remote access should specifically address problems related to electronic connectivity, the transfer of technically complex principles, and shared access (perhaps remotely) to expensive manufacturing equipment and critical information or knowledge.
Any program that provides major funding for academic research in manufacturing will likely increase academia's interest in manufacturing; of course, that is an objective of the federal Advanced Manufacturing Technology initiative. Another mechanism for boosting the prestige of manufacturing would be an NSF postdoctoral program for manufacturing fellows. Matching grants would leverage limited federal funds with industrial resources, using industrial support to signal areas of particular interest to industry. Similarly, funding could be directed to appropriate university-industry partnerships that are launched with explicit mechanisms for technology transfer.
Finally, consortia may be necessary to undertake meaningful manufacturing research, a trend recognized in a number of federal programs.5 The issue of manufacturing is bigger than any company, indeed bigger than any industry. Consortia may be the way to carry out expensive research, especially in areas that transcend specific products. One may question the relevance of, for example, the possible collaboration between an aircraft company and a semiconductor chip
5 For example, the Intelligent Manufacturing Systems (IMS) program is an international program for cooperative research and development in advanced manufacturing that involves the United States, Japan, Canada, Australia, the European Community, and the European Free Trade Association. It is intended to bring together large and small manufacturing companies, research and academic institutions, and public authorities in a structure that facilitates the sharing of intellectual property. In the United States, the IMS program is run out of the office of the assistant secretary for technology policy in the Department of Commerce. This program is not aimed at any particular research topic or direction for attention, though certainly the application of information technology to manufacturing does fall under its umbrella.A second example is the technology deployment activity of ARPA's Technology Reinvestment Project that focuses on manufacturing extension centers (MECs). An MEC is an organization that works with small manufacturers to assist them in using technology (including information technology) to improve performance and productivity. MECs are intended to be consortia consisting of institutions of higher education, nonprofit organizations, or government entities in the service of specified clientele.Other consortia are described in Putting the Information Infrastructure to Work (NIST, 1994, pp. 7-23).
Page 142
company. However, it turns out that there may be many possible benefits of collaboration. Boeing has established a sophisticated design program for its new line of aircraft, and the supplier-customer relationship established to share designs of parts can benefits a chip maker also trying to share design topics. Both companies share concerns about environmental issues, team development, and technology to share information. Interindustry collaboration may indeed provide a rapid path to the introduction of new technology.
The development of the NII is sure to have a significant influence on industry collaboration. Sharing of information will lead to more collaboration and the availability of high-bandwidth communications networks will facilitate more and better ways to collaborate, as will the availability of an infrastructure involving adoption of standards that support interoperability. As graphic and image-rich information management capabilities become available through the NII, for example, the collaboration between industries and suppliers is very likely to be enhanced.
Motivating Introduction And Implementation
Of Information Technology
Understanding Organizational Issues
The introduction of new technologies into an organization almost always affects existing social structures. A factory operation that is as tied to information technology as the vision of Chapter 1 suggests will inevitably have different employment hierarchies, divisions of labor, information flows, and forms of organizing for tasks than those that exist today. To facilitate a smoother integration of new information technology into manufacturing businesses, research is needed in the following areas.
Human Resources
Even in a highly automated and information technology-intensive manufacturing enterprise, human beings will continue to have a critical impact on operations. Important questions regarding the human role must be addressed:
| • | How should responsibilities be divided between people and information technology? At the extremes, this division may be clear (clearly, people should decide if it is socially responsible for a certain product to be produced, and computers should decide at what moment to turn off a cutting torch). However, intermediate cases are less clear, especially as information technology becomes capable of acting with more autonomy and can assume tasks previously thought to require human input. |
Page 143
| • | How should goals and incentives be structured to ensure that the aggregate of individuals' behaviors is most supportive of an enterprise's overall objective? Individuals in a hierarchy may well be much more sensitive to the demands of their immediate environment than to the company's overall goals; most individuals in an organization do what is necessary to satisfy their supervisor, and what the organization wants overall is secondary. In motivating the appropriate behavior, a system that rewards individuals on the basis of satisfying the needs of (internal or external) customers may be more efficacious than current approaches, and information technology may well provide the means to make such a reward system more immediate and thus effective (Holland, 1985, pp. 1-7). |
| • | What mechanisms are needed to give individuals the skills they need? This issue implies training or apprenticeship programs and invites exploration of issues such as just-in-time training for small firms that cannot afford to release staff for more traditional educational environments. Older workers who associate high technology with increased job insecurity may be particularly apprehensive about the introduction of IT. |
Communications
Traditionally, the content of communications is structured as distinctions between data, information, and knowledge. This continuum can be regarded as one of ever-higher abstraction coupled to increasing levels of contextual awareness and relevance. Intuitively, data belong to "low-level" computers while knowledge is the province of "high-level" human actors. But as computers become capable of assuming higher-level functions, what data, information, and knowledge are relevant to which entities? (For example, a person or a computer might have a single function that requires selected communications content from across an entire facility, or it might deal with the entire communications content regarding a geographically and temporally bounded environment. Under what circumstances is each design more appropriate?) How can the communications stream best be made available to workers and for what purposes? How can the differing requirements of individual agents be taken into account? How can information overload on human beings be detected and corrected?
Organization for Ad Hoc Tasks
For substantial tasks, teams are necessary. As enterprise integration for manufacturing expands to include actors not traditionally included in a manufacturing factory complex (e.g., suppliers and customers), issues of team management and organization will assume added importance. For example, the problems of adjusting organizational structures to permit and support flexibility in
Page 144
design and increased interaction with suppliers and customers as they participate in product development constitute one of the fundamental issues manufacturers are wrestling with today. Moreover, the group dynamics of teams with human and computer actors are very poorly understood. Put differently, what should the partnership between person and machine be like?
Informal Hierarchies of Status
In addition to a formal organizational hierarchy (as might be depicted on an organizational chart), most organizations also have informal hierarchies that reflect the greater status or prestige of certain categories or classes of individuals relative to that of others. In manufacturing companies, product designers are often at the top of the pecking order, even though a successful manufacturing operation requires talent and intelligence in other job functions as well. Such status is consistent with the fact that product designers have a much richer array of computer-based tools to help them in their work than do factory managers or process designers. One consequence of this richer array of tools is that product designers can point to computer-supported decisions whereas others usually must rely on experience and intuition, and it would not be surprising if the preferences of product designers were given more (and perhaps undue) weight as a consequence. Preceding chapters have underscored the fact that knowledge about product design is more advanced than that for process design or factory management, but it is also clear that limited resources are more often invested in accordance with the needs of those of higher prestige and influence. Because of a long-standing history of status differentials in U.S. manufacturing organizations, it may be necessary for management to pay explicit attention to differences in status and to make special attempts to listen to the needs of individuals across an organization so that they can contribute on a more equal basis.
Overcoming History and Managing Risk
The history of attempts to introduce computer-integrated manufacturing (CIM) systems into factories demonstrates that companies have tried to adopt these systems without adequate justification or based on the wrong metrics or parameters; the result has been that these large expenditures of dollars have been poorly justified and often wasted. As a result, many decision makers in manufacturing regard CIM technology as technology searching for problems to solve. In some cases, such a perception may taint even good CIM applications.
Moreover, the gap between manufacturing managers (the customers for manufacturing research) and technology ''pushers" (the suppliers of manufacturing research) is often large; technologists at times push technology because it is "neat stuff," whereas managers want and need answers to today's problems. This gap between technology pushers and customers with real problems is significant
Page 145
and is the source of much friction. Many information technology researchers are so immersed in computer technology that they take increasing computerization as a given. These researchers often suffer some hubris about "those people out on the factory floor who cannot write programs or integrate new systems . . . ." Manufacturing managers, including members of this committee, express concern that they already have more technology than they can use. They note that many human obstacles must be overcome before today's, let alone tomorrow's, information technology can be used to benefit manufacturing. Attention to the managerial and cultural preconditions for effective use of information technology is an important element—perhaps the most important element—of the development and application of advanced manufacturing systems (CSTB, 1994a).
The gap between technologists and managers was demonstrated directly to the committee in its interviews with senior manufacturing executives, who spoke in terms of cost, quality, and time to market while the committee spoke in the jargon of information technology. The company executives' view of what was contained in information technology differed strongly from the committee's view; the committee's view of manufacturing was strongly at variance with the executives' view. As one of the interviewees for this report put it, "The managers and the technical experts talk two different languages. The technologists regard the managers as being too tied up in return-on-investment concerns and therefore likely to miss emerging capabilities, while the managers and decision makers regard the technologists as being too uninvolved in real factory problems to understand what the customer really wants and needs." Such discussions demonstrated that both technologists and executives need to make greater efforts to bridge the communications gap before technology can be applied to the real problems faced in manufacturing in companies across the United States.
Manufacturing managers know that realizing the promise of information technologies to help improve factory performance is not easy, and there are many opportunities for mismatched expectations. Expectations may differ between technology supplier and user—indeed, users may have had too little input into the design of new systems. Achieving a solution may take so long that it may come only after the problem has largely disappeared or been fixed by other means, or the conditions have changed so much that the solution is no longer applicable. A new system may not make use of embedded legacy systems that are too expensive to discard or too integral to operations to modify. The skills required may differ significantly from those of the existing work force (see "Education, Training, and Retraining" below).
Thus, the resistance of many senior managers to the introduction of information technology into real manufacturing operations is not unreasonable, and it arises from the fact that many of the new information technologies are untested and unproven in actual manufacturing lines, and in some cases have offered poor results. New systems are not used because they are untested, and they cannot be tested because no one wants to be the guinea pig.
Pages 146
| BOX 7.1 Guidelines for Introducing Innovative Information Technology in Real Production Lines | |
| • | Take small steps. Deliver improvements incrementally, rather than striving to deliver "the solution" all at once. Deliver improvements that can be installed with lower levels of management authority. |
| • | Lower the risk. Avoid serious disruption of manufacturing facilities while applications are being tested or while systems are being introduced onto the factory floor. |
| • | Deliver real solutions to real problems; focus on the customer, not the technology. Indeed, there is a great deal of good technology already waiting to be used. Develop the customer "pull" in concert with the technology "push." |
| • | Develop a vision and a set of strategies that have temporal longevity. Create a system that withstands the ups and downs of yearly budget plans. For example, "lean" manufacturing is still not 100 percent utilized, in spite of the fact that it was started over 30 years ago! |
| • | Work in teams with customers and suppliers. Since having a viable supplier base is a key to success, it is important to make suppliers a part of the solution, rather than regard them as part of the problem. Similarly, the known (and perhaps anticipated) needs of customers should be the ultimate focus of installations of information technology. |
| • | Develop systems that build on legacy systems, unless it appears that the legacy system is so hopelessly useless that it must be jettisoned. It is hard to quickly change the momentum of manufacturing enterprises; try to make incremental improvements. |
| • | Where possible, buy off-the-shelf information technology from commercial suppliers. In general, it is better to use commercial vendors than to undertake custom development projects on one's own. Doing the latter forces one to inherit all the issues that relate to product and application support for the product life cycle. |
Another contributor to risk aversion is time pressure. Bringing in new systems consumes valuable production time, which may not easily be recovered. It is easier (and therefore common) for a factory to use an old, perhaps less efficient, system than to suffer "downtime" as the result of installing and testing a new system. Box 7.1 provides some guidelines for introducing new information technologies that are responsive to the difficulties that line managers face in approaching these technologies.
The scope of change inherent in manufacturing systems can confound financial analyses. The financial impact may appear too large to justify for a single factory, even if the new technology appears worthwhile for the enterprise overall. In addition, the benefits of investing in a new system may be hard to quantify, especially if they relate to quality; metrics for improved quality, service, and accessibility are particularly problematic. Research is needed to support more appropriate financial and business justifications and to better define the cost of
Page 147
quality, the cost of insufficient quality, and the trade-offs among quality, time, and money (Box 7.2).6
One mechanism that could contribute to better planning for the time and real cost involved in implementing new systems (although not necessarily changing those quantities) would be an extension of the "beta test" concept into manufacturing environments.7 Beta tests can assess the fidelity of a product to its design specification, augmenting a real production environment with special instrumentation to assess errors or relay feedback to developers. Although pilot plants have been used for years as laboratories to assess production concepts, research is needed to explore how actual production facilities, operating at normal volumes, can be used directly for achieving continuous improvements or innovation. Legitimizing beta test periods would disclose the need for and real costs of a practice to manufacturing management at the proposal stage. The beta test concept could also provide a basis for links to research facilities, with the benefit of enhancing the understanding of what does and does not work in a given type of system or application. Research could help to define the tools, techniques, and practices appropriate to beta testing in manufacturing contexts.
Finally, the development of the NII may, for entirely psychological reasons, stimulate the incorporation of information technology into manufacturing. To the extent that the NII is a medium that ordinary people use every day (e.g., to make purchases, to read newspapers, to order and watch movies, and to transmit pictures, images, and voice and other messages), manufacturing managers will become more familiar with information technology, and the part of managerial resistance to information technology in factories that is attributable to unfamiliarity will likely diminish.
Providing for Technology Demonstrations
In the absence of convincing evidence that the benefits of information technology far outweigh the costs, effects of factory disruption, and time required to do the job, many managers and executives in manufacturing enterprises will be understandably risk-averse. For these individuals, there is a world of difference between a promise and a concrete demonstration of a technology's feasibility and
6 For example, the economic production of customized manufactured products (i.e., production lot sizes of one or a few) may well depend on the adoption of new accounting systems in which the entire life-cycle cost of a product is taken into account; life-cycle costs include transportation costs, inventory costs, obsolescence costs, reengineering and retrofitting costs, and the cost of product recall, if any.
7 Software developers draw on objective third-party users in beta tests of new products or versions to verify functional capabilities and ensure lack of defects. Explicitly applying the concept and practice of a beta testing period into production launch activities would provide for significant first trials of the custom software typical of manufacturing information systems.
Page 148
| BOX 7.2 Information Technology Deployment and Justification | |
| Information technology was originally used in fairly narrow ways: accounting and clerical functions were automated, personnel records were placed in rudimentary databases, and controllers were developed for individual machines. For these applications the financial justification could be made on fairly traditional grounds, such as cost reduction (usually by machines and systems replacing people) or the improved quality associated with better machine control. As information technology has matured, as the cost mix (hardware, software, and labor) associated with it has changed, and as its applications have broadened, it has become more difficult to comprehend the impact that an information system will have and, consequently, to judge whether or not an investment in the technology will pay off (CSTB, 1994a; Kaplan, 1986, 1989). The benefits of better understanding would extend to organizations of all sizes. It has proven particularly difficult to justify the costs of providing the infrastructure required to maintain and enhance information technology. There is no straightforward relationship between information services and the yield or quality of the finished product. Consequently, justifications for the infrastructure have to be based on more speculative, qualitative rather than quantitative, arguments. Better metrics are needed for measuring the value of support systems. These could also contribute to better cost management. Recent efforts to develop activity-based costing may be helpful to manufacturers in this context; additional research into applications of this approach in the manufacturing context may be useful. As companies attempt to become more agile, reduce the time required to respond to market needs, and explore the merits of "virtual enterprises," they will need tools to help determine what should be measured. How, for example, should one translate the metrics of agility into realistic financial terms? How can one measure the costs at organizational boundaries and judge whether information technology will improve an operation enough to justify the costs of the system? The speed with which enterprise-wide systems will be deployed will, to a large degree, depend on how well companies are able to understand these issues. That understanding would benefit from research into "exchange rates" between agility and financial metrics as well as case studies of enterprises that have implemented complex information systems. |
desirability. For this reason, the committee is drawn strongly to the notion of an advanced long-range technology demonstration (ALRTD).8
ALRTDs focus on a variety of needs, covering both technical and cultural issues. ALRTDs are intended to provide an active means for transferring promising research into practice and move technological concepts closer to business
8 An advanced long-range technology demonstration differs from the Advanced Technology Demonstration program of the National Institute of Standards and Technology in that the latter is focused primarily on demonstrations of technology just before it is mature enough to support product development. The technology associated with an ALRTD would generally be at an earlier stage of development at the time the ALRTD was proposed.
Page 149
opportunities; connect researchers to problems, reveal knowledge gaps, and test research ideas; and bridge a variety of "believability gaps" that keep new ideas from being tried in the real world. Successful ALRTDs would be vehicles that could help to change manufacturing management skeptics into supporters and believers without making them suffer through a painful process of education within their own factory domain, and they would help answer the question, Is there a performance benefit to making changes on the factory floor?
For a candidate ALRTD to be viable and effective, it must be focused and bounded; Box 7.3 provides a number of areas in which ALRTDs might prove helpful. Since some investment in technology will almost certainly be required (in some cases, as much as that required for a pilot manufacturing plant), it must be funded well beyond the level of technical research grants. So that its success (or lack thereof) can be determined, quantitative goals or metrics that can serve as yardsticks should be established in advance. Moreover, each ALRTD should have core nontechnological content, such as a specific business activity, so that the technology will be the means to a clearly defined end rather than an end unto itself. Finally, each ALRTD should have some growth or follow-on potential and should not be a dead end.
Each ALRTD proposed should be based on a vision of what will be possible if the ALRTD is implemented, including what knowledge will be obtained if the ALRTD is successful. It should have available to it the necessary physical and information infrastructure, including other ALRTDs that could be under way at the same time or should have been completed first. It should clearly state the capabilities that will be demonstrated. Finally, it should describe expected consequences of a successful ALRTD that go beyond the explicit things to be demonstrated.
Appendix C describes several possible ALRTDs that are consistent with this description.
Stimulating The Adoption Of Open Standards
As discussed in previous chapters, the lack of standards is a major inhibitor to the use of existing information technology for manufacturing. From computer-aided product design tools to real-time machine controllers, the lack of standards in various forms prevents data interchange, increases training costs, and impedes would-be product developers who have to build basic services into their products from scratch. Smaller manufacturers, especially, stand to benefit from the wide promulgation and acceptance of standards for interfaces and interconnections, because such standards are likely to lead to a much wider selection of information technology products for manufacturing applications (as more vendors will be able to develop products based on a common standard) and lower prices for such products (as vendors will be able to avoid the cost of developing unique solutions to common problems).
Page 150
| BOX 7.3 Areas for an Advanced Long-range Technology Demonstration | |
| • | Executive level Strategic analysis and evaluation of alternatives; Strategy formulation; Planning for implementation of strategies |
| • | Finances Debit management; Cash management; Currency and commodities hedging; Profitability |
| • | Human resources Hiring metrics; Team management; Training; Rewards; Multinational teams |
| • | Sales and marketing Gathering customer input; Forecasting sales by product and geography |
| • | Product design and development R&D, design, prototyping, and verification; Design methodologies and management; Computer-aided design tools; Process design and verification |
| • | Production Fabrication, assembly, logistics; Distribution; Inventory control; Resource procurement and use planning; Procurement, source qualification; Maintenance and repair of facilities |
| • | Legal concerns Alliances; Contracts; Safety; Environment; Liability |
What prevents standards from being adopted more broadly? One answer is that the premature imposition of standards tends to freeze technological development and progress, and manufacturing is certainly one arena in which the use of information technology is relatively immature. But a second important barrier to adopting standards is not rooted in technical issues—a company trying to differentiate its products from those of other vendors may well choose to make them incompatible with others. Representing data in ways that facilitate interoperability may subject a vendor to infringement liability or require it to pay royalties.
Page 151
Finally, once many competing products employ different representations and interfaces, a company may well resist adopting a standard proposed by another company for fear of ceding a commercial advantage to that company.
Previous chapters suggest the need for technical work on the development of open standards that will still allow technological progress. But there is a sociological dimension to standards as well—even when the underlying science and generic enabling technology are genuinely capable of supporting useful commercial applications, the vendor community needs to find some way to agree on standards that will expand the market for all.
Many problems need to be overcome if agreement is to be reached. For example, trends toward distributed intelligence, as well as the need to form alliances with other business entities (including outside design and engineering services as well as material and equipment suppliers), raise important knowledge management issues. How should companies determine what proprietary knowledge to share through distributed intelligence? What are ways to share such knowledge that protect knowledge assets and avoid disruptive interactions? How can commercial, off-the-shelf systems and technology best be combined with proprietary knowledge, perhaps in the form of software and knowledge bases? Such difficulties have already appeared in the very large scale integration applications-specific integrated circuit industry: vendors that own significant design libraries (for cells or "macrocells") are very reluctant to disclose detailed models of these cells to their customers, for fear that their customers' ability to do reliable simulations will hurt them competitively.
A second problem is how to achieve an appropriate balance between reducing the costs of variation and enabling rapid implementation of new technology—in short, the fundamental tension over whether and when to standardize. Research could help to establish criteria for determining when the benefits of restricting options by the adoption of standards outweigh the costs.
A third problem is how to generate widespread acceptance of and conformance to standards. Because of the large number of parties with vested interests in the particulars of any given standard, a careful and deliberative approach with wide representation is necessary. However, while wide representation is often a facilitator of consensus and acceptance, it is a factor that tends to slow down the consensus-forming process and also sometimes discourages the exploitation of technological opportunities that may spontaneously arise. In addition, once a standard is in place, a substantial amount of education is often necessary to familiarize potential users with its benefits and its technical implications.
Developing Better Metrics Of Performance
Many metrics used today measure factory performance at high levels of aggregation that do not provide managers with insight into what is going on at the activity level. For example, metrics often used today include inventory turns (a
Page 152
measure of how rapidly inventory flows through the manufacturing system), defect rates (how often manufacturing problems occur), and on-time deliveries (the extent to which manufacturing schedules are met). These metrics are useful to signal the existence of problems somewhere within a manufacturing operation but cannot by themselves identify sources of the problems, which are affected by many factors.
Moreover, when the use of new technologies and approaches leads to new combinations of previously separate activities, many traditional measures of performance are problematic at best—and perhaps outright misleading or useless. For example, repetitive activities are much easier to measure than others, and performance metrics are often based on what is easy to measure. How does one use metrics based on repetition when the very point of information technology is to increase flexibility and the ability to cope with one-of-a-kind new situations and new sets of circumstances? Understanding in detail the impact of information technology is crucial for obtaining the best value, because only a detailed analysis can separate the impact of information technology from the effects of all of the other factors that may also impinge on overall performance.
Research is needed to develop good metrics that reflect the impact of information technology on manufacturing activities; a good beginning would be research on measuring the effect of information technology applications using existing metrics. Progress in this area will require both sophisticated technical understanding and a shared understanding within the community on the appropriateness of these metrics.
A related issue is that trustworthy cost-benefit analyses require reliable data. Obtaining meaningful data on costs is typically much more difficult than budget analysts anticipate. Consistent definitions are lacking, and typical cost categories and allocations are designed in accordance with conventional accounting rules rather than for tracking the performance of integrated business practices. New accounting categories based on the specifics of manufacturing may well be needed if an accurate picture is to be drawn of dollar flows in manufacturing.
Developing New Models For
Accounting And Valuation
The complexity of a large manufacturing business makes it very difficult for managers to identify operations that cost more than their value warrants. It is thus necessary to develop detailed models of a business that truly relate causes and effects, costs and benefits, and problems and solutions. Without such models, it will be difficult to convince anyone of the wisdom of taking certain courses of action, especially those that may be counter-intuitive.
Better management accounting systems are an integral part of these models. Timely activity-based accounting systems can provide explicit guidance regarding where costs can be trimmed and information on design parameters that demand
Page 153
resources.9 With traditional accounting, the overhead costs of operating a plant are spread over all items produced in proportion to the volume of each type of item. But with activity-based accounting systems, the actual cost of small-volume production can be properly attributed and thus recovered.
Finally, generally accepted accounting principles that businesses use to audit their finances and operations are derived from a business philosophy in which capital expenditures (i.e., expenditures that relate to the long-term value of a company) are associated with buildings and pieces of equipment. Although this philosophy was appropriate in manufacturing 30 years ago, its relevance has decreased as the basis for competitive advantage has increasingly become the ability to exploit information more effectively. In particular, under old accounting principles the contribution of various types of knowledge (e.g., skilled personnel, software, and organizational structure) to the book value of a firm is essentially zero, even if that knowledge is the primary enabler of that firm's success.
It is admittedly difficult to place a book value on assets that are intangible and inscrutable to the layperson or that seemingly lack durability. Senior management is often legitimately concerned about the potential for misleading financial reporting as a product of estimating the value of intangible assets such as knowledge, and a variety of stakeholders (e.g., analysts, investors and owners, managers of organizations using software, managers of organizations producing software, tax authorities) are likely to have different judgments on what such assets are "truly" worth. Nonetheless, given the importance of knowledge to the future competitive environment, serious research to find valuation schemes that appropriately account for the contribution of knowledge to value is well worthwhile.10
Emphasizing Education, Training, And Retraining
The obstacles described above are among the factors motivating the current concern of management theorists and analysts with the "management of change." According to this perspective, successful enterprises of the future must change rapidly and frequently. Information, types of media, tools, processes, organizational structures, and so on will change, rendering knowledge and skills obsolete several times in an individual's career. Continuing education and training will have to inculcate adaptability and the expectation of change. These outcomes will be difficult to achieve because change provokes anxiety and stress. Research into techniques and tools to help enterprises and individuals understand and
9 See, for example, Nanni et al. (1988).
10 Similar concerns were raised for service industries in Information Technology in the Service Society: A Twenty-first Century Lever (CSTB, 1994a).
Page 154
manage the process of change could facilitate the kinds of evolution of organizations and skills that many analysts anticipate.
Expectations for frequent change and greater employee autonomy imply that the manufacturing enterprise of the future will depend more on the knowledge and skill sets of each employee. Several needs follow from this assumption:
| • | Education of senior managers, to help them make informed decisions about using new, expensive, and perhaps risky information technology; |
| • | Education and training of middle managers, to help them reduce barriers to trying new technology that may seem only to consume time and resources or imperil their own jobs; |
| • | Education of designers, engineers, manufacturing engineers, and toolmakers to give them an increased technical base so that they can better understand and use computer technology and design relevant software; |
| • | Training of factory workers, who are expected to use this new technology to improve their performance and that of their facility; and |
| • | Retraining of factory workers whose skills are not compatible with the information technology paradigm and who may have to seek alternative jobs. |
Even the term ''frequent change" may understate the rapidity with which individuals will have to learn new things. Indeed, the volume of information that is relevant to successful individual performance is both so large and so rapidly changing that a learning model based on discrete phases of "retraining" may not be entirely sufficient. Another notion—what might be called "just-in-time learning"—is based on the idea that people should learn new skills and information as they need them to do what they must do. While such a notion has applicability that goes beyond manufacturing (indeed, it has ramifications for all of education), the manufacturing domain may be one that is particularly suitable for exploring the potential of this notion through intelligent tutoring systems, long-distance learning systems, and multimedia experiential learning tools.
Education issues are ubiquitous and topical; in addition, the renewal and currency of employee process- and product-specific knowledge have long been a major problem for all manufacturing enterprises. Needed are better means of delivery and better understanding of human issues such as motivation (or fear based on inadequate technical education and the resulting wish to fall back on intuition and experience).
Possible approaches to addressing the need for ongoing education include conferences and seminars for senior and middle managers; publication of educational materials; close collaboration among academicians, designers, engineers, and factory workers; and television and audio tape courses.
Computer aids—computer-aided instruction—also may be brought to bear on the education and training requirements of an enterprise, providing efficient and customized education that allows a person to learn quickly and selectively.
Page 155
Instructional programs of the future will be more interactive and user-friendly than the lecture and fixed-sequence documentaries of today. Computer aids for learning may be built into new manufacturing technologies themselves, as well as installed in stand-alone workstations for educational purposes. Multimedia and virtual reality technology may hold promise for providing a flexible, socially acceptable, and nonthreatening interface for educational and skill-building programs; an NII and future home and factory systems offer the means to distribute and use such programs. The delivery systems will have to provide overview data as well as fundamental knowledge. NSF support of these technologies, perhaps through Education and Human Resources Directorate programs, could increase the skill base of the manufacturing sector, as well as of other sectors of society in general.
The challenge of providing new educational technology can easily be underestimated. But it is a complex, multifaceted task. The target population spans a wide range of educational backgrounds, capabilities, and needs. Some individuals will want a high-level overview of the material, and others will want in-depth courses with detailed practice sessions. The work force today spans multiple languages in the domestic United States, and specific topics will also be made available to people in other countries and from different cultures. Thus the content of educational programs and information is critical, the quality of presentation must meet modern expectations, and usability must be tunable to purpose.
