New Manufacturing Technologies

byCharles H. Fine

WP #3009-89-MS May 1989

To appear in:Strategic Manufacturing: Dynamic New Directions for the 1990'sNew York: Dow Jones Irwin; Patricia Moody, editor.

NEW MANUFACTURING TECHNOLOGIESCharles H. Fine

INTRODUCTION

Driven by international competition and aided byapplication of computer technology, manufacturing firms havebeen pursuing two principal approaches during the 1980's:

* automation, and* integration.

Automation is the substitution of machine for humanfunction; integration is the reduction or elimination ofbuffers between physical or organizational entities. Thestrategy behind manufacturing firms' application of newautomation technologies is multidimensional:

* to liberate human resources for knowledge work,* to eliminate hazardous or unpleasant jobs,* to improve product uniformity, and* to reduce costs and variability.

The execution of that strategy has lead firms automate awaysimple, repetitive, or unpleasant functions in their offices,factories, and laboratories.

Integration, when used as an approach to improvequality, cost, and responsiveness to customers, requires thatfirms find ways to reduce physical, temporal, andorganizational barriers among various functions. Such bufferreduction has been implemented by the elimination of waste,the substitution of information for inventory, the insertionof computer technology, or some combination of these.

In most process industries - oil refining and paper-making, for example - automation and integration have beencritical trends for decades. However, in discrete goodsmanufacture - electronics and automobiles, for example -significant movement in these directions is a recentphenomenon in the United States.

This chapter defines, examines, and illustrates theapplication of technologies that support the trends towardmore automation and integration in discrete goodsmanufacturing. We begin with a discussion of thetechnological hardware and software that has been evolving.We then look at six management challenges that must beaddressed to support these trends. And, finally, we look atthe issue of economic evaluation the new technologies.

AUTOMATION IN MANUFACTURING

As characterized, for example, by Toshiba, in their OMEWorks facility, automation in manufacturing can be dividedinto three categories:

1

*factory automation,*engineering automation, and*planning and control automation.

Automation in these three areas can occur independently, butcoordination among the three, as is being pursued by thisToshiba facility, drives opportunities for computer-integrated manufacturing, discussed below.

Factory AutomationAlthough software also plays a critical role, factory

automation is typically described by the technologicalhardware used in manufacturing: robots, numericallycontrolled (NC) machine tools, and automated materialhandling systems. Increasingly, these technologies are usedin larger, integrated systems, known as manufacturing cellsor flexible manufacturing systems (FMSs).

The term robot refers to a piece of automatedequipment, typically programmable, that can be used formoving material to be worked on (pick and place) orassembling components into a larger device. Robots are alsoused to substitute for direct human labor in the use of toolsor equipment, as is done, for example, by a painting robot,or a welding robot, which both positions the welder and weldsjoints and seams. Robots can vary significantly incomplexity, from simple single-axis programmable controllersto sophisticated multi-axis machines with microprocessorcontrol and real-time, closed-loop feedback and adjustment.

A numerically-controlled (NC) machine tool is amachine tool that can be run by a computer program thatdirects the machine in its operations. A stand-alone NCmachine needs to have the workpieces, tools, and NC programsloaded and unloaded by an operator. However, once an NCmachine is running a program on a workpiece, it requiressignificantly less operator involvement than a manually-operated machine.

A CNC (computer numerically-controlled) machinetool typically has a small computer dedicated to it, so thatprograms can be developed and stored locally. In addition,some CNC tools have automated parts loading and toolchanging. CNC tools typically have real-time, on-lineprogram development capabilities, so that operators canimplement engineering changes rapidly.

A DNC (distributed numerically-controlled) systemconsists of numerous CNC tools linked together by a largercomputer system that downloads NC programs to the distributedmachine tools. Such a system is necessary for the ultimateintegration of parts machining with production planning andscheduling.

Automated inspection of work can also be realizedwith, for example, vision systems or pressure-sensitivesensors. Inspection work tends to be tedious and prone toerrors, especially in very high volume manufacturingsettings, so it is a good candidate for automation. However,

2

-------�11�1`��---�1"-""�� �1�----

automated inspection (especially with diagnosis capability)tends to be very difficult and expensive. This situation,where automated inspection systems are expensive to develop,but human inspection is error-prone, demonstrates the valueof automated manufacturing systems with very highreliability: In such systems, inspection and teststrategies can be developed to exploit the high-reliabilityfeatures, with the potential to reduce significantly thetotal cost of manufacture and test.

Automated material handling systems move workpiecesamong work centers, storage locations, and shipping points.These systems may include autonomous guided vehicles,conveyor systems, or systems of rails. By connectingseparate points in the production system, automated materialhandling systems serve an integration function, reducing thetime delays between different points in the productionprocess. These systems force process layout designers todepict clearly the path of each workpiece and often make iteconomical to transport workpieces in small batches,providing the potential for reduced wait times and idleness.

A flexible manufacturing system (FMS) is a systemthat connects automated workstations with a material handlingsystem to provide a multi-stage automated manufacturingcapability for a wider range of parts than is typically madeon a highly-automated, nonflexible, transfer line. Thesesystems provide flexibility because both the operationsperformed at each work station and the routing of parts amongwork stations can be varied with software controls.

The promise of FMS technology is to provide thecapability for flexibility approaching that available in ajob shop with equipment utilizations approaching what can beachieved with a transfer line. In fact, an FMS is atechnology intermediate to these two extremes, but goodmanagement can help in pushing both frontiers simultaneously.

Automated factories can differ significantly withrespect to their strategic purpose and impact. Two examples,Matsushita and General Electric, may be instructive.

In Osaka, Japan, Matsushita Electric Industrial Companyhas a plant that produces video cassette recorders (VCRs).The heart of the operation features a highly automatedrobotic assembly line with 100-plus work stations. Exceptfor a number of trouble-shooting operators and processimprovement engineers, this line can run, with very littlehuman intervention, for close to 24 hours per day, turningout any combination of 200 VCR models. As of August 1988,the facility was underutilized; Matsushita was poised toincrease production, by running the facility more hours permonth, as demand materialized.

In this situation, the marginal cost of producing moreoutput is very low. Matsushita has effectively created abarrier to entry in the VCR industry, making it verydifficult for entrants to compete on price.

The second example is General Electric's Aircraft EngineGroup Plant III, in Lynn, Massachusetts. This fully

3

automated plant machines a small set of parts used by theAircraft Engine Group's assembly plant. In contrast toMatsushita's plant, which provides strategic advantage in theVCR product market, the strategic advantage provided by GE'splant seems to address its labor market. Plant III'sinvestment is now sunk. Eventually, it will run around theclock at very high utilization rates with a very small crew.As volume is ramped up, GE has the ability to use Plant III'scapacity and cost structure as leverage with its unionizedlabor force which is currently making many of the parts thatcould eventually be transferred to Plant III. Thus, factoryautomation can address a variety of types of strategic needs,from product market considerations to labor market concerns.

Engineering AutomationFrom analyzing initial concepts to finalizing process

plans, engineering functions that precede and supportmanufacturing are becoming increasingly automated. In manyrespects, engineering automation is very similar to factoryautomation; both phenomena can dramatically improve laborproductivity and both increase the proportion of knowledgework for the remaining employees. However, for manycompanies, the economic payback structure and thejustification procedures for the two technologies can bequite different.

This difference between engineering automation andfactory automation stems from a difference in the scaleeconomies of the two types of technologies. In manysettings, the minimum efficient scale for engineeringautomation is quite low. Investment in an engineeringworkstation can often be justified whether or not it isnetworked and integrated into the larger system. The first-order improvement of the engineer's productivity issufficient.

For justification of factory automation, the reverse ismore frequently the case. The term "island of automation"has come to connote a small investment in factory automationthat, by itself, provides a poor return on investment. Manyfirms believe that factory automation investments must bewell integrated and widespread in the operation before thestrategic benefits of quality, lead time, and flexibilitymanifest themselves.

Computer-aided design is sometimes used as anumbrella term for computer-aided drafting, computer-aidedengineering analysis, and computer-aided process planning.These technologies can be used to automate significantamounts of the drudgery out of engineering design work, sothat engineers can concentrate more of their time and energyon being creative and evaluating a wider range of possibledesign ideas. For the near future machines will nottypically design products. The design function remainsalmost completely within the human domain.

4

III

Computer-aided engineering allows the user to applynecessary engineering analysis, such as finite elementanalysis, to propose designs while they are in the drawing-board stage. This capability can reduce dramatically theneed for time-consuming prototype workup and test during theproduct development period.

Computer-aided process planning helps to automatethe manufacturing engineer's work of developing process plansfor a product, once the product has been designed.

Planning and Control AutomationPlanning and control automation is most closely

associated with material requirements planning (MRP).Classical MRP develops production plans and schedules byusing product bills of materials and production lead times toexplode customer orders and demand forecasts netted againstcurrent and projected inventory levels. MRP II systems(second-generation MRP) are manufacturing resource planningsystems that build on the basic MRP logic, but also includemodules for shop floor control, resource requirementsplanning, inventory analysis, forecasting, purchasing, orderprocessing, cost accounting, and capacity planning in variouslevels of detail.

The economic considerations for investment in planningand control automation are more similar to that forinvestment in factory automation than that for engineeringautomation. The returns from an investment in an MRP IIsystem can only be estimated by analyzing the entiremanufacturing operation, as is also the case for factoryautomation. The integration function of the technologyprovides a significant portion of the benefits.

INTEGRATION IN MANUFACTURING

Four important movements in the manufacturing arena arepushing the implementation of greater integration inmanufacturing:

* Just-in-Time manufacturing (JIT),* Design for Manufacturability (DFM),* Quality Function Deployment (QFD), and* Computer-integrated Manufacturing (CIM).

Of these, CIM is the only one directly related to newcomputer technology. JIT, QFD, and DFM, which areorganization management approaches, are not inherentlycomputer-oriented and do not rely on any new technologicaldevelopments. We will look at them briefly here because theyare important to the changes that many manufacturingorganizations are undertaking and because their integrationobjectives are very consonant with those of CIM.

Just-in-Time Manufacturing (JIT)JIT embodies the idea of pursuing streamlined or

continuous-flow production for the manufacture of discrete

5

III

goods. Central to the philosophy is the idea of reducingmanufacturing setup times, variability, inventory buffers,and lead times in the entire production system, from vendorsthrough to customers, in order to achieve high productquality (conformity), fast and reliable delivery performance,and low costs.

The reduction of time and inventory buffers between workstations in a factory, and between a vendor and itscustomers, creates a more integrated production system.People at each work center develop a better awareness of theneeds and problems of their predecessors and successors.This awareness, coupled with a cooperative work culture, canhelp significantly with quality improvement and variabilityreduction.

Investment in technology, that is, machines andcomputers, is not required for the implementation of JIT.Rather, JIT is a management technology that relies primarilyon persistence in pursuing continuous incremental improvementin manufacturing operations. JIT accomplishes some of thesame integration objectives achieved by CIM, withoutsignificant capital investment. Just as it is difficult toquantify the costs and benefits of investments in (hard)factory automation, it is also difficult to quantify costsand benefits of a "soft" technology such as JIT. A fewrecent models have attempted to do such a quantification, butthat body of work has not been widely applied.

Design for Manufacturability (DFM)This approach is sometimes called concurrent design or

simultaneous engineering. DFM is a set of concepts related topursuing closer communication and cooperation among designengineers, process engineers, and manufacturing personnel.In many engineering organizations, traditional productdevelopment practice was to have product designers finishtheir work before process designers could even start theirs.Products developed in such a fashion would inevitably requiresignificant engineering changes as the manufacturingengineers struggled to find a way to produce the product involume at low cost with high uniformity.

Ouality Function Deployment (OFD)Closely related to Design for Manufacturability is the

concept of Quality Function Deployment (QFD) which requiresincreased communication among product designers, marketingpersonnel, and the ultimate product users. In manyorganizations, once an initial product concept was developed,long periods would pass without significant interactionbetween marketing personnel and the engineering designers.As a result, as the designers confronted a myriad oftechnical decisions and tradeoffs, they would make choiceswith little marketing or customer input. Such practicesoften led to long delays in product introduction becauseredesign work was necessary once the marketing people finally

6

got to see the prototypes. QFD formalizes interactionbetween marketing and engineering groups throughout theproduct development cycle, assuring that design decisions aremade with full knowledge of all technical and market tradeoffconsiderations.

Taken together, Design for Manufacturability and QualityFunction Deployment promote integration among engineering,marketing, and manufacturing to reduce the total productdevelopment cycle and to improve the quality of the productdesign, as perceived by both the manufacturing organizationand the customers who will buy the product.

Like Just-in-Time, Design for Manufacturability andQuality Function Deployment are not primarily technologicalin nature. However, technologies such as Computer-aidedDesign can often be utilized as tools for fosteringengineering/manufacturing/marketing integration. In a sense,such usage can be considered as the application of computer-integrated manufacturing to implement these policy choices.

Computer-intearated Manufacturing (CIM)

Computer-integrated manufacturing refers to the use ofcomputer technology to link together all functions related tothe manufacture of a product. CIM is therefore both aninformation system and a manufacturing control system.Because its intent is so all-encompassing, even describingCIM in a meaningful way can be difficult.

We describe briefly one relatively simple conceptualmodel that covers the principal information needs and flowsin a manufacturing firm. The model consists of two types of

system components:* departments that supply and/or use information, and* processes that transform, combine, or manipulate

information in some manner.The nine departments in the model are:

1. production2. purchasing3. sales/marketing4. industrial amd manufacturing engineering5. product design engineering6. materials management and production planning7. controller/finance/accounting8. plant and corporate management9. quality assurance.

The nine processes that transform, combine, ormanipulate information in some manner are:

1. cost analysis2. inventory analysis3. product line analysis4. quality analysis5. workforce analysis6. master scheduling7. material requirements planning (MRP)8. plant and equipment investment

7

9. process design and layout.

To complete the specification of the model for aspecific manufacturing system, one must catalog theinformation flows among the departments and informationprocesses listed above. Such an information flow map canserve as a conceptual blueprint for CIM design, and can aidin visualizing the scope and function of a CIM informationsystem.

Design and implementation of a computer system to linktogether all of these information suppliers, processors, andusers is typically a long, difficult, and expensive task.Such a system must serve the needs of a diverse group ofusers, and must typically bridge a variety of differentsoftware and hardware subsystems.

The economic benefits from such a system come fromfaster and more reliable communication among employees withinthe organization and the resulting improvements in productquality and lead times.

Since many of the benefits o a CIM system are eitherintangible or very difficult to quantify, the decision topursue a CIM program must be based on a long term, strategiccommitment to improve manufacturing capabilities.Traditional return-on-investment evaluation procedures thatcharacterize the decision-making processes of many U.S.manufacturing concerns will not justify the tremendous amountof capital and time required to aggressively pursue CIM.Despite the high cost and uncertainty associated with CIMimplementation, most large U.S. manufacturing companies areinvesting some resources to explore the feasibility of usingcomputerized information sytems to integrate the variousfunctions of their organizations.

TECHNOLOGY ADOPTION CONSEQUENCES: FLEXIBILITY ANDCAPITAL INTENSIVENESS

As explained above, investments in factory automationand CIM move a firm in the direction of more automation andintegration. To fully evaluate such investmentopportunities, and to weigh the potential pay-offs againstthe costs,' one must consider two consequences of thesetechnologies:

1) the flexibility of the manufacturing operation, and2) the capital intensiveness of the operation.

In this section, we look briefly at these two effects beforediscussing six challenges created by the new manufacturingtechnologies.

Manufacturing flexibility - flexibility to changeproduct mix, to change production rate, and to introduce newproducts - is achieved by shortening lead times within themanufacturing system and by automating setups and changeoversfor different products. The importance of manufacturingflexibility for firm competitiveness has become apparent overthe past decade as the rate of economic and technological

8

change has accelerated and as many consumer and industrialmarkets have become increasingly internationalized.

As a consequence of this increased competition, productlife cycles shorten as each firm tries to keep up with thenew offerings of a larger group of industrial rivals.

To survive, companies must respond quickly and flexiblyto competitive threats. Therefore, firms must pay particularattention to evaluating the flexibility component of the newmanufacturing technologies.

Increased capital intensiveness follows directly fromautomation on a large scale that replaces humans withmachines. A transformation to a capital intensive coststructure has two important effects.

First, the manufacturing cost structure changes, fromone with low fixed investment and high unit variable costs,to one with high fixed investment and low variable costs.This change will affect significantly a firm's ability toweather competitive challenges, because low variable costsallow a firm to sustain short-term profitability even in theface of severe price wars.

Second, the changes in both employment levels and workresponsibilities brought about by automation requiresignificant organizational adjustment. Challenges broughtabout by this type of change are discussed below.

SIX CHALLENGES CREATED BY THE NEW MANUFACTURINGTECHNOLOGIES

1.Desian and Development of CIM SystemsBecause of their ambitious integration objectives, CIM

systems will be large, complex information systems. Ideally,the design process should start with the enunciation of theCIM mission, followed by a statement of specific goals andtasks. Such a top-down design approach insures that thehardware and software components are engineered into acohesive system.

In addition, since the foundation of CIM consists of anintegrated central database plus distributed databases,database design is critical. Also, since many people in theorganization will be responsible for entering data into thesystem, they must understand how their functions interactwith the entire system. Input from users must be consideredat the design stage, and systems for checking databaseaccuracy and integrity must be included.

Hardware and software standardization must also beconsidered at the system design stage. At many companies,computing and database capabilities have come from a widevariety of vendors whose products are not particularlycompatible. Either retooling, or developing systems to linkthese computers together, requires significant resources.

Obviously, designing a system that will be recognized asa success, both inside and outside the organization, is a

9

- --` - - ... .

formidable challenge. Few, if any, companies have fullyaccomplishedthis task to date.

2. Human Resource Management SystemAs mentioned above, significant adjustment is required

for an organization to coalesce behind the implementation ofnew factory automation and CIM technology. If the newtechnology is not installed in a greenfield site, thenlayoffs are often one consequence of the change. Reductionsin force are inevitably associated with morale problems forthe remaining employees who may view the layoffs as a sign ofcorporate retreat rather than revitalization.

Furthermore, human resource problems are not typicallylimited to simply laying off a set number of people and thenjust moving forward with the remaining group. CIM andautomation technologies place significantly greater skilldemands on the organization. Retraining and continuouseducation must be the rule for firms that hope to becompetitive with these technologies; the firm must undergo acultural transformation.

Requirements for retraining and continuous education areat least as strong for managers and engineers who work withthese new technologies as for the factory workers on theplant floor. Designing automated factories, managingautomated factories, and designing products for automatedfactories all require supplemental knowledge and skillscompared with those required for a traditional, labor-intensive plant. Senior managers, who must evaluate CIMtechnologies, as well as the people who work with them, alsocan benefit significantly from education about thetechnologies.

3. Product Development SystemFactory automation and CIM can make product designers'

jobs more difficult. Human-driven production systems areinfinitely more adaptable than automated manufacturingsystems. When designers are setting requirements for amanually built product, they can afford some sloppiness inthe specifications, knowing that the human assemblers caneither accomodate unexpected machining or assembly problemsas they occur, or at least can recognize problems andcommunicate them back to the designers for redesign.

In an automated setting, designers cannot rely on themanufacturing system to easily discover and recover fromdesign errors. There are severe limits to the levels ofintelligence and adaptability that can be designed intoautomated manufacturing systems, so product designers musthave either intimate knowledge of the manufacturing system orintimate communication with those who do. Developing such adesign capability in the organization is a difficult, butnecessary step for achieving world-class implementation ofthe manufacturing system.

10

-��-�---��'�II�~-~"�~11���'"��""1�"�-"

4. Managing Dynamic Process ImprovementIn most well-run, labor-intensive manufacturing systems,

continuous improvement results from a highly motivatedworkforce that constantly strives to discover better methodsfor performing its work. In a highly automated factory,there are few workers to observe, test, experiment with,think about, and learn about the system and how to make itbetter. As a consequence, some observers claim that factoryautomation will mean the end of the learning curve as animportant factor in manufacturing competitiveness. Such anassertion runs counter to a very long history of progress inindustrial productivity, resulting from a collection ofradical technological innovations, each followed by anextensive series of incremental improvements that helpperfect the new technology. Most students of the subjectestimate that the accumlation of such incrementalimprovements accounts for as much total productivity growthas do the radical innovations. In essence, any radicalinnovation may be thought of as a first pass innovation whichrequires much more innovation before it reaches its maximumpotential.

To presume that factory automation and CIM will reversethis historic pattern is premature at best, and potentiallyvery misleading to managers and implementers of thesetechnologies. Because these technologies are so new aiid socomplex, one cannot expect to capture all of the relevantknowledge at the system design stage. If a firm assumes thatonce it is in place, the technology will not be subject tovery much improvement, it will evaluate, design, and managethe system much differently than if it assumes that muchbenefit can be achieved by learning more about how best touse and improve the system once it is in place. One mightexpect to observe self-fulfilling prophecies in this regard.Even though an automated factory has far fewer people(potential innovators) in it, firms who invest in thistechnology would be wise to assure that those people who represent are trained to discover, capture, and apply as muchnew knowledge as possible. In fact, discovering andexploiting opportunities for continuous improvement might bethe primary reasons for firms to avoid completely unattendedfactories.

5. Technology Procurement

Before evaluating a specific technological option, thatoption must be reasonably well defined. A firm needs tochoose equipment and software vendors, and to decide how muchof the design, production, installation, and integration ofthe technology will be performed with in-house staff. Manyobservers argue for doing as much technology development inhouse as possible, to minimize information leaks about thefirm's process technology, and to assure a proper fit between

11

the firm's new technology and its existing strategy, people,and capital assets.

For external technology acquisition, technologicaloptions must be generated before they can be evaluated. Indeveloping these options, a firm must consider its currentassets, environment, and market position, as well as those ofits competitors. Equipment vendors must be brought into thedecision process. Vendor and technology evaluation criteriamust be developed and utilized within the organization.

6. System Control and Performance EvaluationOnce a technology investment choice has been

implemented, managers typically want to track the efficacy ofthat investment. The shortcomings of the traditional methodsfor measuring manufacturing performance are widelyrecognized. Many of these methods can be manipulated to makecurrent results look good at the expense of potential futureresults. When managers spend only a small fraction of theircareers in one facility or position, they often have anincentive to engage in such manipulations. In addition, inmany settings, the appropriate performance yardstick for afacility requires information on one or more competitors'facilities, on which timely, accurate data may beunavailable.

Increasingly, firms are using multidimensional measuresof manufacturing performance. Rather than depending on justa profitability summary statistic, measures of quality, leadtimes, cost of quality, delivery performance, and totalfactor productivity are being utilized to evaluateperformance. Despite this trend, firms could benefit frommore research on how, for example, to set standards forproductivity and learning rates in a highly automated,integrated environment.

ECONOMIC EVALUATION FOR NEW TECHNOLOGY ADOPTION

The technology adoption costs that are the most visibleand easiest to estimate in advance are the up-front capitaloutlays for purchased hardware, software, and services. Mostmodels consider only these costs. Also important, however,are (1) costs of laying off people whose skills will not beused in the new system, (2) costs of plant disruption causedby the introduction of new technology into an operatingfacility, and (3) costs of developing the human resourcesrequired to design, build, manage, maintain, and operate thenew system.

The benefits that flow from investment in factoryautomation and CIM are both tactical and strategic. Thesebenefits relate to changes in a firm's cost structure,increased process repeatability and product conformance,lower inventories, increased flexibility, and shorter flowand communication lines.

12

With respect to cost structure, investment in CIM andfactory automation tends to represent a large up-front costthat leads to a reduction in variable costs per unit ofoutput. This characteristic results primarily fromreplacement of labor by machines. Low variable costs canprovide significant competitive advantage when interfirmrivalry is high. In addition, reduced variable costssometimes lead firms to cut prices, potentially increasingmarket share and revenues.

The advantages arising from the increased repeatabilityand product conformance afforded by CIM and factoryautomation can also have significant competitive impact.Decreased process variability reduces scrap and rework costs,a source of variable cost savings that can be as important asthe reduction of direct labor costs by automation. Inaddition, improved product conformance can providesignificant sales gains in product markets.

Secondary effects of improved process control includeimproved morale (and consequent reduced absenteeism andturnover) of employees happy to work in a system that runswell.

Inventory reduction following automation and integrationinvestments can originate from several sources. First,factory automation can reduce setup times for some types ofoperations, reducing the need for cycle stocks. Second,decreased process variability can decrease uncertainty in theentire manufacturing system, reducing the need for safetystocks. Third, factory integration can shorten manufacturingcycle times, reducing the in-process inventories flowingthrough the system.

Manufacturing flexibility is another key strategicadvantage offered by CIM and factory automation. Rapid tooland equipment changeovers enable firms to quickly changeproduct mix in response to varying market demands. Inaddition, NC programming and computerized process planningshorten the time to market and time to volume for newproducts introduced into the factory. Fully-automatedmanufacturing systems provide volume flexibility as well.The highly-automated Matsushita VCR factory mentioned earliercan change its output rate with relatively low adjustmentcosts by increasing or decreasing the number of hours it runseach month. Because the factory's direct labor force isquite small, output declines will not lead to dramaticunderemployment, and increases do not require major hiringand training efforts.

Finally, reduced lead times between work stations willlower the flow times of work between stations, thusdecreasing the need for WIP in the system. As inventoriesand lead times are reduced, firms may increase their profitmargins by charging more for rapid delivery or may increasemarket share by offering better service and holding pricesconstant.

13

SUMMARY AND CONCLUSIONS

Increased global competition and environmentalvolatility require that firms adapt quickly or face thepossibility of extinction. Investment in automation andintegration, including hardware such as automated machinesand flexible manufacturing systems, software such as CIMsystems, and managerial approaches such as just-in-time anddesign for manufacturability, can help firms to achieve andmaintain competitiveness.

Of course, the assets always in shortest supply aremanagerial vision and leadership. Manufacturing strategycreation must precede technology investment decisions,because good technology rarely saves poor management.Therefore, firms must complement their learning abouttechnology options with information and insights about theirbusiness challenges and opportunities.

Charles H. Fine teaches operations management andmanufacturing strategy at MIT's Sloan School of Management.He holds Ph.d and Master's degrees from Stanford University,and a Bachelor's degree from Duke University. He hasauthored a number of articles on manufacturing issues,including pieces on the economics of quality improvement andon models for investment in flexible manufacturing systems.His industrial consulting, executive teaching, and researchproject experience includes work at Digital EquipmentCorporation, Eastman Kodak, General Electric, IBM, andMotorola.

Parts of this paper have been excerpted, with permission,from the author's longer paper, "Developments inManufacturing Technology and Economic Evaluation Models," tobe published in Logistics of Production and Inventory, S.CGraves, A. Rinnooy Kan, and P. Zipkin (eds.) in the NorthHolland Series of Handbooks in Operations Research andManagement Science. The author gratefully acknowledgeshelpful comments from Paul Adler, Steve Eppinger, SteveGraves, Patricia Moody, Karl Ulrich, and Larry Wein.

14

_�1�1�_ __________l__rrmX__��_a__r_*I �pm_�__�/\illl___�gI�Ulis�JFI_�LFI�Pn�