No. 98-06

Sustaining New Coordination Methods:
 The Case of World Class Manufacturing



 
 
 
 

Jonathan Allen
Yannis Bakos
Rob Kling
School of Management
Leonard N. Stern School of Business
SLIS
Purdue University
New York University
Indiana University
West Lafayette, IN
New York, NY10012-1126
Bloomington, IN 47405
Phone:765-494-2188
Phone: 212- 998-0841
Phone: 812-855-9763
Fax:
Fax:212-202-4130
Fax: 812-856-6166
E-mail: allenj@mgmt.purdue.edu 
E-mail: bakos@stern.nyu.edu
E-mail: kling@indiana.edu

September, 1998

Center for Social Informatics
SLIS
Indiana University
Bloomington, IN 47405


Abstract

A popular philosophy of manufacturing reform commonly referred to as "World Class Manufacturing", calls for the adoption of organizational practices that significantly alter coordination within and between manufacturing firms.  These practices are intended to enable continuous improvement, speed up response time, improve product quality, and create closer relationships with customers and suppliers.

The adoption of coordination methods advocated by World Class Manufacturing, such as cross-functional teams and vendor certification, has been uneven.  This is not surprising, considering the diversity of technical and institutional demands faced by different manufacturing firms, and even by different groups within firms.  But the literature on World Class Manufacturing coordination reforms has yet to describe the key contextual factors and processes which make these methods more or less sustainable for different organizations, continuing to prescribe the same solutions to every firm regardless of their situation or history.

In this study, we turn to organizational theory to help explain the sustainability or abandonment of different World Class Manufacturing coordination methods.  Using a case study of three different coordination reforms in a Southern California aerospace firm, we compared the explanatory power of two popular theoretical perspectives on organizational coordination and action -- one rational perspective (structural contingency theory) and one natural perspective (institutional theory). One of the abandoned coordination  reforms depended upon a complex computerized information system.

Our study indicates that the use of organizational theories adds substantially to our explanation and understanding of the practical barriers faced by World Class Manufacturing coordination practices.  Since few of these coordination innovations are justified using traditional techniques, it is especially important to have rich conceptual tools for thinking through exactly which coordination innovations will be sustainable in particular manufacturing organizations.  The combination of rational and natural systems models accounts for both the technical and the (often neglected) institutional dimensions of the manufacturing environment.
 



1.0. Introduction:  The Coordination Reforms of World Class Manufacturing

A  philosophy of manufacturing activity, commonly referred to as "World Class Manufacturing", has become increasingly popular in the United States over the past decade (e.g., Schonberger, 1986, Steudel and Desruelle, 1991).  What is most remarkable about the World Class Manufacturing (WCM) philosophy is its emphasis on organizational innovations and techniques, rather than the traditional focus on manufacturing competitiveness that dominated the post-World War II era -- issues such as better automation technology (Diebold, 1952), more sophisticated computerization (Harrington, 1973), and product R&D.  WCM has reoriented the conventional explanation of Japanese manufacturing success from more automation and robots to the reduction of "wasteful" (non-value-added) activities, including administrative overhead. While WCM advocates have often been critical of complex computerized manufacturing information systems, there have been recent efforts to show how such systems can be compatible with WCM techniques (Pun, Chin and Wong, 1998; Weng, 1998).  This article examines the sustainability  both computerized and non-computerized coordination strategies with a new way of conceptualizing coordination problems and data from a case study to anchor the analyses.

The WCM philosophy argues for a set of organizational innovations designed to improve coordination practices within and between firms. Different authors emphasize different features of WCM , but common elements include:

Just-in-time-inventory control - tightens coordination between the manufacturing firm and its suppliers.
Vendor certification/partnerships - tightens coordination between the manufacturing firm and its suppliers, typically resulting in a smaller number of suppliers.
Cellular manufacturing - tightens coordination within individual production areas, but reduces the need for coordination between production areas.
Employee involvement - tightens horizontal coordination, while reducing the need for vertical coordination.
Cross-functional teams - tightens coordination laterally across functional areas.
Concentration on core competencies - reduces coordination at many levels within the firm, but tightens coordination between purchasing and suppliers.

A core idea of the WCM philosophy that ties these practices together is the notion of simplicity in production.  By reducing coordination demands through more self-contained, focused, and flexible production units with fewer non-value-adding activities, WCM firms are able to increase coordination in their value-adding production activities, rather than the traditionally defined functional specialties.

The poor coordination of production activities, caused by an overly rigid division of labor and bureaucratic inertia, has become a leading explanation of poor US manufacturing performance in the 1980's (Dertouzos et al, 1990; Hayes et al, 1988; Harmon and Peterson, 1991).  WCM innovations are intended to eliminate the consequences of poor coordination around production, including unbuildable product designs, adversarial relations with suppliers and customers, overly complex shop floor layouts, and slow response times.

Despite the widespread enthusiasm for the WCM philosophy, and evidence of dramatic performance improvements in some manufacturing firms that adopted its innovations (Schonberger, 1986; Harmon and Peterson, 1991), implementation has been uneven and sometimes problematic (e.g., Plossl, 1988; Zammuto and O'Connor, 1992).  Uneven adoption of major organizational innovations like WCM is not surprising, considering the wide array of technical and institutional demands faced by different manufacturing firms.  The professional literature rarely contextualizes the principles of WCM. WCM strategies are usually recommended for all manufacturing firms, regardless of their situation or history.

WCM practices presuppose a number of major changes in the way manufacturing work is coordinated that may be more or less feasible in the organizational contexts of specific manufacturing firms.  For example, WCM prescribes closer working relationships between high status design engineers and lower status operations workers, more open relationships with suppliers based on trust, and the ability to group parts together in related families that can be manufactured in more self-contained clusters of equipment (manufacturing cells).  Because the decision to adopt WCM practices often cannot be justified on the basis of traditional criteria such as return on investment, it is especially important to develop an understanding of the conditions under which these new coordination mechanisms are viable and likely to be sustained, and when they are likely to be unsustainable and abandoned.

We turn to organizational theory to study under what conditions the WCM coordination reforms are likely to be sustainable, and when are they likely to be abandoned.  Using a case study of three different WCM coordination reforms in a Southern California aerospace firm, we compare the explanatory power of two different theoretical perspectives on organizational coordination and action -- a rational perspective (structural contingency theory) and a natural perspective (institutional theory).  Our analysis suggests that the use of organizational theories improves our ability to understand and explain the practical barriers faced by WCM coordination techniques.  By including both rational and natural systems models of organizations to examine the technical and institutional demands created by a manufacturing firm's environment, we can more clearly explain the key contextual conditions that manufacturing firms should not ignore in their decision to adopt WCM coordination reforms.

2.0 Using Organizational Theory:  Rational and Natural Systems Models

Can organizational theories help us explain important differences in the extent to which firms can sustain WCM coordination methods? We have selected two organizational theories that have been used to explain coordination within organizations:  structural contingency theory (Thompson, 1967; Galbraith, 1977; Van de Ven et al, 1976), and institutional theory (Powell and DiMaggio, 1991; Meyer and Rowan, 1977) (Table 1).
 

Theoretical Perspective 
Coordination Question
 Type of Model (Scott, 1992)
Structural Contingency Theory 
What are the coordination structures with sufficient information processing ability to accomplish tasks at a reasonable cost?
 Open/Rational
Institutional Theory 
How do organizations maintain legitimacy and obtain resources by conforming to coordination structures considered legitimate by the outside environment? 
Open/Natural
 
Table 1:  Two Organizational Theoretical Perspectives on Coordination Practices
In particular, we examine both rational and natural systems models of organizations (Scott, 1992).  Rational and natural systems models, according to Scott, are the major competing paradigms in current organizational theory.  Rational systems models focus on the pursuit of specific organizational goals through relatively formalized structures.  Natural systems models, in contrast, consider the common interest of members in the continuing survival of the organization the most important aspect of organizational action.  Rational and natural systems models differ in the way they ask why organizations coordinate the way they do.  Rational systems models stress the ability of coordination structures to meet the performance demands of formally defined tasks or goals. In contrast, natural systems models focus on the survival value of using (or appearing to use) coordination structures for different groups and coalitions.  We have selected open systems forms of each model -- they each include the influence of an extraorganizational environmental context, because of the broad consensus within organizational theory that the environment is important.

The following summaries briefly describe these themes.  (They do not capture all of the rich insights that each of these theoretical perspectives bring to the study of organizations.)  For analytic purposes, we will treat theory as homogenous wholes and focus on the basic question they ask about coordination within organizations, as well as the key environmental factors they identify.

2.1  Structural Contingency Theory

Structural contingency theory views organizations as bureaucracies designed to complete identifiable tasks.  The structure of any particular bureaucracy is determined largely by the uncertainty involved in their formal tasks (Galbraith, 1977).  The greater the task uncertainty, the greater the amount of information processing required.  Each kind of coordination structure, from standardized rules to cross-functional teams, can cope with a different level of uncertainty at a specific cost to the organization.

Structural contingency theory is a common model within organizational theory for studying coordination (e.g., Van de Ven et al, 1976; Ensign, 1998).  Though it is often treated as a single whole, there are two important variants of coordination explanations within structural contingency theory that are strongly related.  Galbraith's (1973; 1977) view of structural contingency theory explains the different coordination mechanisms available to organizations as their information processing needs (or uncertainty) increase.  When uncertainty increases to the point where traditional coordination methods used in bureaucracies -- including rules, hierarchy, and goal setting -- become overloaded, the organization has two sets of choices.  The organization can reduce the need for information processing (by creating slack resources or more self-contained tasks). Alternatively, the organization can increase its information processing capacity (by investing in vertical information systems or by creating new lateral relations).  The theory does not try to predict which choice an organization will make, but it does claim that the default choice is creating slack resources through reduced performance requirements (Galbraith, 1973; Ensign, 1998).
A second variant of structural contingency theory considers the different kinds of interdependence demanded by task activities (Thompson, 1967).  In general, the greater the interdependence, the more resources that must be devoted to coordination.  This variant of structural contingency theory identifies lists of coordination mechanisms (most commonly, four) that are appropriate for increasing levels of task interdependence:  standards, plans and schedules, mutual adjustment, and teams (Van de Ven et al, 1976; Adler, 1993).  Scott (1992) considers uncertainty, interdependence, and complexity as the three underlying dimensions of task requirements used by the structural contingency theory tradition.

2.2  Institutional Theory

Institutional theory views organizations as overlapping sets of practices, procedures, meanings, and explanations that are to varying degrees persistent, taken for granted beliefs about how organizations should coordinate and be structured.  Institutional theory first arose as a direct challenge to rational systems models such as structural contingency theory, which assumed that formal coordination structures were adopted by organizations because they were functional and rational (Meyer and Rowan, 1977).  Institutional theorists see the need for maintaining legitimacy in the eyes of powerful external groups as the main explanation for the adoption of formal coordination structures.  By maintaining legitimacy, organizations are able to obtain the resources and cooperation they need for continued survival.
According to institutional theory, organizational practices such as coordination methods are institutionalized (or accepted as objective and real by organizational actors) to match the legitimacy expectations of the perceived environment.  Coordination methods are diffused more because of social expectations than because of rational or technical properties (Meyer and Rowan, 1977).  Coordination methods are adopted through three different mechanisms (DiMaggio and Powell, 1991):  coercive isomorphism (forced by a powerful external actor, such as the government), mimetic isomorphism (copying a perceived leader in the face of high uncertainty and few other alternatives), and normative isomorphism (considered good practice by outside professional and/or educational associations).  Organizational activities are internally decoupled from institutionalized coordination methods if they conflict with local efficiency criteria (Meyer and Rowan, 1977).

2.3  Technical and Institutional Environments

Structural contingency theory and institutional theory identify different sources of environmental pressure to adopt coordination methods -- efficiency vs. conformity.  Structural contingency theory highlights the technical dimension of environmental demands, and institutional theory the institutional dimension as identified by Scott and Meyer (1991) (Figure 1).
 
 

 
 
Institutional
Environments
Technical        Environments
Stronger
Stronger
Weaker
 
Weaker
utilities, banks
general manufacturing
 
 
schools, churches
restaurants, health clubs
 
 Figure 1:  Technical and Institutional Environments

All organizations face technical and institutional demands from their environment, though in varying degrees.  In technical environments, organizations are rewarded for effective and efficient control of their production systems as their products or services are exchanged in a market (Scott and Meyer, 1991).  In institutional environments, organizations must conform to an elaborate set of rules and requirements if they are to receive support and legitimacy.  Institutional requirements may come from regulatory agencies, professional or trade associations, or from general belief systems held by society.  Scott and Meyer predict that organizations in strong technical environments will succeed to the extent that they "develop efficient production activities and effective coordinative structures", while organizations in strong institutional environments will succeed to the extent that they "acquire types of personnel and develop structural arrangements and production processes that conform to the specifications of established norms and/or authorities" (Scott and Meyer, 1991, p. 125).

Manufacturing firms have traditionally been associated with strong technical and weak institutional environments.  In this situation, Scott and Meyer would predict that focusing on the technical dimensions of coordination would be adequate for explaining coordination activity within manufacturing.  However, we have observed that the institutional environment of  certain manufacturing industries is stronger than is generally suspected.  Government regulations, powerful professional and trade organizations, and strong normative beliefs about appropriate manufacturing practices play an important role in manufacturing firms.  Institutional environments are especially strong, as Scott and Meyer recognize, in manufacturing firms that deal with public safety issues, such as health care and aerospace.
Because we are investigating an aerospace manufacturer with a strong technical and institutional environment in this study, we would expect that both rational and natural systems models of organizations would be needed to explain the sustainability of different coordination methods.  When either technical or institutional demands dominate, it may be easier to identify the relevant issues.  But what if both the technical and the institutional demands are strong?  Scott and Meyer will only predict in this situation that organizations will "develop more complex and more elaborate administrative systems and will experience high levels of internal conflict."  (Scott and Meyer, 1991, p. 126).  We will expand upon and clarify this generic prediction with data about the ways that one manufacturing firm adopted (and rejected) various coordination methods.

3.0  Study Design

Our study examines how well two different organizational theories, selected from both the rational and natural systems models of organizations, explain the sustainability or abandonment of three separate World Class Manufacturing coordination reforms at one case study site.  Using multiple theoretical perspectives allows us to examine the descriptive and explanatory adequacy of these different frameworks.  From this theoretical examination, we can contribute to explaining the larger patterns and key contextual conditions that facilitate or inhibit the workability of  coordination methods advocated by WCM enthusiasts.
In our case study of a one division of a large Southern California aerospace manufacturer, we tracked the  adoption and organizational implementation of three separate WCM coordination reforms over an 18-month period beginning in the summer of 1991.  The division employed about 1,200 people in two sites and had annual sales of about $150M.  In the early 1990s, the firm was losing sales because of increased foreign competition and  cutbacks in military aircraft purchases the first time since the "boom period" in the mid 1980's.  The division had recently laid off 700 workers, and its managers and staff were seeking ways to be more competitive under new market conditions. The firms managers were enthusiastic about three new coordination practices that they initiated around 1991, as means of reducing manufacturing costs and/or improving the quality of the firm's products:

a.  Lead time reduction -- the amount of time it takes to produce an already designed part from order to delivery -- Mangers tried to tighten schedules through a complex computerized Material Requirements Planning (MRPII) system . The effort was abandoned during our study and its leading advocates were either reassigned or fired.

b.  Focusing on core competencies by regrouping equipment to manufacture complex items into manufacturing cells and subcontracting for the purchase of simpler components. This reforms was partially sustained during our study, insofar as many simple items were purchased from outside suppliers and some concrete core competency objectives had been met. But the manufacturing cells were still not in routine operation.

c. Cross-functional teaming, was not only sustained during our study, but had spread to many other areas of the company.
Originally, we had no a-priori reasons for seeing any of these reforms in coordination practices as more or less viable for WCD. We collected our data over an 18-month period in three temporal waves of interviews with individuals and teams.  These 22 interviews each lasted 60-120 minutes and followed semi-structured protocols. Our informants ranged from machinists and shop floor supervisors to diverse managers and technical professionals to senior plant managers.
In a study of coordination, it is especially important to obtain coverage of as many different functional areas as possible.  We interviewed people who work in 14 different functional areas during the course of our study.  Purchasing, information systems, and configuration management personnel were interviewed once.  Capacity planning, contracts, design engineering, engineering technical support, manufacturing engineering, operations management, production control, program management, quality assurance, shop floor supervisors, and shop floor workers were interviewed multiple times.  The complexity of the coordination interactions between different functional areas makes it very difficult for the study design to be anything other than focused case studies, if the analyst hopes to include the perspectives of all relevant functional areas.

Our first wave of interviews was oriented towards improving our general understanding of the manufacturing environment. In the later waves we focused on collecting data to help examine the two organizational theories.  To learn about cross-functional teaming reform, we observed actual team meetings.  We collected key historical memos, pamphlets, policy and procedure manuals, and computer reports associated with the coordination reforms whenever possible, and used them as discussion points in our interviews.  Tape recording interviews was impractical in a defense-oriented firm where visitors cannot easily carry tape-recorders in and out of the organization, or record conversations without elaborate approval. We insisted on having two or more team members present at every interview or observation session to "triangulate" our notes and observations immediately after each session.

4.0  WCD/AIRTECH Background and Coordination Changes

In this section, we present some contextual information about AIRTECH, and briefly describe the key developments in each of the three attempts to adopt WCM coordination reforms.

4.1 WCD/AIRTECH -- Context and History

We conducted our study at the Wing Control Division (WCD) of AIRTECH, a Southern California aerospace manufacturer.  WCD produces sophisticated control equipment for airplanes, helicopters, and missiles that requires the integration of mechanical, hydraulic, and electronic technologies.  WCD's 10-12 product lines were evenly split between commercial and military markets.  WCD has a reputation for high-tech design skill and high prices.  As one design engineer reported, "we'll win [the contract] on technology if the price doesn't kick us out."

WCD copes with a highly complex design and manufacturing environment.  They manufacture equipment that can controls the motion of an airplane's flaps and tail fins. These control units are built from a solid block of aluminum. The aluminum is machined to include a complex three-dimensional network  of pathways through which fluid flows control the movements of mechanical arms that are attached to the block. A typical controller assembly includes about 500 parts and is machined to a few 10,000-inch tolerances in a some places.

The three dimensional routing problems WCD's staff confront in the design of hydraulic controllers tests the limits of IBM's mainframe-based CAD package.  Strict manufacturing requirements led another designer to remark, "life is not straightforward at 1/10,000th of an inch tolerances".  In addition to the task complexity, the institutional requirements of complying with outside regulation on both the military and commercial side add to the complexity of performing and documenting daily activities.

Many studies of manufacturing firms focus on assembly lines, where sufficiently high production volumes enable a firm to dedicate specialized equipment and workers to a relatively fixed operating sequence.  Partially-built assemblies generally flow down a line where equipment is already set up for the next operation. In contrast, WCD's manufacturing operation is organized as a job shop, where small batches of partially customized parts are manufactured by travelling a unique path through the machines and testing stations. Each batch of partially manufactured goods is accompanied by descriptions of the operations to be performed at a given work station and a description of its routing through the plant. A person who walks through a mass production assembly line can readily assess the state of production by counting the number of subassemblies in various stages of manufacture and seeing the assembly line's velocity. In contrast, it is much more difficult to assess the status of  any specific product in a job shop; work is being done at various workstations, and boxes of subassemblies for a specific product can be distributed at different workstations throughout the plant.

Our case study at WCD began in 1991, a time of upheaval in the US aerospace industry.  WCD was acquired by AIRTECH in the late 1970s.  The 1980s were a "boom period" in US aerospace, and WCD grew quickly to almost 2000 employees by 1990.  In the process, WCD changed from a smaller, "family style" company started by an aviation pioneer to a more formal organization in a stricter corporate environment.  By 1990, the "boom" was over -- the industry contracted, and for the first time foreign competition became a serious concern in WCD's major markets.  Civilian customers were able to quote quality (rate of rejection) figures from Japanese aerospace manufacturers that were significantly better than WCD's.  In competitive bidding for contracts, cost became much more of a concern, and WCD's high overhead costs became a major issue. In 1991, WCD's upper management, under strong pressure from AIRTECH headquarters, began a number of initiatives to adopt "World Class Manufacturing" changes that would improve coordination between activities in WCD, speed up their response time, and hopefully return them to high returns on net assets.  At the same time WCD had three rounds of layoffs in 1990 and 1991, reducing their workforce from around 1900 to 1200.

4.2  Lead Time Reduction through Material Requirements Planning (Not Sustained)

The coordination reforms of World Class Manufacturing stress the strategic benefits of reducing the amount of time it takes to design and produce products (e.g., Schonberger, 1986).  The production lead times -- the amount of time it takes to produce an already designed part from order to delivery -- of WCD's sophisticated aerospace components was too long in the view of AIRTECH's management.  In some cases, lead times of up to 24-30 months were causing scheduling delays, and increased foreign competition had AIRTECH's customers "screaming" for shorter lead times.  Shipment delays were impacting revenue flows, sending inventory "through the roof", and keeping production work in a constant "crisis mode".  Heavy pressure came down from AIRTECH on WCD's operations management to reduce lead times.

Reducing production lead times for small batch job-shop production is largely a matter of coordination and scheduling, rather than improving the physical production process.  Typically, 90-95% of the time parts spend on the factory floor in a job shop is idle time, waiting for an operation to finish on another machine (e.g., Susman and Chase, 1986).  The major tool available to WCD's operations management for shop floor coordination and scheduling is an Material Requirements Planning (MRPII) information system.  Given a desired "master schedule" of how many final products should be produced and when, the MRPII system uses information about the materials and subassemblies needed to produce a final product, how long it takes to make or buy each of the necessary materials and subassemblies, and the sequence of manufacturing operations materials and subassemblies must go through to create a schedule of purchases and manufacturing operations.

The basic MRPII shop floor activity model assigns standard times to each work center, broken down into queue (waiting) time, setup (changeover) time, and actual run time.  WCD's operations management, who had jurisdiction over these standard scheduling times, were certified by, and active in, the American Production and Inventory Control Society (APICS).  In the 1970's, APICS played a major role in launching the computerized factory scheduling movement through MRP/MRPII, but more recently has become a major proponent of Japanese-inspired World Class Manufacturing techniques, including Just-In-Time (JIT) inventory control.  At the beginning of our study, WCD's Material Manager had just returned from an APICS sponsored JIT seminar.  One of the major JIT metaphors he described to us was the analogy between excess inventory and other organizational slack that is used to mask fundamental problems in the factory, and the water level in a lake.  By reducing the water level (i.e., the amount of slack), the rocks (i.e. wasteful practices) would be exposed and the problems could be dealt with.  While APICS stresses that the factory schedule must always be "realistic" (e.g., Vollmann et al, 1993), cutting slack in the schedule and excess inventory is not only desirable, but has become one of the main measures of operations success.

The Director of Operations was convinced that he knew the main source of WCD's problems -- inadequate quality.  In his view, operations management had launched an array of quality related improvements in the factory, including increasing the visibility of quality problems, creating new computer reports to report scrap and rework, vendor certification programs, removing a layer of shop floor management, manufacturing cells, employee involvement, and most importantly, the removal of "quality-related buffers" from the MRPII scheduling model.  Operations management substantially reduced the lead times built into the MRPII scheduling model by removing an automatic one-month delay (called a pad) had been added to the front of each order, and by reducing the queue (wait) times at each work center to near zero.  The Director of Operations and the Material Manager both reported how these efforts had resulted in dramatically reduced lead times -- from 24-30 months down to 14-16 months.

This solution to the problem of excessive lead times, however, caused significant new coordination problems in the eyes of the other functional areas.  Capacity planners, while not disagreeing with the goal of reduced lead times, considered the time standards too variable and poorly measured for such tight scheduling.  Purchasing agents objected to the new "arbitrary and unrealistic" standards that did not consider the time it takes for competitive bidding.  Quality assurance inspectors objected to the lack of time for inspections.  But the harshest comments came from the shop floor supervisors.  The reduction of slack in the production schedule caused the manufacturing priorities to move back and forth "like an accordion." As jobs became more behind schedule, the marketing department pressured the shop floor managers to give priority to specific contracts, regardless of the official schedule; in practice, a significant fraction of the scheduling became ad-hoc.  Workers in their view "didn't know what was going on" when a job they had worked overtime on one day became a low priority item the next. Some machinists became demoralized when they worked hard to machine part of a batch for Product A one day, and on the next morning were told to stop the job, set aside Product A, reconfigure their equipment, and start a machining a batch for Product B. One supervisor described their "playing" with the queue times in the MRPII model as like "an earthquake - very unsettling afterwards".  Other terms used to describe the actions of operations management included "distant", "visionary", "autocratic", "controlling", and "theoretical".

By the end of our study period, the crises on the production floor had significantly affected WCD's ability to deliver products on schedule, and in a timely manner.  When combined with bad times in the aerospace industry overall, the result for WCD's operations management was fatal.  The General Manager was reassigned, and both the Director of Operations and the Material Manager were replaced.  A significant fraction, but not all, of the slack was reintroduced in the MRPII model, along with a tentative commitment to consult more functional areas (including shop floor workers) before making major changes in lead times.

4.3 Core Competencies through Manufacturing Cells (Partially Sustained)

The main idea behind the core competencies coordination reform is to focus and build expertise in a few critical, high-value activities (e.g. Prahalad and Hamel, 1990).  Core competencies is a corporate strategy that is often implemented on the factory floor by creating distinct manufacturing cells for each family of products relating to a competency, and by outsourcing other parts that do not fall under the strategic competencies.  Manufacturing cells organizes factory machinery and workers in terms of product families, rather than functional specialties.  By creating more self-contained manufacturing cells, some coordination issues between operations can be resolved, resulting in dramatic reductions in scheduled lead times (e.g., Schonberger, 1986; Harmon and Peterson, 1991).  At the same time, the firm increases its reliance upon subcontracting, and creates new coordination demands between purchasing agents and its suppliers.

In the 1980s, WCD manufactured most of the parts for each controller. Controller's include some high-precision components, such as the housing with complex hydraulic channels and high-grade pistons that require machining to 10-thousandths of an inch. These items were WCD's specialty and typify the five major production activities in which the firm's managers believed that they held special competitive advantages. They hoped that manufacturing cells for each kind of activity (ie., a set of diverse equipment grouped to manufacture pistons for different aircraft controllers) would substantially improve their production efficiencies. However, other parts, such as brackets that could be effectively machined to hundredths of an inch, could be subcontracted to numerous machine shops in the region.
Just prior to the beginning of our study, AIRTECH encouraged WCD's upper management to begin core competency, manufacturing cells, and outsourcing initiatives.  These coordination reforms were seen by WCD's management as a means for reducing scrap and rework rates (by concentrating on specific skills), and as a way of dealing with the issue of WCD's ever-increasing overhead rates (by focusing upon high value-added activities, and subcontracting the machining of the simpler products).  Soon after our study began, WCD management cited impressive lead time reduction figures for one of their new cells -- from six to nine months down to three to six weeks.

By the end of our study, however, WCD had disorganized all of their manufacturing  cells.  Production controllers considered the main problem to be the low utilization (as traditionally measured) of expensive five-axis high-precision metalworking machines.  These expensive machines could only be justified financially within the company by being used almost continuously; but the manufacturing cell concept optimizes around the benefits of reduced lead time and skill building rather than the intensive utilization of a specific piece of equipment.

Surprisingly, the skilled shop floor workers objected to the manufacturing cells because they viewed the cells as less stimulating, a "dead end".  Running a variety of jobs on a single type of machine (the traditional functional layout) was viewed by some workers as more interesting than "just being an operator, sitting and watching the machines".  Manufacturing cells are also dependent on the ability of the manufacturer to group their parts into sufficiently similar product families that can be run through a similar sequence of operations.  WCD's diverse product lines did not easily lend themselves to such groupings, as indicated by the earlier failure of a Group Technology encoding project intended to create part families.

Though the manufacturing cells coordination reform was not sustained over the 18 month period, an increase in outsourcing was sustained.  The number of parts manufactured inside WCD was permanently reduced from around 2000 to 500, leaving manufacturing operations the time to concentrate their attention on fewer high-value parts of strategic value.  The increase in outsourcing raised new coordination issues that were only partially addressed, despite the success of the outsourcing.

Although operations managers recognized that the pressure on purchasing would increase, purchasing managers complained that they did not have an increased involvement in production decisions made at the monthly scheduling meetings.  Changes to the schedule now involved many more transactions outside of WCD, and the purchasing manager complained that he "doesn't have the resources to change the order every time someone hiccups".  Most of the purchasing department's resources have to be spent on maintaining government records, and preserving the ability to link specific purchases (and their costs) to the specific controllers in while they are used.  Purchasing agents also complained that the rest of WCD saw them as merely "ordering out of a catalog", while they believed that Federal regulations and purchasing from foreign countries made their jobs complex and difficult.  For instance, all machined parts bought from outside vendors still had to be 100% inspected on arrival, and competitive bids are required for all parts that affect the direct cost of the product.  Fifty to one hundred Federal regulations might apply to any one purchase.  These requirements add time and complexity to outside transactions.

Outsourcing is supposed to take place with regards to maintaining core competencies, but at WCD there are still pressures to outsource core competency parts because of overloading in the factory (after three rounds of layoffs, the factory was running well over 100% capacity), and because the overhead costs tend to be higher.  Design engineers were particularly critical of WCD's managers for making outsourcing decisions on the basis of cost reduction without considering the added complexity and loss of control caused by performing the work outside of WCD.  Design engineers spent a significant fraction of their time guiding complex outsourced parts through the process, adding the "seasoning" that isn't completely captured in the formal blueprints.  Engineers felt that they had trouble dropping a supplier for poor technical performance, because purchasing agents would defend the supplier because they meet purchasing department's price, delivery, and documentary criteria for being a good supplier.

By the end of our study, the core competency coordination reform had been partially sustained.  Manufacturing cells were judged as impractical, and died a quiet death.  But as the study finished, WCD's project managers were convinced that they could make cells work in a new project because reduced demand in the industry has left WCD with excess machining capacity.  Outsourcing increased significantly, though occasionally not for the right reasons, but created new coordination problems within WCD that have yet to be addressed.

4.4. Cross-Functional Teams (Sustained)

Cross-functional teaming is a coordination reform advocated by World Class Manufacturing to improve and speed up interactions between functional areas that, while crucially dependent on each other, have become organizationally distant (e.g., Steudel and Desruelle, 1991; Schonberger, 1986).  During the 1980's, WCD had grown from a "Mom and Pop" independent into a large division of an industrial conglomerate.  Design engineers, referred to WCD transforming from a "you help me, I'll help you" atmosphere into "sterile bureaucracy."  Formal communication lines were becoming increasingly clogged ---- one engineer reported a "moderate six month backlog" on formal information requests.

At the same time, WCD was experiencing cost overruns on their large fixed price military contracts.  Project managers attributed the cost overruns to a lack of coordination up front that led to expensive redesign activity.  Because of the frequency of cross-functional interactions on one large military project, the first cross-functional team was created by the employees and project managers themselves.

By the beginning of our study, WCD's upper managers declared in a series of memos that all project activity was to be reorganized around "multi-disciplinary, collocated, empowered, ad hoc teams".  A close reading of the memos showed that, in addition to the call for greater employee involvement and empowerment, there was a strong push towards greater cost controls --"the ceremonial approach to approving cost overruns must stop!" they declared.  This observation was supported by project managers, who saw WCD's main objective for cross-functional teams as getting everyone involved in cost containment.  This was especially needed because AIRTECH's computerized financial accounting system could not provide each project with real time cost information, particularly in the early design and testing stages, and there was little chance of the huge financial system being substantially modified in the next couple of years.

Over the next 18 months, the staff in eight other product lines created cross-functional teams.  Our team observations and interviews suggest that enthusiasm for teaming remained high during our study.  The teams facilitated coordination in various ways. For example, tracking the status of parts in meetings was simplified with people from different groups co-present. Everyone from shop floor workers to design engineers were able to learn about the progress of key parts through production, as well as problems, such as supply bottlenecks or machining failures.    MRPII reports were clarified (how broken is this part?), advanced warning could be given for "hot" jobs, and contract questions about what WCD would have to pay for vs. the customer could be answered immediately.  In addition, the teams seemed to be rapidly discuss diverse options and make immediate decisions when conditions changed. Proposals to customers for new products were also created in a much shorter time.  One project manager's explanation for teaming's success where other change programs have failed is because of the "immediate results" it brings for flexibility, responsiveness, and cost containment.  He lauded the "spirit of camaraderie" encouraged by teaming as the different "social strata" "let their hair down" and worked together.

Developing cross-functional teams is not simple, since they require people in diverse occupational specialties and statuses to develop trust and mutual understanding.  Cross-functional teams also raised new coordination issues within WCD.  Team members still belonged to their established functional areas, and performance was still rewarded primarily by their functional area.  These career and compensation issues were only being gradually dealt with by WCD, but they eventually planned to base rewards primarily on team performance .   Another area of heightened coordination need was between the project teams and the shop floor areas which, because of the failure of the manufacturing cells initiative, were still organized into functional areas.  Resolving priority conflicts between different project teams became a delicate negotiation question.

The benefits of increased lateral coordination could not benefit functional areas that were not included on the teams.  In our observations of the early project teams, for example, we could not help but notice that the most complained about, joked about, and generally abused functional area (in this case, purchasing) was one of the few areas that did not have a representative on the team.  As our study came to a close, WCD made initial attempts to reorganize purchasing agents as members of project teams. Cross-functional teaming was the one coordination reform advocated by World Class Manufacturing enthusiasts that was sustained, and grew, during our study.

5.  Explaining the Sustainability of WCM Coordination Reforms

Why were some World Class Manufacturing coordination reforms sustained at Wing Control Division, and others abandoned?  In this section, we investigate how well each of our organizational theories answers this question.

Though specific coordination reforms are sometimes described as changing the amount or quality of coordination (Malone and Crowston, 1994), we view them as coordination transformations.  Any coordination reform transforms one set of coordination problems into another, that may be more or less tractable for the organization.  For example, the lead time reduction through MRPII improved scheduling coordination in the short run, but created new coordination problems for prioritizing daily shop floor activities.  Increased outsourcing relieved coordination problems in production, but increased the coordination demands faced by purchasing agents.  And the move to cross-functional teams increased coordination between different production specialists, but caused new coordination problems between team members and their functional areas.  For a coordination reform to be sustainable, the organization must be able to cope with the new set of coordination issues that are created.

The concept of coordination transformations implies that a coordination reform can fail even if the reform itself appears to be doing well, because the coordination reform might create another coordination difficulty in some other part of the organization that is less tractable.  The two organizational theories used below focus on different aspects of the coordination issues that remain after the transformation.

5.1. Structural Contingency Theory

For both variants of structural contingency theory, coordination reforms are unlikely to be sustained when the coordination method does not have the information processing capacity to handle the uncertainty of the primary technical task, or if the coordination method is excessively costly to the organization given their information processing needs.

5.1.1. Structural Contingency Theory and Lead Time Reduction

WCD's operations managers changed the scheduling model in the MRPII system to reduce the scheduled lead time for producing parts in the factory.  In the language of structural contingency theory, this move reduces slack without a corresponding reduction in the uncertainty of the task.  Removing the slack created a need for more information processing capability to coordinate task operations, which the system was not able to provide.  Consequently, the coordination of shop floor tasks became more difficult, and task performance was reduced by default.

But why did the computerized MRPII system fail to supply the additional information processing capacity?  In Galbraith's scheme, vertical information systems are a major strategy for increasing an organization's information processing ability.  Galbraith (1973) discusses the coordination choices available to an aerospace company similar to WCD, where he assumes that a computerized scheduling system would lead to better coordination -- with the only downside being the cost of developing the system.  Unfortunately, he assumes that all vertical information systems are of roughly equal quality and appropriateness.  The value of the MRPII system for coordinating real factory activity, however, is highly dependent on a number of implementation choices, the most crucial being the time standards built into the system.  But the time standards are not only very difficult to create technically, but are an extremely sensitive political issue because they influence job assignments, work priorities, and rewards.

Structural contingency theory, particularly the Galbraith variant, usefully describes the internal technical failure of the lead time reduction coordination reform.  The content-free description of coordination methods is the basis for a clear and concise general theory of coordination, but also makes it difficult to apply to specific organizations without some evaluation of the quality or appropriateness of the specific implementation of the coordination method.  All vertical information systems, like all teams, are not of equal quality.

5.1.2. Structural Contingency Theory and Core Competencies

The attempt to create manufacturing cells can be seen as trying to make task coordination less costly by creating more self-contained tasks, and reducing interdependence.  But structural contingency theory requires the coordination method to be adequate for people to effective process information to carry out a  task (whether uncertainty or interdependence).  The failure of the cell-manufacturing ("group technology") project, that tried to arrange the parts into common families that could undergo similar manufacturing operations, strongly suggests that their was not sufficient commonality between parts to make manufacturing cells work.  Other organizational shifts, such as changes in product lines or changes in design, would be needed before manufacturing cells could be organized around families of parts, such as pistons.

At the same time, WCD did succeed in outsourcing many of their less crucial parts.  Although structural contingency theory does not deal with the make/buy question explicitly, the move to outsource can be seen as an attempt to redeploy scarce coordination resources around the most important technical tasks.  The fact that less complex parts were usually outsourced is compatible with structural contingency theory, since simpler tasks can be coordinated at a distance by relatively inexpensive rules and plans.  When WCD did outsource more complex parts, the coordination problems increased significantly.

5.1.3. Structural Contingency Theory and Cross-Functional Teams

Teams are a lateral coordination structure that increase the information processing capacity of an organization, according to Galbraith.  The "interdependence" variant of structural contingency theory considers teams as appropriate for the most highly interdependent, highly complex tasks that organizations cope with.
WCD operates in an extremely complex task environment.  Team-like structures arose informally at WCD, before the proclamation of upper managers, because a few managers found that they helped to cope with the complex communication patterns  of large aerospace projects.  Structural contingency theory would predict that teams (or, more precisely, well-implemented teams) would thrive in a highly uncertain task environment such as WCD, and they have.

In the same case study of an aerospace manufacturer as mentioned previously, Galbraith (1973) describes how the company chose to cope with their increased uncertainty by adopting scheduling teams, rather than new vertical information systems or reducing the need for information processing.  Teams may be a costly coordination method in terms of the total amount of staff time and attention, but at WCD the "up-front" cost of teams was very low compared to other coordination reform alternatives.  Teams did not require (at first) permanent rearrangements in the hierarchy, or expensive new technologies, and could be successfully organized without changing factors that lie outside of the organization's control.  The "immediate benefit" of teams for coordination was particularly important in the crisis-driven, scarce resources situation that manufacturer's often find themselves in before they decide to pursue massive coordination reforms.

5.1.4  Structural Contingency Theory Summary

The strength of structural contingency theory is in its ability to sharply characterize the internal technical performance barriers to World Class Manufacturing coordination reforms.  As a rational model of organizations, it concentrates on the organization's ability to fulfill its formal goals -- in this case, efficient product manufacturing.  Because Wing Control Division has a task environment with strong technical demands, the impact of task uncertainty and interdependence is significant and needs to be accounted for in any explanation of coordination reform sustainability.

The variants of structural contingency theory have a strong, clear characterization of the need for coordination in terms of task uncertainty, interdependence, and complexity.  The concepts used to describe the solutions to this need -- generic coordination methods, the information processing capacity of each method, and the costs of each method -- we find are still too general to guide decision-making in particular organizations.  It is difficult to evaluate, for example, the concrete information processing capacity of a particular MRPII implementation, or whether the costs of other coordination reforms to make manufacturing cells work would be worth the coordination gains.  Associating each coordination method with a single cost abstracts out important implementation details, and obscures the distribution of costs over time which may be an important determinant of sustainability.

5.2. Institutional Theory

According to institutional theory, coordination reforms are initially adopted out of a desire to conform to what powerful external actors, or society at large, consider as legitimate behavior.  The decision to adopt "World Class Manufacturing" reforms, for example, were a legitimate response, in the eyes of AIRTECH headquarters.  Coordination reforms are more likely to be sustained when the legitimacy they contribute helps the organization to receive the resources and cooperation it needs to survive.

5.2.1. Institutional Theory and Lead Time Reduction

Why did the attempted reduction in lead time take the form that it did?  Institutional theory would explain the MRPII modifications as an example of normative isomorphism.  An influential professional organization, in which the key operations managers people were active, advocated a combination of managers through technology and the latest Just-In-Time reform metaphors (i.e., reducing the water level in the lake, getting rid of "fat").  Immediate, dramatic reductions in lead time were stressed in this professional world as the primary measure of manufacturing success.  WCD's operations managers drew upon these institutionalized beliefs to cope with their coordination and scheduling problems.

In particular, the Director of Operations used the term, "quality," as a slogan to resolve the problems that he saw in WCD's operations.  His strong belief allowed him to remove all "quality-related" slack time in the MRPII model as he simultaneously tackled quality problems with a barrage of reform initiatives.  This attempt to reduce lead times transformed the set of coordination problems faced by WCD, making shop floor, quality assurance, and purchasing work more difficult to stabilize.  The reform attempt convinced other functional areas that operations managers was too "distant" and "theoretical".  The centralized MRPII scheduling system allowed operations managers to enact their institutionalized beliefs unilaterally, and on a massive scale -- enforcing a kind of virtual reality that did not allow routine work to be decoupled from ceremonial demands in the way that institutional theory predicts.

Institutional theory provides a strong explanation for why the lead time coordination reform took place the way it did.  The ultimate abandonment of this coordination reform, however, was strongly dependent on the technical problems that resulted from the coordination transformation. Therefore it lies outside the strict view of institutional theory.  The strongest institutional explanation for the abandonment would be that the normative legitimacy (in the eyes of the professional association) gained by adopting these practices in no way compensated for the loss of task efficiency.  In other words, is it worth changing a supplier to please APICS, only to upset the U.S. Government or a major customer, such as an airplane  manufacturing firm?

5.2.2. Institutional Theory and Core Competencies

Coordination reforms such as manufacturing cells are culturally accepted as legitimate in the manufacturing world today, and were encouraged by the executives at AIRTECH as a solution to WCD's internal and environmental problems.  The idea of manufacturing cells comes bundled with a set of beliefs about greater organizational performance and upskilled workers.  Cells were abandoned, however, as the low machine utilization rate could not be justified (i.e., made legitimate) to the financial accountants who controlled resource flows.  Using machine utilization as a key performance measure is an institutionalized belief that is being increasingly discredited in manufacturing (Schonberger, 1986), but is still an important part of the financial accounting world.  WCD's installed base of expensive machinery, justified financially on the basis of high utilization, could not be easily adapted to the needs of cells.  In addition to the measurement problem, the belief in upskilled workers was not considered legitimate by a significant fraction of the floor workers, who saw cell work as less attractive than functional specialization.

The outsourcing coordination reform was more sustainable, though significant institutionalized barriers prevented the outsourcing from taking the form most advocated by World Class Manufacturing.  Heavy government regulations prohibited WCD from forming close relationships with suppliers, because of the requirements for competitive bids and 100% inspection of incoming parts.  Both design engineers and purchasing complained that the coordination problems that remained after the transformation to outsourcing made their work more difficult.  But the legitimacy gain in the eyes of upper managers, and AIRTECH, from reduced overhead costs in WCD made the coordination transformation sustainable regardless of the technical wisdom of the move.

5.2.3. Institutional Theory and Cross-Functional Teams

Why were teams a sustainable coordination reform?  Institutional theory would first ask, why were teams chosen at all?  Though teams are the culturally-preferred innovation of choice for empowering workers and decentralizing decision-making, WCD appeared to be primarily interested in teams to improve cost controls in their large aerospace projects.  Teams, though not usually touted as the optimum cost control method (indeed, structural contingency theory describes teams as the most costly method of coordination), were a legitimate innovation available to WCD upper managers.

The coordination reforms brought by teams quickly spread within WCD.  Teams matched the task completion needs of the projects without needing substantial changes in the rest of the organization, or the outside environment.  The coordination problems that remained between the product teams and the functionally oriented shop floor units were not as significant as the gains from close, up-front coordination between team members.  Though important functional areas were left out of the early product teams because of institutionalized beliefs, such as purchasing, within the 18 month period WCD had already moved to add purchasing employees to the teams.

Conformance to legitimate practice explains the use of teams as a primary cost-control method within WCD (rather than new accounting controls).  And the legitimacy of cross-functional teaming, combined with a significant local technical need, contributed to breaking down the barriers between established functional areas without lengthy up-front negotiations.

5.2.4  Institutional Theory Summary

The strength of institutional theory is in explaining the behavioral process of deciding which coordination reform will be pursued, and how it is pursued.  Institutional theory is also useful for explaining situations in which powerful legitimizing forces in the outside environment, such as government regulators, accounting practices, or professional organizations, have a large impact.

Explaining the sustainability of coordination reforms along both technical and institutional dimensions is more effective than looking at the technical attributes of the task alone in this case.  We would predict the same for any manufacturing organization that faces both strong technical and institutional demands from its environment.  We would find it very difficult, for example, to explain coordination in complex manufacturing solely on the basis of institutional theory.  In contrast to the rational structural contingency theory view, institutional theory offers a clear portrait of the solutions to the coordination needs (culturally legitimate structures), but is vague on describing the coordination needs themselves.  How can we differentiate between stronger and weaker institutional environments?  How can we compare the legitimacy needs of an organization with regards to different government, professional, and societal actors?  These concepts need to be clarified before institutional theory could provide a coherent, if necessarily incomplete, account of coordination in complex manufacturing organizations.

6. Discussion and Conclusions

6.1  Practical Implications

In our case study of an aerospace manufacturer attempting to change the way it coordinates, we found that organizational theories do help identify the conditions that make World Class Manufacturing coordination reforms sustainable.  If a manufacturing firm faces strong technical demands, a rational model of organizations such as structural contingency theory can help predict when the coordination structure will be sufficient to handle the uncertainty, complexity, and interdependence of the task.  If a manufacturing firm faces strong institutional demands, such as government regulations, professional associations, or strong societal beliefs, a natural model of organizations such as institutional theory can help predict whether a coordination reform will satisfy a manufacturing firm's need for legitimacy.

Every World Class Manufacturing coordination reform we studied in this case had to be adapted to the particular needs of WCD.  Despite the universal claims of World Class Manufacturing advocates, manufacturing cells may not be technically feasible for a given task configuration, or close supplier relations may face institutionalized barriers that have to be surmounted -- or may not be worth surmounting.  Our results suggest that a careful examination of the firm's technical and institutional environment is required before attempting elaborate coordination reforms.  As practical analysis tools, the organizational theories we use provide a starting point -- a guide for thought -- rather than a methodology.  But we suspect that many similarly complex manufacturing firms, facing strong technical and institutional demands, need a structured way of thinking through these problems.

We cannot emphasize enough that the institutional dimension of manufacturing environments must not continue to be neglected, especially in industries with strong regulatory oversight such as health care and aerospace.  When government regulations, professional associations, or societal beliefs are critical to a manufacturing firm's environment, they should not be ignored simply because everyone knows manufacturing is mostly technical.  We see no shortage of evangelistic manufacturing philosophies -- SPC, TQM, Deming, World Class Manufacturing, etc. -- but we do see a shortage in the tools of analysis available to a company deciding to pursue one of these reform movements.  If new coordination methods are not selected exclusively on the basis of task efficiency, and we see evidence in complex manufacturing environments that they are not, then they cannot be reasoned about exclusively in terms of task efficiency.

We also introduced the idea of coordination transformations, because we found evidence that a coordination reform transforms the entire set of coordination activities performed by a manufacturing firm.  A coordination reform may not be sustainable because it creates new, difficult coordination problems elsewhere.  This point should not be forgotten when evaluating the success or failure of an isolated reform.

6.2  Theoretical Implications

Our findings suggest that both rational and natural systems models of organizations are necessary for explaining organizational outcomes in strong technical and institutional environments such as aerospace manufacturing. Rational systems models address key internal technical barriers that must be overcome, but natural systems models are able to cope with key external demands on the manufacturing firm's legality and legitimacy, and better explain the cognitive processes through which powerful coalitions in firms decide to launch various coordination reforms.

Organizational theories that can cope with both technical and institutional dimensions are particularly important when action along one dimension affects another.  For example, the institutionalized professional belief system which led WCD operations managers to change the MRPII scheduling model led to severe task accomplishment problems of a largely technical nature.  Attempting to improve task coordination performance through manufacturing cells, on the other hand, led to a clash in legitimate beliefs of how to evaluate manufacturing performance (machine utilization vs. other measures).  These barriers to sustainability would be poorly explained from the viewpoint of either single theory.

The two particular rational and natural systems theories of organizations that we chose for this study, however, represent only one of many possible choices.  They complement each other, in the sense that the rational systems theory had a strong conceptual view of the need for coordination (task uncertainty, complexity, and interdependence) and a relatively weak conceptual view of proposed coordination methods (generic types with roughly defined information processing capacities and costs). In contrast, the natural systems theory had a relatively weak accounting of the need for coordination (institutional demands from many sources) but a strong conceptual view of the coordination methods themselves (packages of legitimate cultural beliefs).  But other theories could be substituted, or added, depending on the research question.  The key point is that institutional demands from the environment suggest including natural systems theories of organizations, while technical demands suggest including rational systems  theories.

Our study confirmed Scott and Meyer's (1991) hypothesis that firms with both strong technical and strong institutional environmental demands do create complex administrative structures, and do experience high levels of internal conflict in some areas.    But can we say more?  Successful manufacturing organizations with strong technical and institutional demands seemingly have to do the impossible:  create technically efficient and effective coordination structures, and develop structures that conform to established norms of legitimacy.  One solution manufacturing companies can turn to is decoupling -- separating formal structures from actual shop floor practice (Zetka, 1991; Meyer and Rowan, 1977).  In the WCD case, decoupling was very possible for the cross-functional teams (since the formal structure was very loosely defined anyway). However, decoupling was almost impossible in the case of the centralized computerized MRPII scheduling system, which tightly linked shop floor performance to the normative scheduling model.  In short, the computerized communication system was much less flexible and effectively unmanagable in this case. Our approach to viewing coordination technologies, such as MRP, in terms of the ease or difficulty of managing the resulting coordination problem is very different (and stronger than) a generic critical success factors approach (see, for example, Sum, Ang, and Yeo, 1997).

Another strategy is to optimize the coordination transformations so that the most attention and resources are devoted to the most crucial sources of resources and cooperation in the environment, whether they be from the marketplace or from key societal sectors, and that the remaining poorly coordinated activities are relatively unimportant.  A third possibility is to lower the barriers to experimentation, so that new coordination methods can be quickly evaluated and either spread, be discarded, or maintained formally.  The sustainability of cross-functional teaming was helped substantially by its low up-front costs to the organization.  These three explanations should be further evaluated in future work.

We also have shown the value of studying the dynamic, time-sensitive aspect of coordination, especially in times of organizational upheaval.  Rather than studying fixed, on-going operations (e.g., Van de Ven et al, 1976; Crowston, 1997), coordination studies should be sensitive to the shifting nature of the task over time (Adler, 1993) and to the maturity of the coordination method within the organization.  New coordination reforms solve some existing coordination problems, but also create new ones.  This process is difficult to study using the traditional static view of coordination mechanisms.

In summary, we have shown that organizational theories can help  explain when organizational reforms that are expected to improve coordination are sustained.  While the computerized reform (MRP) was not sustained in this case, we do not argue that non-computerized coordination reforms are automatically easier to sustain. In the case of WCD, the shift to core competencies was only partially sustained as well.  Many manufacturers, especially those with important public safety components such as aerospace, face strong institutional as well as technical demands from their environments.  This requires that natural systems models of organizations, such as institutional theory, are needed as well as rational systems models to adequately explain why advanced manufacturers coordinate activities the way they do.  The concepts of structural contingency theory and institutional theory help us practically and theoretically understand the likelihood of a specific coordination reform being sustained in a particular organization.

ACKNOWLEDGEMENTS:
The research for this paper was partially supported by NSF grant IRI 9015 497.  We also benefitted significantly from our discussions about coordination early in the project with our colleague Professor John King. Dr. Werner Beuschel helped conduct the field work that supports this paper.

----------------------
ENDNOTES:

References

Adler, Paul S.  (1995). "Interdepartmental Interdependence and Coordination: The Case of the Design/Manufacturing Interface." Organization Science. 6(2):147.

Crowston, Kevin. 1997. A coordination theory approach to organizational process design.  Organization Science 8(2) (Mar/Apr):157?175.
Dertouzos, Michael, Lester, R., and R. Solow. (1989).  Made in America. New York:  Harper Perennial.

Diebold, John. (1952).  Automation: The Advent of the Automated Factory. New York:  Van Nostrand.

DiMaggio, Paul J., and Walter W. Powell. (1991). "The Iron Cage Revisited:  Institutional Isomorphism and Collective Rationality in Organizational Fields."  In Powell and DiMaggio.

Ensign, Prescott C. 1998. Interdependence, coordination, and structure in complex organizations: Implications for organization design. Mid?Atlantic Journal of Business 34(1) (March):5?22.

Galbraith, Jay. (1973). Designing Complex Organizations. Reading, MA: Addison-Wesley.

Galbraith, Jay. (1977). Organizational Design. Reading, MA: Addison-Wesley.

Harmon, Roy D., and Leroy D. Peterson. (1990). Reinventing the Factory:  Productivity Breakthroughs in Manufacturing Today. New York: The Free Press.

Harrington, Joseph  (1973). Computer Integrated Manufacturing. Huntington, NY:  Krieger.

Hayes, Robert H., Wheelwright, Steven C., Kim B. Clark. (1988). Dynamic Manufacturing: Creating the Learning Organization. New York: The Free Press.

Huber, Vandra L., and Karen A. Brown. (1991). "Human Resource Issues in Cellular Manufacturing: A Sociotechnical Analysis."  Journal of Operations Management. 10(1)(January):138-159.

Karlsson, Christer and Par Ahlstrom. (1995). "Change processes towards lean production: The role of the remuneration system."  International Journal of Operations & Production Management 15(11):80?99.

Kling, Rob and Tom Jewett. (1994). "The Social Design of Worklife With Computers and Networks: An Open Natural Systems Perspective." in Advances in Computers. Volume 39. Marshall C. Yovits (ed.) San Diego: Academic Press.

Kling, Rob, Kraemer, Kenneth L., Allen, Jonathan, Bakos, Yannis, Gurbaxani, Vijay, and John King. (1992). "Information Systems in Manufacturing Coordination:  Economic and Social Perspectives." in Proceedings of the Thirteenth International Conference on Information Systems. Dallas, TX.

Kling, Rob Kenneth L. Kraemer, Jonathan P. Allen, Yannis Bakos, Vijay Gurbaxani and Margaret Elliott. (in press)  Transforming Coordination: The Promise and Problems of Information Technology in .  Tom Malone, Gary Olson, and John Smith (Eds).  Coordination. in: Coordination Theory and Collaboration Technology Mahwah, NJ: Lawrence Erlbaum.

Malone, T. W. & K. Crowston. (1994). "The interdisciplinary study of coordination."  ACM Computing Surveys. 26(1):87-119.

Meyer, John W., and Brian Rowan (1977). "Institutionalized Organizations:  Formal Structure as Myth and Ceremony." American Journal of Sociology. vol. 83: 340-363.

Motwani, Jaideep, Ashok Kumar and Yunus Kathawala. (1994). "WCM practices of North American manufacturing organizations."  Industrial Management & Data Systems. 94(7):18?23.

Plossl, George W. (1991). Managing in the New World of Manufacturing: How Companies Can Improve Operations to Compete Globally.  Englewood Cliffs, NJ:  Prentice Hall.

Powell, Walter W., and Paul J. DiMaggio. (eds.) (1991). The New Institutionalism in Organizational Analysis. Chicago: University of Chicago Press.

Pralahad, C.K., and Gary Hamel. (1990). "The Core Competence of the Corporation."  Harvard Business Review. 68(3):79-93.

Pun, Kit?Fai,  Kwai?Sang Chin and  K H. Wong. 1998.  Implementing JIT/MRP in a PCB manufacturer.  Production & Inventory Management Journal 39(1):10?16.

Rastogi, P.N. (1995). Management of Technology and Innovation: Competing Through Technological Excellance. London: Sage Publications.

Schonberger, Richard J. (1986). World Class Manufacturing: The Lessons of Simplicity Applied. New York:  Free Press.

Scott, W. Richard. (1992). Organizations: Rational, Natural, and Open Systems. (3rd edition). Englewood Cliffs, NJ: Prentice Hall.

Scott, W. Richard, and John W. Meyer. (1991). "The Organization of Societal Sectors." In Powell and DiMaggio.

Steudel, Harold J., and Paul Desruelle. (1992).  Manufacturing in the Nineties:  How to Become a Mean, Lean, World-Class Competitor.  New York: Van Nostrand Reinhold.

Sum, Chee, Chuong,  James S K Ang and Lei?Noy Yeo. (1997). Contextual elements of critical success factors in MRP implementation.  Production & Inventory Management Journal 38(3):77?83.

Thompson, James D. (1967). Organizations in Action: Social Science Bases of Administrative Theory. New York: McGraw-Hill.

Van de Ven, Andrew H., Andre L. Delbecq, and Richard Koenig, Jr. (1976).  "Determinants of Coordination Modes within Organizations."  American Sociological Review. vol. 41: 322-338.

Vollman, Thomas E., Berry, William L., and D. Clay Whybark. (1993).  Integrated Production and Inventory Management: Revitalizing the Manufacturing Enterprise.  Homewood, IL: Business One Irwin.

Weng, Z Kevin. 1998.  Tailored just?in?time and MRP systems in carpet manufacturing.  Production & Inventory Management Journal 39(1):46?50.

Zammuto, Raymond M., and Edward J. O'Connor. (1992). "Gaining Advanced Manufacturing Technologies' Benefits: The Roles of Organizational Design and Culture." Academy of Management Review. 17(4):701?728.

Zetka, James R., Jr. (1991). "Automated Technologies, Institutional Environments, and Skilled Labor Processes: Toward an Institutional Theory of Automation Outcomes." The Sociological Quarterly. 32(4):557