Masatoshi Kano
Ram Duvvuru Sriram
Amar Gupta
WP #3778 August 1994
PROFIT #94-17
Productivity From Information Technology
"PROFIT" Research Initiative
Sloan School of Management
Massachusetts Institute of Technology
Cambridge, MA 02139 USA
(617)253-8584
Fax: (617)258-7579
Copyright Massachusetts Institute of Technology 1994. The research described herein has been supported (in whole or in part) by the Productivity From Information
Technology (PROFIT) Research Initiative at MIT. This copy is for the exclusive use of
PROFIT sponsor firms.

The Productivity From Information Technology (PROFIT) Initiative was established on October 23, 1992 by MIT President Charles Vest and Provost Mark Wrighton "to study the use of information technology in both the private and public sectors and to enhance productivity in areas ranging from finance to transportation, and from manufacturing to telecommunications." At the time of its inception, PROFIT took over the Composite Information Systems Laboratory and Handwritten Character
Recognition Laboratory. These two laboratories are now involved in research related to context mediation and imaging respectively.
LABORATORY
ELECTRONICS
CENTER FOR
\_CNA
COORDINANTION
SCIENCES .
RSAC)
TCambridge, MA 02ATIO142-1247
R Ioom

Masatoshi Kano,
2
Ram Duvvuru Sriram,
3 and Amar Gupta
4
ABSTRACT-- Concurrent/collaborative engineering (CE) often requires the collaborative participation of several companies. Management of such multi-company collaboration in engineering becomes more difficult when companies from different countries are involved, as is the trend in the Japanese engineering and construction industry (E&C). This paper focuses on the conditions required for successful management of such CE. The authors first propose a conceptual time- and cost-based model of multi-company CE work as a framework. Then using this framework, current practices of Japanese E&C firms are analyzed. Major findings are that (1) international multi-company CE should be designed with careful consideration, not only to task spilt and allocation but also to inter-firm task dependencies; and (2) Japanese E&C firms need to alter their inter-firm coordination scheme significantly in order to derive full benefits in the global marketplace.
1-1 Motivation
In plant engineering and construction (E&C) projects, tasks are divided into several sub-tasks such as plant design, component design, and construction. The sub-
'Support of Masatoshi Kano for this study was provided by the JGC Corporation. Funding for the DICE project was provided by the MIT Intelligent Engineering Systems Laboratory affiliates program and
National Science Foundation PYI Award No. DDM-8957464, with matching grants from NT Data, and Digital Equipment Corporation.
2
JGC Corporation, 1-14-1 Bessho, Minami-ku, Yokohama, 232 Japan
3Principal Research Scientist, Intelligent Engineering Systems Lab., 1-253, Massachusetts Institute of
Technology, Cambridge, MA 02139
4Senior Research Scientist, Sloan School of Management, E53-3 11, Massachusetts Institute of
Technology, Cambridge, MA 02139

tasks are shared among joint venture partners, equipment manufacturers, and subcontractors. This multi-company task-sharing scheme for Japanese E&C firms has become increasingly internationalized during the last decade under fierce market competition for lower costs. Because engineering activities require close coordination and collaboration, especially under strong market demand for shorter project periods, efficient organization and management of international multi-company collaborative engineering work is critical to such E&C firms.
For faster and better-quality product development, concurrent/collaborative engineering (CE) has become a major focus of engineering management. Information technology (IT) is considered a key tool which enables effective CE. Responding to this situation, the second author and his group, in 1986, initiated research on a computer-based architecture program called DICE, an acronym for Distributed and Integrated
Environment for Computer-aided Engineering [1,2]. In addressing coordination and communication problems in engineering, key research issues of CE and IT application identified and addressed in the DICE project are as follows: (i) Framework for dealing with problem-solving architectures, (ii) Representation issues for the development of product models needed for communicating information across disciplines, (iii)
Organizational issues to be addressed in organizing engineering activities for effective use of computer-aided tools, (iv) Negotiation/constraint management techniques for conflict detection and resolution between various agents, (v) Transaction management issues relating to interaction between the agents and the central communication medium, such as long-duration transactions and concurrency control, (vi) Design methods to support various kinds of design activities of individual agents, such as hierarchical refinement and constraint propagation, (vii) Visualization techniques including user interfaces and physical modeling systems, (viii) Design rationale records of the justifications generated
2

during design and other engineering activities, (ix) Interfaces between agents to support information transfer between various agents, and (x) Communication protocols to facilitate the movement of objects between applications. Within this broad perspective of
CE and IT application, this paper concentrates on organizational issues (item (iii) above) in international multi-company settings.
Organizational issues of engineering projects across country and company borders have been examined by several researchers. For example, the organization for large-scale development projects has been studied by Sayles and Chandler [3], the organization of multi-national companies by Ghoshal & Westney [4] and Egelhoff [5], information technology in multi-national companies by Roche [6], international operations by Hollier, et al. [7] and organization and supports of design projects by Eppinger, et al. [8]
However, no study has yet addressed how best to organize and manage international multi-company CE projects.
1-2 Objectives and Organization of the Paper
This paper addresses four basic questions:
1) Is a cross-national distribution of engineering tasks advantageous to E&C firms?
2) How should engineering work be split?
3) How should tasks be allocated?
4) What aspects make such collaborative work efficient and effective, thus feasible?
3

To answer these questions, the motivating factors for and issues in international multi-company CE of Japanese E&C firms are discussed (Section 2), and a conceptual workload model is presented (Section 3). Using this framework, the current engineering processes, intra- and inter-firm engineering coordination work, and information systems of
Japanese E&C firms are analyzed (Section 4) to identify the issues in splitting and allocating the engineering tasks. The paper concludes with suggested measures and a future scenario for making international multi-company CE efficient and effective.
2-1 Drivers of International Multi-Company CE
E&C firms rely exclusively on awarded contracts, and are highly prone to fluctuations in workload. (See Fig. 1.) Because of difficulties in recruiting and laying off a competitive workforce that can respond to the changes in workload, Japanese E&C firms have focused instead on: (1) geographical market expansion, (2) product family expansion,
(3) arranging finance for clients, and (4) outsourcing work, including engineering and design. The last option constitutes a major reason for E&C firms to practice multicompany CE.
4

Fi2ure I WORLD MARKET FOR PETROCHEMICAL PLANTS IN THE 80S z
-J
-J
0
-i
4*
M EUROPE
· NORTH AMERICA
· SOUTH & CENTRAL AMERICA
O AFRICA
* MIDDLE EAST
X ASIA
1981 1982 1983 1984 1985 1986 1987
YEAR
1988 1989 1990
(Source: Engineering News Record [9])
Japanese E&C firms typically own neither manufacturing facilities nor assembly factories; they purchase equipment and construction material from outside suppliers.
Because equipment manufacturers (vendors) possess rapidly evolving equipment design/manufacturing technology and often share the engineering work of customizing their equipment for a particular project, E&C firms are motivated to enter into multicompany CE relationships with vendors.
On the other hand, the appreciation of the Japanese Yen initiated by G5's Plaza
Agreement of 1985 boosted Japan's engineering fees to the world's highest level in a short period (see Table 1). Also during the 1980s, worldwide investment in the oil refining and petrochemical industries became sluggish (Fig. 1), leading to fierce competition among
E&C firms worldwide.
5

Table I
ENGINEERING FEE
Fee
9
(Source: The Engineering Business [ 10])
Both these developments prompted Japanese E&C firms to increase their purchase of equipment and material from abroad, and to shift part of the work to developing countries using quality but less expensive engineers. The JGC Corporation, for example, established a new engineering subsidiary, Technoserve International Co., Inc., in The Philippines during 1989, and Toyo Engineering Corporation set up Techno Management Co. in
Korea, in 1988. In other words, the new and turbulent global market has served to catalyze internationalization of CE among E&C firms.
In addition, in order to foster industrialization at home, the public policy of developing countries often mandates foreign firms to work with local firms in plant construction projects. Such industrialization policy serves as yet one more motivator for international multi-company CE.
6

2-2 Issues in International Multi-Company CE of Japanese E&C Firms
After a shortened project period became critical for E&C firms, a task overlapping technique, or CE, was adapted to reduce the engineering period. Unlike sequential task execution which provides a complete set of information at the end of each task, a task overlapping technique requires bilateral or multi-lateral flow of information on an evolving basis, enabling a downstream task to start earlier, even when its upstream task is still underway [11]. Such flow and coordination of information becomes difficult and costly in an international scenario because of the greater distances involved. The difficulties in coordination across both company and country borders have occasionally resulted in design changes and rework at the construction site as well. Although engineering accounts for only nine to fifteen percent of total project costs [12], engineering work determines most of the overall project costs. So, each factor that impacts engineering quality, engineering time, or costs is critical to the overall competitiveness of E&C firms.
What is unique about international multi-company CE? The answer possibly lies in distances between companies. Although telecommunication and transportation technologies have altered traditional views about physical distance, there still exist distances in (i) cultures, (ii) time zones and seasons, (iii) meteorological conditions, (iv) technological expertise, and (v) communication/coordination modes, or methods and tools for communication. While these differences can provide new perspectives and approaches to an engineering problem and can lead engineers to more effective solutions, they nevertheless make communication and coordination difficult.
7

3-1 Engineering and Coordination Workload Model
In order to analyze the difficulties in communication and coordination encountered in international multi-company CE, we must review the characteristics of coordination.
When engineering work is broken down into tasks that involve several engineering disciplines, three kinds of interdependencies among different tasks emerge [13]: (i) pooled, where the activities share or produce common resources but are otherwise independent;
(ii) sequential, where the initiation of some activities depends on the prior completion of others; and (iii) reciprocal, where each activity requires input from the others.
Coordination offers a means to manage these task interdependencies, so the overhead involved in coordination between companies is determined primarily by the method used to divide the engineering work and to allocate the decomposed tasks among companies. That is, overhead is determined by the number of inter-firm task dependencies of the companies, and the level of difficulty in communication encountered when managing the task dependencies across company-borders. The number of inter-firm task dependencies can be approximated by the number of firms involved. Communication difficulties generally arise from encoding and decoding information (code in Miller's terminology [22]) exchanged by engineers of the task-dependent companies. Difficulties in encoding and decoding are determined by the communication tools available and by differences in language, culture and technological standards. The more limited the tools or the greater the differences, the more difficult the encoding and decoding operations become. In short, communication difficulties can be represented by distance that denotes, in this paper, differences in culture, language, and technologies, rather than for physical distance.
8

Thus, the coordination workload in multi-company CE can be modeled as a function of the number of firms involved and the distances between task-dependent companies. On the other hand, the effort involved in performing sub-divided engineering tasks is considered to be constant, because engineering tasks are based on scientific and technological knowledge and skills, and are relatively immune to differences in language, culture, and geography.
Based on these assumptions, the total workload F is deemed to comprise of a constant engineering workload K and a variable coordination workloadf as follows.
F(n,d) = K + f(n,d) (3-1-1) where n is the number of firms, and d is the distance between the task-dependent firms.
The functionf(n,d) is considered to increase in value as either n or d increases.
3-2 Productivity and its Effects on Multi-Company CE
If multi-company CE merely adds extra coordination workload, as in Eq. 3-1-1, why does one bother with multi-company collaboration? The key reason is that we can reduce total engineering time and costs despite an increase in total workload. To relate workload to engineering time and costs, let us consider the company's engineering productivity p, which is defined as a ratio of the total workload F(n,d) to required person-
hour; the person-hour is the product of the average number of engineers m and the total engineering time t in hours. Formally, p is given by:
P
F(n,d) xt (3-2-1)
9

In multi-company CE, the productivity of i-th firm is given as:
Pi =
F(n,d) m i x t i
(3-2-2) where F(n,d) = E F (n,d). By using a unit engineering fee per engineers' person-hour c, and productivity pi of i-th firm for the workload F, the total engineering cost C becomes:
Cl~(n~d)X Ci i Pi
Total engineering period T is:
(3-2-3)
T= (n,d) i Pi x
1 mi
- if all the firms' tasks are sequential, or (3-2-4a)
1 mi
) if all the firms' tasks are in parallel.
5
(3-2-4b)
Equations 3-2-3 through 3-2-4b highlight the importance of the term
F (n,d)
, which denotes the number of person-hours, in reducing engineering time and costs.
3-3 Task-Resource Fit and Coordination Fit: Strategic Dimensions
We will elaborate on the person-hour in this section. Person-hour
F(n,d) can be
re-written by using Eq. 3-1-1.
5
In case that firms have mixed task connections (both sequential and parallel), Twill have a value between Ts given by Eqs. 3-2-4a and 3-2-4b, depending on the scheme of the task connections.
10

F(n,d)
Pi
K i
-- +
Pi f(n,d)
Pi where K = E K i and ftn,d)= Z f(n,d).
i i
(3-3-1)
The first term of the right-hand side of Eq. 3-3-1 denotes the engineering personhours, and the second term the coordination person-hours. As engineering person-hours represent the level of skillfulness of engineers of a company with regard to engineering tasks assigned to them, we call it as task-resource fit. Similarly, the coordination personhour represents how smoothly task-dependencies are processed by a company, or the skillfulness used in performing coordination work, so we call it as coordination fit. Hence, the total cost C can be stated as follows:
C=
Pi xc + f(nXd)C .
Pi
(3-3-2)
First, we check the effect of task-resource fit
Pi
Assuming that two companies have the same engineering project, the same task breakdown, and can perform all brokendown tasks on an interchangeable basis, multi-company collaboration always gives better fit to the overall workload. Suppose that Company A and Company B have four disciplines (civil, machine, piping, and electrical engineering), and for each discipline, j, a project has engineering workload Ki, (30,15,35, and 20, respectively). If each company, i, has a set of productivity pi and the unit engineering fee c for each disciplinej as is shown in Fig. 2, the cost of the project is calculated as shown in the parentheses in Fig. 2. For machinery design, Company B has a cheaper unit engineering fee and involves less cost.
11

For piping design, Company A has better productivity with the same unit engineering fee, thus provides less cost. For civil design, Company B has a more expensive unit engineering fee but has much better productivity, resulting in less cost. For electrical design Company B has less productivity but much less unit engineering fee, yielding less cost. When two companies split the project tasks and work together (bottom one in the figure), the overall engineering cost becomes lower because we can find better productivity pij or lower unit engineering fee cj while total engineering workload K
(= E
Ki ) remains the same value.
Fiure 2 BENEFITS OF MULTI-COMPANY COLLABORATION
K
1
=30 K
2
=15 K
3
=35 K4=20
s
190 .7.-(30) l.
P : 83.-'.:,..-
( 42 (50<,s °--K Be'tt u}<A.:\ Its}, o62 dt
Note: ( ) cost
12

In this way, frther task decomposition and involvement of more companies, or a bigger n, will yield a better chance of reducing the overall engineering cost through (i) task-resource fit or (ii) involvement of cheaper unit engineering fee through multi-company collaboration. Similarly, greater cost reductions are likely when companies with greater d are involved, because of growing science and technology parity across borders and complementarity of national systems in science and technology [14]. (See Fig. 3.) Better engineering productivity, p, and cheaper engineering fee, c, constitute two motivations for multi-company collaboration to enhance overall engineering cost competitiveness.
Figure 3 ENGINEERING COST CURVE
j L
-EN
Note: The horizontal axis is number of firms n or distance d.
On the other hand, as coordination workload fi(n,d) increases with number of firms n or distance d, the overall coordination cost also increases with n or d even with better coordination productivity or lower unit coordination fee. (See Fig. 4.)
13

Fi2ure 4 COORDINATION COST CURVE
Mftkm.
Note: The horizontal axis is number of firms n or distance d.
There are, however, possible options to suppress the increase in coordination cost through better coordination fit f (n,d)
. Let us check two factors of coordination fit
P f (n,d)
Pi coordination workload fi(n,d) and coordination productivity pi.
6
The coordination workload is determined by inter-firm task dependencies related to engineering work process, or task sequence, and the method of splitting the tasks for each company. The examples of Figure 5 illustrate how task sequence and the method of splitting impact interfirm task dependencies.
6
Coordination productivity Pci is actually different from engineering productivity pi. This paper uses the notion of Pi in the sense of Pci for coordination, for simplicity's sake.
14

Figure 5
COMPANY A
INTER-FIRM TASK DEPENDENCIES
SE UENTIAL
COMPANY B
REIPC2
(0)
L 3 ID
(
4
RECIPROCAL (UNIDIRECTIONAL)
COMPANYB ..
5IL
_
%:
.,..:-+
M- .
~ ~
._
.
~ ~~~
I
I
RECIPROCAL WITH SYNCHRONIZATION (UNIDIRECTIONAL)
,Lo;; ,COMPANY B
':'- ·r; - sV.X.
U
,~·. t.
,
-
·
(1)~~~·
-1-~·:: t!
.
COMPANY A
·. 44_1 1~~~ i
I I
(1)
1
Note: ( )= number of inter-firm task dependencies, j = number of iteration cycles
15

Let us measure inter-firm task dependencies by the number of inputs required for each company to perform its tasks. A simple sequential task transfer from Company A to
Company B (top one in the figure) gives one inter-firm task dependency for Company B, while Company A will also get one inter-firm task dependency if task sequence transfers back to Company A on a reciprocal basis (second from the top). If Company A has an independent task to be done in parallel with tasks of Company B, and Company A also needs to consolidate the design results from both task sequences, each Company gets one inter-firm task dependency (second from the bottom). If Company A's and Company B's tasks need an iterative process between them, both companies get j inter-firm task dependencies, wherej represents the number of required iteration cycles (bottom one).
As we can see from these examples, inter-firm task dependencies can be reduced by: (i) avoiding inter-firm task split across the iteration process [8], (ii) avoiding task transfer reciprocity, or (iii) minimizing the number of firms to be involved. These are important rules for limiting the increase in coordination workload under multi-company collaboration.
Now let us examine coordination productivity, which is the company's proficiency in handling required coordination work. One way to improve it is to identify the inter-firm differences that cause encoding and decoding difficulties, and to devise countermeasures to improve encoding and decoding productivity by: (i) wide exchange of engineers, or (ii) transferring liaison engineers to other firms as technological gatekeepers [15].
To summarize, better coordination fit can be attained through three steps: (i) splitting and sharing engineering tasks so that the inter-firm task dependencies are minimized, (ii) identifying the difficulties in task coordination that collaborating companies have between them, and (iii) devising strategies for overcoming these difficulties.
16

3-4 Tradeoff between Engineering Cost and Coordination Cost
From the discussions thus far, when total cost for a project is considered under multi-company collaboration, a tradeoff exists between reduced engineering cost and increased coordination cost. In other words, an increase in the number of firms or distance among firms is advantageous so long as marginal coordination cost is less than marginal engineering cost. (See Fig. 6.) Taking the number of firms and their distances as variables, the point where the marginal coordination cost is equal to marginal engineering cost represents the ideal operating situation characterized by the lowest total cost (Point
A-a in the figure).
Figure 6 TRADEOFF BETWEEN ENGINEERING COST AND
COORDINATION COST
TOTAL COST
-...... ENGINEERING COST
.................. COORDINATION COST
I
r(n,d) = Vn '+ d2 n: number of firms d: distance between firms
17

Exactly the same argument holds in case of engineering time T. If the number of firms or the distance increases, engineering time decreases through better task-resource fit and higher probability of finding companies with a greater number of available engineers.
The tradeoff between engineering time and coordination time can determine the optimal number of firms and the distance.
In Figure 6, when coordination cost and marginal coordination cost are reduced, the minimum total cost will decrease by more than the amount of coordination cost reduction. This is achieved through engineering cost reduction enabled by the increase in number of firms or their distance (Point B in Fig. 7). We term this effect as leverage by
task-resource fit. Similarly, when engineering cost and marginal engineering cost are reduced, the minimum cost will decrease by more than the amount of engineering cost reduction through coordination cost reduction enabled by a decrease in number of firms or
The conceptual model described here suggests two strategic dimensions to improve performance through multi-company CE: (i) manage improvement of coordination productivity among collaborating companies, if the leverage by task-resource fit through increase in ii and d is great, or if the marginal coordination cost curve is steeper than the marginal engineering cost curve (from Point A to Point B in Fig. 7); and (ii) manage improvement of engineering productivity of collaborating companies, if the leverage by coordination fit through reducing n and d is great, or if the marginal engineering cost curve is steeper than the marginal coordination curve (from Point A to
Point C in Fig. 8). (See Fig. 9.)
18

Figure 7 TOTAL COST REDUCTION BY
COORDINATION IMPROVEMENT
$
$ I
\
Ii
4-
A B
TOTAL COST
-*-.-----*--- ENGINEERING COST
I- r(n,d)
.................. COORDINATION COST
COORDINATION
IMPROVEMENT
TOTAL COST
IMPROVEMENT
- -- -- - -- - - - - - - r(n,d) = n:+ d
Figure 8 TOTAL COST REDUCTION BY
ENGINEERING PRODUCTIVITY IMPROVEMENT
TOTAL COST
............. ENGINEERING COST
................ COORDINATION COST
PRODUCTIVITY
V IMPROVEMENT
TOTAL COST
IMPROVEMENT
C A r(n,d)
19

STRATEGIC FRAMEWORK FOR MULTI-COMPANY CE
IMPROVE
COORD.
PROD'TY
1
LEVERAGE BY
TASK-RESOURCE
FIT
4
CURRENT
POSITION
(A)
LEVERAGE BY
COORDINATION
FIT
(C)
IMPROVE
ENGINEERING
PRODUCTIVITY
4-1 Engineering Process
The engineering process of E&C firms can be divided into three main phases: (i) the basic engineering phase; (ii) the detailed engineering phase; and (iii) the vendor's/subcontractor's engineering phase. In the basic engineering phase, process
engineers specializing in a certain product group that requires product-specific expertise design the overall system to client specifications. In the detailed engineering phase, the engineering work is decomposed into several tasks to design different types of subsystems and components, e.g., steel structure, building, piping, compressor systems, boiler
20

systems, control systems, and substation. While project engineers provide coordination support, each component is engineered by component engineers grouped into functional departments mainly by categories of engineering disciplines or manufacturing industries.
The component engineers maintain technological expertise about components by specialization, creating smoother interfaces with vendors and subcontractors for effective engineering coordination. The component design is transferred to vendors, with each functional department retaining responsibilities for supervising the vendor's design of the components in both performance and ease of site installation and maintenance.
Document Approval Procedure
Although this sequential engineering process appears to be simple, the actual taskconnection is often not straightforward. Engineering tasks subcontracted to vendors or subcontractors are generally subject to a reciprocal and iterative approval procedure based on documents which enable the E&C firm to keep track of the responsibilities of each company. Moreover, regardless of inter-firm task dependencies, vendors and subcontractors are often subcontracted directly by the E&C firm so that the E&C firm can easily take control of the whole engineering process. In this way, even a simple task sequence flows through a complex, reciprocal and iterative path (see the Current Practice in Fig. 10), although the tasks need to be connected directly from engineering subcontractors to vendors or construction subcontractors (see the Ideal Task-Flow in Fig.
10).
One way to simplify the path, for example, is to subcontract tasks directly from project engineers to final vendors by improving the engineering capabilities of project engineers and vendors so that component engineering departments and engineering
21

subcontractors can be eliminated from the path (Possibility (1) in Fig. 10). A second method of simplification would be to establish an IT-based coordination platform among companies based on a long-term close relationship. This will enable once-through engineering processes through preliminary checks and earlier comments (Possibility (2) in
Fig. 10). The engineering process may remain sequential with these options, however the authors believe that this is the first step toward designing an effective CE framework in a multi-company and multi-country situation.
22

Figure 10 INTER-FIRM ENGINEERING TASK-FLOW
IDEAL TASK-FLOW
FOR
CURRENT PRACTICE i ..
POSSIBILITY (1)
x~~~~~~~~~
i
.
vZ,
aVENDOR
, ,,<A_
.
..
.
.
POSSIBILITY (2)
23

Rationale for Current Engineering Practice
The surge in new technology and the advent of a global economy motivate E&C firms to assign work to companies with the lowest quotations in the free market, even with no previous work relationship and on an individual project basis. Because the subcontractors' capabilities is not clear under such short-term company relationship, a typical E&C firm i likely to exercise greater control through centralized subcontracting and strict document approval procedures. So, a short-term company relationship based on free market competition creates extra coordination workload through an iterative process, and reciprocal task dependencies among collaborating firms. Although the competitive market remains attractive to reduce costs of E&C firms, it may be appropriate for these firms to evaluate the free market's efficiency, especially in terms of coordination cost.
One explanation for the emergence of multi-national companies (MNCs) is that
MNCs internalize international economic transaction costs previously curried out on the open market, and can make a profit through doing them more efficiently than the global market [16,17,18,19]. If this transaction-cost theory is correct, long-term partnership among companies may be more efficient than a one-time relation. This is because a longterm partnership will eliminate work relating to each subcontract and allow closer relations between engineering subcontractors and construction subcontractors/vendors, which will then enable them to streamline the task-flow between companies with less E&C firm intervention.
To attain efficiency even in a one-time inter-firm relationship, a global infrastructure is needed to support frequent exchange of the unstructured information required for engineering processes, especially in CE. While the current infrastructure is
24

sufficient for international transfer of funds, goods, people, and structured data, the problem of insufficient market infrastructure for CE cannot be mitigated by frequent faceto-face coordination in an international setting. To overcome this shortcoming, we recommend that companies establish long-term strategic alliances to substitute marketbased relations, and establish IT-based infrastructure to support smoother exchange of unstructured information for engineering.
Adherence to the document approval procedure and E&C firm-centered subcontracting imply that E&C firms are extending their Japan-based organization and operation to the global marketplace only by internationalizing their peripherals. While their project management activities have become international, the corporate management structure of E&C firms remains domestic. To manage effective international operations especially in engineering work, Japanese E&C firms need to depart from the old paradigm of home-country-centered corporate management, to a new paradigm of international alliance-based corporate management (see Section 4-4).
4-2 Engineering Coordination -- Project Engineer
In an E&C firm, project engineers play the pivotal role of intra- and inter-firm engineering coordination with responsibility for resolving inter-disciplinary engineering conflicts from various viewpoints of project management. At one specific E&C firm, project engineers accounted for 36 % of the total engineering resources (see Fig. 11); 37
% of the time of these project engineers was devoted to engineering work (see Fig. 12), of which three-quarters was spent on engineering coordination alone.
25

Fi2ure 11 HUMAN RESOURCE ALLOCATION IN A JAPANESE E&C FIRM
(Percentages are based on the number of employees)
R&D & DATA PROCESSING
ENGINEERS
9%
BASIC ENGINEERING
13%
PROJECT
36%
DETAIL ENGINEERING
25%
- GINEERS
.Y70
CONSTRUCTION
ENGINEERS
8%
PROCUREMENT &
INSPECTION
6%
(Adapted from Outline ofJGC [20])
In other words, project engineers spend nearly one-third of their total time in engineering coordination. From Fig. 11 and 12, roughly 12% of an E&C firm's total human resource is devoted to engineering coordination by project engineers, while 38% is allocated to basic and detailed engineering; i.e., engineering coordination consumes about one-quarter of the total engineering and coordination person-hours.
26

Figure 12 PROJECT ENGINEER'S WORK
(Percentages are based on used time.)
PROCUREMENT CONSTRUCTION
5% 5%
OJECT MANAGEMENT
37%
BASIC & DETAIL
FMUrIM:FRIM(G
PROPOSAL PREPARATION
16%
The intra- and inter-firm coordination work of project engineers has been handled mostly through meetings held at the E&C firm, and by telephone (see Fig. 13). Letters and facsimiles are as popular as such meetings and telephone calls for international projects, while business trips or meetings at another firm's office are much more common for domestic projects. The physical inter-firm distance involved in international projects makes personal meetings less effective in terms of both time and cost, and one relies instead on written communications, including drawings, which can reduce chances of miscommunication across country borders and have the added advantage of leaving a record for future reference. In any case, inter-firm engineering coordination by data communication remains negligibly small.
27

100%
Figure 13 COMMUNICATION MODES OF PROJECT ENGINEERS
80%
60%
40%
OBUSINESS TRIP
=LETTER & FAX mMEETING & TELEPHONE
20%
0%
INTL PJ DOMESTIC PJ
(Vertical axis is based on used time.)
If the current human resource allocations for project engineering and the project engineer's workload are appropriate for E&C firms, overall productivity of E&C firms cannot be effectively improved without first improving the productivity of the engineering and coordination work handled by project engineers. Their communication modes, especially, of letters and facsimiles, which require more time and effort for encoding and decoding than face-to-face communication, need improvement.
4-3 Information Systems
We now examine the state of computer-aided engineering (CAE) systems currently employed by Japanese E&C firms. A survey was conducted by Nippon Kikai Kogyo
Rengo Kai (Japan Machinery Federation) and Engineering Shinko Kyokai (Engineering
Advancement Association of Japan). Of 27 Japanese E&C firms that replied to a survey,
28

90 % use CAE systems for basic and detailed engineering work, 50% intend to improve engineering quality and productivity by introducing CAE systems, and 30% intend to reduce project periods and costs [21]. As shown in Figure 14, the two major problems with current systems are as follows: (i) historical data of completed projects was not yet installed in the CAE database, making it unusable for ongoing projects, and (ii) data was incompatible with other CAE systems. The first problem originated tfrom relatively recent installation of CAE systems and fast evolving CAE technology. The second is of a different origin: the CAE systems in an E&C firm have been developed department by department, and each department uses a unique CAE platform best suited to its particular engineering tasks; e.g., a piping design CAE for a piping engineering department. Data incompatibility and poor hardware inter-connectivity among different CAE platforms have resulted in inefficient repetition of data output and input via paper at each CAE platform associated with extra checking for input error. This problem is true also for inter-firm
CAE data transfer. About 60% of CAE data communication is done by paper and 20% by diskettes for both domestic and international projects [21].
29

Figure 14 PROBLEMS OF CURRENT CAE SYSTEMS
IN THE JAPANESE E&C INDUSTRY
NO CAD SYSTEM
TIN IE FOR INPUT
DIF FICULT INPUT
LIMITED USAGE FOR 3D MODEL
USER UNFRIENDLINESS
HARDWARE INCOMPATIBILITY
DATA ACCESSIBILITY
DUPLICATED INPUT
LIMITED USE OF STANDARD
DATA
INFLEXIBILITY
TIME FOR 3D INPUT
DATA INCOMPATIBILITY
COMPLETED JOB DATA
NOT ON CAE
0
OOTHERS
03
U2
X1
2 4 6 8 10 12 14 16
Notes: The number shown in the legend box denotes respondent's first, second, third, and other minor choice. The horizontal axis represents number of companies who replied.
(Source: [21])
In order to derive higher benefits, an engineering database (EDB) needs to be established, as highlighted in Figure 15. An EDB is the design database originating from process engineers and transferred to and evolved by component engineers; it is considered to be a new engineering framework to replace the current engineering document, such as plot plan and P&ID. It eliminates the need for data input and output repetition among different computer systems, including the CAE systems of each engineering department.
30

Figure 15 ISSUES TO GET MOST OUT OF A CAE SYSTEM IN
THE JAPANESE E&C INDUSTRY
ORGANIZATIONAL CHANGE
HARDWARE/SOFTWARE
COMPATIBILITY
MANAGEMENT COMMITMENT
FUNCTIONALITY
HARDWARE/SOFTWARE COST
INTEGRATION OF Al, NEURO-
COMPUTER, &
EWS PROMOTION
PROMOTION OF
COMPUTER NETWORK
DATA INTEGRATION
TRAINING
ENGG PROCESS
REENGINEERING
HISTORICAL DATA STORAGE
AND QUERY TECHNOLOGY
USER FRIENDLY INTERFACE
ESTABLISHMENT OF EDB
0
E2
0 2 4 6 8 10 12 14 16 18
Notes: The number shown in the legend box denotes respondents' first, second, third, and other minor choice. The horizontal axis represents number of companies who replied. EWS and EDB are the abbreviations for
engineering work station and engineering database, respectively.
(Source: [21])
4-4 Future Scenarios: Toward a Global Matrix Operation of a Multi-Company
Alliance
From the above analyses, a future scenario of international multi-company CE can be constructed. First, we review the service location in current E&C projects. The left half of Table 2 shows the current service location of Japanese E&C firms. Almost every service, except for a part of detailed engineering, construction, and operation & maintenance (O&M), is located in Japan where Japanese E&C firms have a major technology base. From a business viewpoint, services will need to be located near clients or near the local market, while the best technology remains at its centers of excellence around the world; in other words, the technological base of Japanese E&C firms needs a
31

drastically enhanced global perspective regarding service locations, as depicted in the right half of Table 2.
Table 2 E&C FIRMS' SERVICE LOCATION. NOW AND FUTURE
SERVICES
NOW FUTURE
MARKET TECHNOLOGY MARKET TECHNOLOGY
SALES
BASIC ENGG
DETAILED ENGG a
() ,
(v') v
PROCUREMENT
CONSTRUCTION
' ()
/ o
O&M V
R&D , (')
Notes: Columns denote locations: market means client's country or region, while
technology means countries or regions where technological expertise locates. denotes the applicable location(s) of the service in each row.
(/) indicates locations where a part of each service exists or will exist.
To attain this goal, E&C firms are looking at establishing two organizational hierarchies, one for geographic markets and another for globally distributed technologies.
This means that each client's country or region needs a local company having functions at least of sales, construction, and O&M services, seeking for the best local task-resource fit and coordination fit, and possibly organized around the organization operating in each
32

regional market today. However, basic engineering, detailed engineering, procurement, and R&D services which require both state-of-the-art technology and local knowledge of the client's country, such as codes and regulations, need to be organized both domestically and globally. This can be approached using a global engineering coordination platform, possibly through strategic alliance of leading E&C firms worldwide, seeking for the best global task-resource fit and coordination fit. These two organizational hierarchies then need structural coordination, most probably through a matrix (see Fig. 16).
M
A
E
T
R
K
Figure 16 GLOBAL MATRIX COORDINATION STRUCTURE
..
SERVICES
[COUNTRY A I SALES
COUNTRY B SALES
COUNTRY C I SALES rI
I REGION D SALES
! REGION E SALES i
BASIC
ENGG
DETAIL.
ENGG
PROC
_ -----
I i
CONST.I 0 & M
CONST. & M
CONST. & M
CONST. &M
1
CONST. O & M
_o~tl i_
R& D
I
I
I
>j F
~--
M
R
ALLY
Allen's framework of the rationale for matrix organizations [15] supports the scenario above. According to Allen, a functional organization is more appropriate if technology factors are critical in a business, while a project task-force organization is more appropriate if market factors are critical. Where both technology and market factors are critical in a business, a matrix organization is appropriate, as in the case of Japanese
E&C firms today (see the left half of Fig. 17). As both technology and market factors are now globalizing (see the right half of Fig. 17), a new coordination platform possibly
33

through global alliance and matrix coordination will be required to bind these dispersed functions of CE.
Figure 17 GLOPALIZING TECHNOLOGY AND MARKET FACTORS
TECHNOLOGY
GEOGRAPHICAL DISPERSION
TECHNOLOGY
I 11 III I IV
E&C
FIRM
1t 2 3 14
Ib
A B C I
MARKET
MATRIX ORGANIZATION
(PRESENT)
I, II, III, IV: TECHNOLOGICAL INNOVATION
1, 2, 3, 4: FUNCTIONAL TEAM a, b, c : PROJECT TEAM
A, B, C : MARKET SEGMENT
I
.
COORDINATION PLATFORM
E&C
ALLY a
I II
GEOGRAPHICAL DISPERSION
MARKET
GLOBAL MATRIX ORGANIZATION
(FUTURE)
34

5-1 Key Issues
The proposed engineering and coordination workload model illustrates how E&C firms can benefit from international multi-company CE. Through collaboration with companies having complementary technological specialties or engineering skills, E&C firms can reduce engineering cost as described in Section 3-3. Increase in the number of companies or distance involved in CE increases coordination cost while reducing engineering cost. The coordination cost is determined by two factors: coordination workload and coordination productivity. Coordination workload depends heavily on the way of splitting and allocating engineering tasks, which determines inter-firm task dependencies, and coordination productivity depends on communication tools and previous work experience between the collaborating firms. After seeking the optimum combination of collaborating firms which yields the lowest total cost, there are two possibilities for further cost reduction: improvement of engineering productivity and improvement of coordination productivity.
Japanese E&C firms encounter coordination workload and productivity problems based on their reliance on document-based approval procedures, E&C firm-centered subcontracting, use of traditional coordination/communication methods, and inefficient physical and logical connectivity between different CAE platforms. The current problems can be traced to: (i) inefficient market-based short-term company relationships under insufficient market infrastructure; and (ii) home-country-centered corporate management style of Japanese E&C firms which prevents them from realizing the full potential of international multi-company CE. A scenario for a future international multi-company CE was formulated in this paper.
35

5-2 Suggestions for Future Research
Because of the many attributes involved, the measurement of the workload and the distance conceptualized by the model described in Section 3 is difficult; as such, the discussion in this paper has been largely qualitative. Future research on metrics for workload and distance will enable more quantitative analyses of engineering cost and coordination cost curves.
[1] D. Sriram, "Computer Supported Collaborative Engineering: Workshop Overview,"
Symposium Handout for Collaborative Engineering: toward a More Efficient Product
Development Process, October 19-20, 1993, Industrial Liaison Program,
Massachusetts Institute of Technology, Cambridge, MA, 1993
[2] A. Wong, and D. Sriram, "SHARED: An Information Model for Cooperative Product
Development," Research in Engineering Design, Fall, 1993
[3] L.R. Sayles, and M.K. Chandler, Managing Large System: Organizations for the
Future, Harper & Row, New York, NY, 1971
[4] S. Ghoshal, and D.E. Westney, Organization Theory and the Multinational
Corporation, St. Martin's Press, New York, NY, 1993
[5] W.G. Egelhoff, Organizing the Multinational Enterprise: An Information-Processing
Perspective, Ballinger Publishing Company, Cambridge, MA, 1988
[6] E.M. Roche, Managing Information Technology in Multinational Corporations,
MacMillan Publishing Company, New York, NY, and Maxwell MacMillan
International, Oxford, 1992
[7] R.H. Hollier, R.J. Boaden, and S.J. New, International Operations: Crossing Borders
in Manufacturing and Service, North-Holland, Amsterdam, The Netherlands, 1992
[8] S.E. Eppinger, D.E. Whitney, R.P. Smith, and D.A. Gebara, "Organizing the tasks in
Complex Design Projects," appeared in Design Theory and Methodology -- DTM '90, edited by J. Rinderle, and Carnegie Mellon, The American Society of Mechanical
Engineers, New York, NY, pp.39-46, 1990
[9] Engineering News Record (ENR), "Top 250 International Contractors", July 1981 to
July 1991
36

[10] The Engineering Business, "Special Report: Nippon no Plant Yushutsu no
Kyosoryoku wo Kenshosuru (Examining Competitiveness of Japan's Plant Export
Industry)," The Engineering Business, Engineering Journal Co., Tokyo, Japan,
Vol. 13, No. 18, pp.4-19, October, 1993
[11] K.B. Clark, and T. Fujimoto, Product Development Performance, Harvard Business
School Press, Boston, MA, 1991
[12] International Trade Center, The Export Marketing of Technical Constructing
Services form Developing Countries, 1986
[13] J.D. Thompson, Organizations in Action, McGrow-Hill Book Company, New York,
NY, 1967
[14] D.E. Westney, "The Globalization of Technology and The Internationalization of
R&D," The MIT Japan Program, MITJP 91-05, Massachusetts Institute of
Technology, Cambridge, MA, 1991
[15] T.J. Allen, Managing the Flow of Technology, The MIT Press, Cambridge, MA,
1977
[16] J.K. Hennart, A Theory of Multinational Enterprise, University of Michigan Press,
Ann Arbor, MI, 1985
[17] J.K. Hennart, "Can the 'New Forms of Investment' Substitute for the 'Old' forms? A
Transaction Cost Perspective," Journal of International Business Studies, Vol.10,
No.2, pp.211-234, 1989
[18] J.K. Hennart, "A Model of the Choice Between Firms and Markets," Working Paper,
Wharton School, Philadelphia, PA, 1989
[19] J.K. Hennart, "Control in Multinational Firms: the Role of Price and Hierarchy,"
Organization Theory and the Multinational Corporation, edited by S. Ghoshal, and
D.E. Westney, The MacMillan Press, London, and St. Martin's Press, New York, NY,
1992, pp.
157 1 81
[20] Outline of JGC, JGC Corporation, Yokohama, Japan, 1993
[21] Nippon Kikai Kogyo Rengo Kai, and Engineering Shinko Kyokai, Engineering Kigyo ni Okeru CAE no Genjo to Shorai Doko: Engineering Nohryoku no Kyoka ni
Kansuru Chosa Kenkyu (Current Status and Future Trends of CAE at E&C Firms:
Survey and Research on the Build-up of Engineering Capabilities), Engineering
Shinko Kyokai, Tokyo, Japan, 1992, ENAA 1991-So 1
[22] E.A. Fleishman, and M.K. Quaintance, Taxonomies of Human Performance,
Academic Press, Orlando, FL, 1984
37

Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
WORLD MARKET FOR PETROCHEMICAL PLANTS IN THE 80S
BENEFITS OF MULTI-COMPANY COLLABORATION
ENGINEERING COST CURVE
COORDINATION COST CURVE
INTER-FIRM TASK DEPENDENCIES
TRADEOFF BETWEEN ENGINEERING COST AND
COORDINATION COST
TOTAL COST REDUCTION BY COORDINATION IMPROVEMENT
TOTAL COST REDUCTION BY ENGINEERING PRODUCTIVITY
IMPROVEMENT
STRATEGIC FRAMEWORK FOR MULTI-COMPANY CE
INTER-FIRM ENGINEERING TASK-FLOW
HUMAN RESOURCE ALLOCATION IN A JAPANESE E&C FIRM
PROJECT ENGINEER'S WORK
COMMUNICATION MODES OF PROJECT ENGINEERS
PROBLEMS OF CURRENT CAE SYSTEMS IN THE JAPANESE E&C
INDUSTRY
ISSUES TO GET MOST OUT OF A CAE SYSTEM IN THE
JAPANESE E&C INDUSTRY
GLOBAL MATRIX COORDINATION STRUCTURE
GLOBALIZING TECHNOLOGY AND MARKET FACTORS
'""~"-"111'"---1---^--11111-1- .1_-_ - -
38

Table 1
Table 2
ENGINEERING FEE
E&C FIRMS' SERVICE LOCATION, NOW AND FUTURE
39