- Email: [email protected]
Is there a best MIS department location?
Briefings A_-_.
----__-____
_-
Is There a Best MIS Department Location? Donald W. Kroeber
I. Introduction
College of Business Administration, James Madison Univcrsity, Harrisonburg, Virginia, USA
One of the most intriguing organizational structure questions associated with management information systems (MIS) is, “Where should the MIS department be located?” Because early computers were recognized chiefly for their ‘prodigious’ arithmetic skills, managers tended to associate computer applications with accounting activities and most commonly located their data processing activity within the accounting
and
Hugh J. Watson r4lege of Business Administration, Universityof Georgia, A thens, Ceorgina, USA An important organizational structure issue involves the positioning of the MIS department. When computers were first introduced, the data processing function was often placed within an accounting department. During the second generation of computer technology, some organizations operated different computers at different divisions; this generally resulted in a decentralized organizational structure. The third generation of computer technology provided a seemingly unlimited storage capacity and remote terminal capability; there was then a resurgance of centralization of control. Frequently, the MIS department was given a higherlevel, separate organizational location. Empirical data attests to the merits of such a positioning. Today, in the fourth generation of computing, there are several options - a central MIS, a distribution MIS, or a dispersed MIS. Each confiiuration has its merits, and the final selection of a configuration depends on the information needs of the organization, the extent of homogeneity of its applications, and the technological and managerial investment the organization is willing to make. Key words:
Donald W. Kroeber is an Assistant Professor of Management and the Assistant Vice President for Academic Affairs at James Madison University in Harrisonburg, Virginia, USA. He holds the B.A. degree in Economics, M.S. in Finance, and Ph.D. in Business Administration. He teaches, conducts research, and publishes in the fields of statistics and management science as well as MIS. His articles on MIS have appeared in The Data Educator, The Encyclqpedia of ProfessiDna!Mart, agement, and MSU Business Topics. He is also the author of an article on statistical quality control and two books on statistics and quantitative methods. This is his first contribution to Information & Management. He is a member of AIDS and the American Society for Quality Control (ASQC).
centralized/decentrauded, data processing, distributed system, MIS, organizational structure.
Dr. I?@ J. Wattin is Professor of Management at the University of Georgia. He has also taught at the - University of Hawaii and the Florida ” State University and is an Affiliate : Faculty Member of the Japan-Amer. ica Institute of Management Science. His articles have appeared in Manage’ ment Science, Interfaces, Omega, : Academy of Management Journal, Journal of Business Research, Journa? of Bank Research, CaliforniaManage~-~_3,L;l,~.:-,i< - ,. . ment f?evieG; .MSU Business Topics, and others. He is also author of -Ccimputersfor Business: A Managerial Emphasis @PI), QuantitativeMethods for Business Decisions (McGrawHill), M&zagement: Making OrganizationsPerform (Macmillan), and others. He is a member of TIMS, AIDS and the Academy of Management.
0 North-Holland Publishing Company Information & Management 2 (1979) 165- 173 165
department or under the o(sntroller. The few exceptions to this practice occurred in highly technical industries where compuOerswere used for scientific applications; in these cases, the computers were located in research and develoF;nent and similar departments. In either case, the! high initial investment and scarcity of personnel to program and maintain the complex machines invariably resulted in the information system being implemented at a single, centralized location. Iater, as second generation computers became available and as computer applications were developed for production, personnel, marketing, and other function;rl areas, an alternative to a centralized location evolved. Because of the limited capacity of second generation hardware,a single co:mputercould not ah~ayg accommodate all demand for storage (memory) and processing (CPU). Consequently, several
I
i
I 1
departments and divisions began to acquire their own computers with the personnel to operate and program them. Of course, some organizations did not take the path of dec#entralization and chose, instead, to increase the capacity and organizational role of the accounting computer to that of an internal service center. The advent of third generation technology brought about a resurgence of computer centralization. Single computers or computer systems had the capacity to handle all the applications for an organization. Remote terminals provided easy access to the computer from many organization locations. In many organizations, a new, high-level department was created to control and coordinate information system activities for the functional departments and, perhaps most important, to provide functionally integrated management information to top-level decision makers.
J
I I
2MD OENERATION
I
Fig. I. The evohltion of the MIS Department IocatJnIl.
8RD
GENERAtION
D. W. Kroeber, H.J. Watson /MIS Departntent Location
The evolution of MIS location through the first three computer generations is shown graphically in Figure 1. It illustrates the change from centralized to decentralized and back to a centralized location. We are now on the threshold of a fourth generation of computer systems and an accompanying reorganization of MIS responsibilities within the organizational structure. What changes will these new capabilities bring to the question of MIS location? Is there a single, best location for every organization? Let us look at some of the factors which enter the decision of where to place the MIS department.
167
which blend second generation capacity and other considerations (ease of operation, reduced maintenance and environmental controls, etc.) with third or fourth generation technology, seem to favor decentralization. Furthermore, the range of MIS-related skills and activities has broadened and spread to such an extent that personnel throughout an organization are now capable of performing tasks that were once the rctsponsibility of a few elite specialists. Almost every recent BBA or MBA graduate has had some for ma1 training in programming, systems analysis, and the quantitative methods typical of computer-assistec decision making.
2. Fourth Generation Hardware 3. Industry Trends While one may argue whether the fourth generation has arrived or not (IBM, whose model 370 is sometimes termed “fourth generation” insists that it has not), the technology of the fourth generation already exists -just as the transistor existed years before the second generation. We can anticipate in the fourth generation grand scale integration (GSI) of electronic components; vast direct-access memory capacity - I ,000 to 100,000 times greater than current core or integrated circuit (IC) memory. perhaps in the form of holographic or magnetic bubble memories; full-color display surfaces of liquid crystal or ferroelectric design; remote solid-state keyboards without moving parts; and ultra-short wave or optical communications systems capable of hundreds of thousands of data channels [6]. We can also anticipate expanded use of late third generation hardware, notably mini- and micro-computers, either as standalone computers or as “smart terminals” in a distributed system. Do these technological advances argue. for centralization or decentralization? The answer is not clear, and there appear to be confhcting forces ;;t \:ork which may dictate a hybrid or compromise :\+x!!,‘an. On the one hand, the enormous capacit!., .i %;:rth generation computers, the ease of remote #,!a&t&ry and output, and the growing use of organization-wide data base management systems for storing, retrieving, maintaining, and reporting data all suggest a centralized system, and the higher the control is placed in the organization, the better. On the other hand, low cost mini-computers,
Before incorporating fourth genera&n c+nsiderations into an organizational model, it is helpful to look at industry reactions to the location problem. After all, the pioneer work in MIS occurred in organizations and preceded, rather than followed, the conceptualization of theoretical MIS considerations by scholars. Therefore it is altogether appropriate that consideration be given to current industry practice. During the 1960’s, a trend away from the traditional accounting department location began [4]. By the late 1960’s, finance and accounting departments could still claim that they housed the “majority” of information systems departments [5]. But by 1976, the authors’ study of over 150 randomly selected large organizations in business and government showed that only a third of the information system still reported to the finance/accounting locations [3]. Table I shows the percentage of locations of information system department by industrial category from that study. Table 1 suggests that the location depends upon the industrial category. Two categories stand out as being significantly different from the others: financial instit utions and government organizations. Financial institutions, among the earliest commercial users of computers, were also among the first to create separate information system departments. Certainly some of the credit for this action must be given to an enlightened management, but there were other contributing factors. Bank!;, which made up a sizeable portion of the surveyed financial institutions, tradi-
of Information System Department LOdOll
Flee/
SepaCatle ACGOUlltiJlg Department
Other
353% 10.0 58.5 47.1 47.4 7.2
52.9 60.0 26.8 41.1 47.4 61.5
11.8 30.0 14.7 11.8 5.2 30.3
33.9
47.2
18.9
parate their various income-producing and tend nut to have accounting departin the same sense as manufacturing or other of firms.A second force leading to a separate the practice of many banks to sell excess apxity to other business. From an organiint af view, this practice is facilitated by ent location for the information system t organizations do have accounting s, or at least controllers (which they are to spell comptroller), and these offices the custodians of early government informants. However, bureaucracies tend to be conhen, led by the Department of Defense, Government decided that separate MIS were the most efficient, virtually ail govemorganizations relocated their systems offices. der is not that only 7.2 per cent are under ccounting control, but that so many still be remaining industries, almost one-half still
locate their MIS department in finance/accounting. However, an organization by organization analysis of these industries reveals a trend, for more sophisticated systems (in terms of both hardware and software technology and information system performance) to be associated with an independent location while more of the less complex systems remain in finance/ accounting.
4. Benefits of a Separate Location The arguments for a separate MIS location have been repeated often in the information system literature: a separate location reduces duplication of personnel and equipment, it facilitates standardization of data codes and reporting procedures, it eliminates friction between the department with MIS responsibility and potential users in other departments, and it is in a better position to provide integrated information to top management. But are these arguments supported by the results of cost/benefit considerations? The data processing managers of the survey sample were asked to evaluate the cost/benefit ratio of their systems. Their answers, tabulated by MIS department location, are shown in Table 2. A total of 79.6 percent of separate location managers felt that MIS benefits exceeded MIS costs (either much greater or slightly greater) while 65.9 per cent of finance/accounting location managers felt that way. Perhaps even more significant is the 12.0 per cent of finance/accounting location managers who felt that benefits were less (either slightly or much less) than costs against only 1.4 per cent of separate location managers who shared that opinion. These are not overwhelming differences, but they do favor the separate location.
maria&s
opinionof _h@Sbenefits vs. MIS investment SS
No opinion
9.6 7.7
3.5 9.0
D. W. Kroeber, H.J. Watson/MIS Depwttnent Location
5. Cumnt Organizational Options
169
assistance from the concerned department. Even though this system is centralized, data preparation, and perhaps even data entry, are accomplished in the functional or staff department with internal resources. Remote job entry temrinals or other appropriate peripherals may be located in the departments for this purpose. Each entry updates Gre companywide data base for subsequent use by all system users. A central MIS is most appropriate in an organization with relatively homogeneous applications and information requirements. Banks, insurance companies, and other Wnancial institutions that engage primarily in “number crunching” would do f,vell to consider a single, separate MIS department high in the organizational structure.
Given the advances of late-third or early-fourth generation technology, the trends noted in industry, (and the perceived advantages of a separate location, three major options are worthy of further consideration: a highly placed (in the organizational structure) central MIS, a distributed MIS with a highly placed “mother” computer and departmental minicomputers, and a dispersedMIS featuring stand-alone minis in operational departments [2]. 5.1. CentralMIS A central MIS, of the type shown in Fig. 2, is usually located under a vice president for services or a comparable top manager. Functional and staff departments perform their own initial systems analysis following guidelines and data base considerations promulgated by the systems branch. Upon approval by the MIS department, the systems branch performs the detailed systems analysis and turns the project over to the EDP branch for programming. Testing is performed jointly by the EDP and systems branches with
5.2. DistributedMIS Figure 3 shows a typical distributed MIS organization. It is similar to the central MIS organization except for the addition of small EDP sections in functional and staff departments. A distributed system operates much like the central system when applications involve changes to the data base. For example, a
PRES
VP .
I
‘-iii
EDP BRANCH
Fig. 2. Centralized MIS.
r-l
EDI UANCH
l-l
SYSTEMS BEANCtl
I
3
Fig. 3. Distributed MIS.
u&on department which uses a computer model scheduling has that applicatio= approved, ed, and programmed by the MIS department to compatibility with inventory and accounting applications- This insures that inventory records will be updated with the depletion of appropriate raw maserial stocks and the addtiion of the correct amount of fmished goods. Automated accounting ledgers for the value of raw material, work in process, and goods must also be updated. But a strictly application, such as a line-balancing model, be developed in the prod-uction department and run on the production minicomputer. Standard pror
and does not permaalter data base content. nnel, but frequently
heterogeneous applications. A ma-
nufacturing firm, wii h its specialized functions in production, finance, accounting, and marketing is an example of such an organization. 5.3. Dispersed MIS A dispersed MIS can be thought of as a distributed MIS without a “mother computer.” As Figure 4 shows, a dispersed MIS still includes a highly placed, central MIS department, but has no central EDP branch. Instead, the MIS department functions in a staff role, giving guidance and assistance to other departments and MIS a.dvice to top management. A few *ystems analysis personnel remain in the MIS department, which also serves as a program library and documentation center for the organization. This facilitates the exchange of ideas betureen departments and helps to reduce ‘duplication of effort in systems analysis and programrfiing. Programs are conceived, designed, tind written in the functional and staff departments in much the same way as they were in the decentralized, second generation phase. Not many organizations are well served by dispersed systems.. The ‘few that are have widely varying
D. W. Kroeber, H.J. Watson/MIS Department Location
171
,
PIES
7
Lr VP
FUNCTIONAL
DEPT EDP
MIS DEW
zrr-~~-~-rr~~~-~.
-
A
1
Fig, 4. Dispersed MIS.
Table 3 Organizational Considerations
of MIS Location
---
Suggested MIS location Characteristic
Centralized
Distributed
Dispersed
Technology
Large main frame, high volume input/output, report oriented system.
Large main frame for transaction processing/report generation, intelligent terminals for managerial problem solving.
Mini-computers, low volume input/output, problem solving orientation.
Personnel
Predominantly operational level, low technical orientation and educational level
Large operational base with some technically oriented middle and top managers.
Large middle management group, innahvative, welleducated.
Users
Traditional
Both traditional and quantitatively oriented managers.
Pr0fessionals, quantitatively oriented managers.
Structure
Centralized
Limited decentralization.
Decentralized
Budget control
Tight, EDP expenditures strictly controlled.
Moderate, EDP expenditures controlled similarly to other capital outlays.
Liberal, EDP and other expenditures allocated to profit centers.
Managerial
Traditional “Theory X”, strong central control.
Moderate, limited delegation of authority.
Liberal, “Theory Y”, delegation of authority.
Operations
Largely homoge;neous, stable technological eqvironment.
Largely homogeneous, unique operations.
Heterogeneous, rapidly changing technological environment.
Size
Small to medium, single location.
Large, scattered or multiple locations,
managers.
a few
Large or very large, multiple locations.
this one might add management’s philosophy, operations, size, and others. Table 3 illustrates how organizational characteristics can combine to suggest either a centralized, distributed, or dispersed system. The characteristics for centralized and dispersed systems are relatively unambiguous, since they are the end points of a continuum. Such is not the case with distributed systems. There are a variety of distributed systems and Table 3 only presents examples of characteristics from the middle of the continuum. Unfortunately for managers faced with the MIS location decision, characteristics seldom align as neatly as shown in Table 3. More often than not, some characteristics argue for one location while others support another. Fig. 5 shows how two characteristics - users and operations - can interact to sug gest anywhere from a centrdlized to a dispersed system. The best approach for management to use when ma&g the location decision is to first decide on the
infomjation and computing requirements in their various departments, with lit‘tlecommon ground. In a variation of a dispersed MIS, a few departments in an organization may be dispersed from the central or distributed system. The business school computer in a university, the researchand development computer in a laboratory, and the environmental control computer in a factory are examples of dispersed facilities in an otherwise centralized or distributed MIS.
6. &ga&ational Considerations The choice between centralized, distributed, or dispersed systems is guided by the interactions of a number of organizational characteristics. There is no definitive list of these characteristics, but Walsh [7] suggests that th.e dominant forces include computer technology, pec**nnel, users, the existing organizational structure, and budgetary ConsideraGons. To ~ ---
-\
Guontitotiw
\
Distributed hardware, programming and analysis; centralized planning
\
\
\
\
\
\
2;Fried \
\
\
\
\ \
Distributed hordwore and analysis; cen!rdlized planning and pro&amming
\ \
\
\
\
\
\
\
\
DMributod hqrdwore; c~ntralix~’ planrrinlp, plairniir@ atid analyrir
\ \.
Tradition01
\
Hommeous ’ OPEdAJlONS Fig. 5. Comkiizations of characteristics.
I
\
D. W. Kroeber, H.J. Watson/MIS Department Location
dominant characteristics of the organization. These characteristics can then be compared to those described in Table 3. This comparison can provide insights as to what placements is best for the organization. It should also be kept in mind that hybrid locations are sometimes possible and desirable. Davis [l] explores a number of combinations, including one where hardware and planning are centralized while systems analysis and/or programrning are decentralized. For example, a university academic computing facility is likely to have centralized hardware, but applications may well be developed by departmental faculty to support unique classroom work or research. It is not possible to describe here all of the combinations which exist and serve various organizations well.
7. Conclusion Which MIS department location is best? It should be clear by now that there is no single best location for all organizations. There are a variety of organizational characteristics which argue for different placements. However, Table 3 should provide assistance in searching through the characteristics to find the op-
173
timal placement for a particular organization. Of course, any of the three locations described here centralized, distributed, or dispersed - can be tailored to the needs of the organization by subtle variations in the placement of MHScomponents.
References [ 1] G.B. Davis, Management Information Systems: Conceptual Foundations, Structure, and Development, (New York: McGraw-Hill, 1974). [2] A.L. Kelsch, “Dispersed and Distributed Data Processing,” Journal of Systems Management, Vol. 29, March, 1978. [3] D.W. Kroeber, An Empirical Study of the Current State of Information Systems Evolution, doctoral dissertation, (University of Ceoigia, Athens, 1976). [4] W.E. Reif and R.M. Monczka, “Locating the Systems Department,” Journal of Systems Management, Vol. 24, December, 1973. [5] D.H. Sanders, Computers in Business: An Introduction New York: McGraw-Hill, 1975). [6] R. Turn, Computers in the 1980’s (New York: Columbia University Press, 1974). [7] M.E. Walsh, “MIS - Where Are We, How Did We Get Here and Where Are We Going?” Journal of Systems Management, Vol. 29, November, 1978.
----__-____
_-
Is There a Best MIS Department Location? Donald W. Kroeber
I. Introduction
College of Business Administration, James Madison Univcrsity, Harrisonburg, Virginia, USA
One of the most intriguing organizational structure questions associated with management information systems (MIS) is, “Where should the MIS department be located?” Because early computers were recognized chiefly for their ‘prodigious’ arithmetic skills, managers tended to associate computer applications with accounting activities and most commonly located their data processing activity within the accounting
and
Hugh J. Watson r4lege of Business Administration, Universityof Georgia, A thens, Ceorgina, USA An important organizational structure issue involves the positioning of the MIS department. When computers were first introduced, the data processing function was often placed within an accounting department. During the second generation of computer technology, some organizations operated different computers at different divisions; this generally resulted in a decentralized organizational structure. The third generation of computer technology provided a seemingly unlimited storage capacity and remote terminal capability; there was then a resurgance of centralization of control. Frequently, the MIS department was given a higherlevel, separate organizational location. Empirical data attests to the merits of such a positioning. Today, in the fourth generation of computing, there are several options - a central MIS, a distribution MIS, or a dispersed MIS. Each confiiuration has its merits, and the final selection of a configuration depends on the information needs of the organization, the extent of homogeneity of its applications, and the technological and managerial investment the organization is willing to make. Key words:
Donald W. Kroeber is an Assistant Professor of Management and the Assistant Vice President for Academic Affairs at James Madison University in Harrisonburg, Virginia, USA. He holds the B.A. degree in Economics, M.S. in Finance, and Ph.D. in Business Administration. He teaches, conducts research, and publishes in the fields of statistics and management science as well as MIS. His articles on MIS have appeared in The Data Educator, The Encyclqpedia of ProfessiDna!Mart, agement, and MSU Business Topics. He is also the author of an article on statistical quality control and two books on statistics and quantitative methods. This is his first contribution to Information & Management. He is a member of AIDS and the American Society for Quality Control (ASQC).
centralized/decentrauded, data processing, distributed system, MIS, organizational structure.
Dr. I?@ J. Wattin is Professor of Management at the University of Georgia. He has also taught at the - University of Hawaii and the Florida ” State University and is an Affiliate : Faculty Member of the Japan-Amer. ica Institute of Management Science. His articles have appeared in Manage’ ment Science, Interfaces, Omega, : Academy of Management Journal, Journal of Business Research, Journa? of Bank Research, CaliforniaManage~-~_3,L;l,~.:-,i< - ,. . ment f?evieG; .MSU Business Topics, and others. He is also author of -Ccimputersfor Business: A Managerial Emphasis @PI), QuantitativeMethods for Business Decisions (McGrawHill), M&zagement: Making OrganizationsPerform (Macmillan), and others. He is a member of TIMS, AIDS and the Academy of Management.
0 North-Holland Publishing Company Information & Management 2 (1979) 165- 173 165
department or under the o(sntroller. The few exceptions to this practice occurred in highly technical industries where compuOerswere used for scientific applications; in these cases, the computers were located in research and develoF;nent and similar departments. In either case, the! high initial investment and scarcity of personnel to program and maintain the complex machines invariably resulted in the information system being implemented at a single, centralized location. Iater, as second generation computers became available and as computer applications were developed for production, personnel, marketing, and other function;rl areas, an alternative to a centralized location evolved. Because of the limited capacity of second generation hardware,a single co:mputercould not ah~ayg accommodate all demand for storage (memory) and processing (CPU). Consequently, several
I
i
I 1
departments and divisions began to acquire their own computers with the personnel to operate and program them. Of course, some organizations did not take the path of dec#entralization and chose, instead, to increase the capacity and organizational role of the accounting computer to that of an internal service center. The advent of third generation technology brought about a resurgence of computer centralization. Single computers or computer systems had the capacity to handle all the applications for an organization. Remote terminals provided easy access to the computer from many organization locations. In many organizations, a new, high-level department was created to control and coordinate information system activities for the functional departments and, perhaps most important, to provide functionally integrated management information to top-level decision makers.
J
I I
2MD OENERATION
I
Fig. I. The evohltion of the MIS Department IocatJnIl.
8RD
GENERAtION
D. W. Kroeber, H.J. Watson /MIS Departntent Location
The evolution of MIS location through the first three computer generations is shown graphically in Figure 1. It illustrates the change from centralized to decentralized and back to a centralized location. We are now on the threshold of a fourth generation of computer systems and an accompanying reorganization of MIS responsibilities within the organizational structure. What changes will these new capabilities bring to the question of MIS location? Is there a single, best location for every organization? Let us look at some of the factors which enter the decision of where to place the MIS department.
167
which blend second generation capacity and other considerations (ease of operation, reduced maintenance and environmental controls, etc.) with third or fourth generation technology, seem to favor decentralization. Furthermore, the range of MIS-related skills and activities has broadened and spread to such an extent that personnel throughout an organization are now capable of performing tasks that were once the rctsponsibility of a few elite specialists. Almost every recent BBA or MBA graduate has had some for ma1 training in programming, systems analysis, and the quantitative methods typical of computer-assistec decision making.
2. Fourth Generation Hardware 3. Industry Trends While one may argue whether the fourth generation has arrived or not (IBM, whose model 370 is sometimes termed “fourth generation” insists that it has not), the technology of the fourth generation already exists -just as the transistor existed years before the second generation. We can anticipate in the fourth generation grand scale integration (GSI) of electronic components; vast direct-access memory capacity - I ,000 to 100,000 times greater than current core or integrated circuit (IC) memory. perhaps in the form of holographic or magnetic bubble memories; full-color display surfaces of liquid crystal or ferroelectric design; remote solid-state keyboards without moving parts; and ultra-short wave or optical communications systems capable of hundreds of thousands of data channels [6]. We can also anticipate expanded use of late third generation hardware, notably mini- and micro-computers, either as standalone computers or as “smart terminals” in a distributed system. Do these technological advances argue. for centralization or decentralization? The answer is not clear, and there appear to be confhcting forces ;;t \:ork which may dictate a hybrid or compromise :\+x!!,‘an. On the one hand, the enormous capacit!., .i %;:rth generation computers, the ease of remote #,!a&t&ry and output, and the growing use of organization-wide data base management systems for storing, retrieving, maintaining, and reporting data all suggest a centralized system, and the higher the control is placed in the organization, the better. On the other hand, low cost mini-computers,
Before incorporating fourth genera&n c+nsiderations into an organizational model, it is helpful to look at industry reactions to the location problem. After all, the pioneer work in MIS occurred in organizations and preceded, rather than followed, the conceptualization of theoretical MIS considerations by scholars. Therefore it is altogether appropriate that consideration be given to current industry practice. During the 1960’s, a trend away from the traditional accounting department location began [4]. By the late 1960’s, finance and accounting departments could still claim that they housed the “majority” of information systems departments [5]. But by 1976, the authors’ study of over 150 randomly selected large organizations in business and government showed that only a third of the information system still reported to the finance/accounting locations [3]. Table I shows the percentage of locations of information system department by industrial category from that study. Table 1 suggests that the location depends upon the industrial category. Two categories stand out as being significantly different from the others: financial instit utions and government organizations. Financial institutions, among the earliest commercial users of computers, were also among the first to create separate information system departments. Certainly some of the credit for this action must be given to an enlightened management, but there were other contributing factors. Bank!;, which made up a sizeable portion of the surveyed financial institutions, tradi-
of Information System Department LOdOll
Flee/
SepaCatle ACGOUlltiJlg Department
Other
353% 10.0 58.5 47.1 47.4 7.2
52.9 60.0 26.8 41.1 47.4 61.5
11.8 30.0 14.7 11.8 5.2 30.3
33.9
47.2
18.9
parate their various income-producing and tend nut to have accounting departin the same sense as manufacturing or other of firms.A second force leading to a separate the practice of many banks to sell excess apxity to other business. From an organiint af view, this practice is facilitated by ent location for the information system t organizations do have accounting s, or at least controllers (which they are to spell comptroller), and these offices the custodians of early government informants. However, bureaucracies tend to be conhen, led by the Department of Defense, Government decided that separate MIS were the most efficient, virtually ail govemorganizations relocated their systems offices. der is not that only 7.2 per cent are under ccounting control, but that so many still be remaining industries, almost one-half still
locate their MIS department in finance/accounting. However, an organization by organization analysis of these industries reveals a trend, for more sophisticated systems (in terms of both hardware and software technology and information system performance) to be associated with an independent location while more of the less complex systems remain in finance/ accounting.
4. Benefits of a Separate Location The arguments for a separate MIS location have been repeated often in the information system literature: a separate location reduces duplication of personnel and equipment, it facilitates standardization of data codes and reporting procedures, it eliminates friction between the department with MIS responsibility and potential users in other departments, and it is in a better position to provide integrated information to top management. But are these arguments supported by the results of cost/benefit considerations? The data processing managers of the survey sample were asked to evaluate the cost/benefit ratio of their systems. Their answers, tabulated by MIS department location, are shown in Table 2. A total of 79.6 percent of separate location managers felt that MIS benefits exceeded MIS costs (either much greater or slightly greater) while 65.9 per cent of finance/accounting location managers felt that way. Perhaps even more significant is the 12.0 per cent of finance/accounting location managers who felt that benefits were less (either slightly or much less) than costs against only 1.4 per cent of separate location managers who shared that opinion. These are not overwhelming differences, but they do favor the separate location.
maria&s
opinionof _h@Sbenefits vs. MIS investment SS
No opinion
9.6 7.7
3.5 9.0
D. W. Kroeber, H.J. Watson/MIS Depwttnent Location
5. Cumnt Organizational Options
169
assistance from the concerned department. Even though this system is centralized, data preparation, and perhaps even data entry, are accomplished in the functional or staff department with internal resources. Remote job entry temrinals or other appropriate peripherals may be located in the departments for this purpose. Each entry updates Gre companywide data base for subsequent use by all system users. A central MIS is most appropriate in an organization with relatively homogeneous applications and information requirements. Banks, insurance companies, and other Wnancial institutions that engage primarily in “number crunching” would do f,vell to consider a single, separate MIS department high in the organizational structure.
Given the advances of late-third or early-fourth generation technology, the trends noted in industry, (and the perceived advantages of a separate location, three major options are worthy of further consideration: a highly placed (in the organizational structure) central MIS, a distributed MIS with a highly placed “mother” computer and departmental minicomputers, and a dispersedMIS featuring stand-alone minis in operational departments [2]. 5.1. CentralMIS A central MIS, of the type shown in Fig. 2, is usually located under a vice president for services or a comparable top manager. Functional and staff departments perform their own initial systems analysis following guidelines and data base considerations promulgated by the systems branch. Upon approval by the MIS department, the systems branch performs the detailed systems analysis and turns the project over to the EDP branch for programming. Testing is performed jointly by the EDP and systems branches with
5.2. DistributedMIS Figure 3 shows a typical distributed MIS organization. It is similar to the central MIS organization except for the addition of small EDP sections in functional and staff departments. A distributed system operates much like the central system when applications involve changes to the data base. For example, a
PRES
VP .
I
‘-iii
EDP BRANCH
Fig. 2. Centralized MIS.
r-l
EDI UANCH
l-l
SYSTEMS BEANCtl
I
3
Fig. 3. Distributed MIS.
u&on department which uses a computer model scheduling has that applicatio= approved, ed, and programmed by the MIS department to compatibility with inventory and accounting applications- This insures that inventory records will be updated with the depletion of appropriate raw maserial stocks and the addtiion of the correct amount of fmished goods. Automated accounting ledgers for the value of raw material, work in process, and goods must also be updated. But a strictly application, such as a line-balancing model, be developed in the prod-uction department and run on the production minicomputer. Standard pror
and does not permaalter data base content. nnel, but frequently
heterogeneous applications. A ma-
nufacturing firm, wii h its specialized functions in production, finance, accounting, and marketing is an example of such an organization. 5.3. Dispersed MIS A dispersed MIS can be thought of as a distributed MIS without a “mother computer.” As Figure 4 shows, a dispersed MIS still includes a highly placed, central MIS department, but has no central EDP branch. Instead, the MIS department functions in a staff role, giving guidance and assistance to other departments and MIS a.dvice to top management. A few *ystems analysis personnel remain in the MIS department, which also serves as a program library and documentation center for the organization. This facilitates the exchange of ideas betureen departments and helps to reduce ‘duplication of effort in systems analysis and programrfiing. Programs are conceived, designed, tind written in the functional and staff departments in much the same way as they were in the decentralized, second generation phase. Not many organizations are well served by dispersed systems.. The ‘few that are have widely varying
D. W. Kroeber, H.J. Watson/MIS Department Location
171
,
PIES
7
Lr VP
FUNCTIONAL
DEPT EDP
MIS DEW
zrr-~~-~-rr~~~-~.
-
A
1
Fig, 4. Dispersed MIS.
Table 3 Organizational Considerations
of MIS Location
---
Suggested MIS location Characteristic
Centralized
Distributed
Dispersed
Technology
Large main frame, high volume input/output, report oriented system.
Large main frame for transaction processing/report generation, intelligent terminals for managerial problem solving.
Mini-computers, low volume input/output, problem solving orientation.
Personnel
Predominantly operational level, low technical orientation and educational level
Large operational base with some technically oriented middle and top managers.
Large middle management group, innahvative, welleducated.
Users
Traditional
Both traditional and quantitatively oriented managers.
Pr0fessionals, quantitatively oriented managers.
Structure
Centralized
Limited decentralization.
Decentralized
Budget control
Tight, EDP expenditures strictly controlled.
Moderate, EDP expenditures controlled similarly to other capital outlays.
Liberal, EDP and other expenditures allocated to profit centers.
Managerial
Traditional “Theory X”, strong central control.
Moderate, limited delegation of authority.
Liberal, “Theory Y”, delegation of authority.
Operations
Largely homoge;neous, stable technological eqvironment.
Largely homogeneous, unique operations.
Heterogeneous, rapidly changing technological environment.
Size
Small to medium, single location.
Large, scattered or multiple locations,
managers.
a few
Large or very large, multiple locations.
this one might add management’s philosophy, operations, size, and others. Table 3 illustrates how organizational characteristics can combine to suggest either a centralized, distributed, or dispersed system. The characteristics for centralized and dispersed systems are relatively unambiguous, since they are the end points of a continuum. Such is not the case with distributed systems. There are a variety of distributed systems and Table 3 only presents examples of characteristics from the middle of the continuum. Unfortunately for managers faced with the MIS location decision, characteristics seldom align as neatly as shown in Table 3. More often than not, some characteristics argue for one location while others support another. Fig. 5 shows how two characteristics - users and operations - can interact to sug gest anywhere from a centrdlized to a dispersed system. The best approach for management to use when ma&g the location decision is to first decide on the
infomjation and computing requirements in their various departments, with lit‘tlecommon ground. In a variation of a dispersed MIS, a few departments in an organization may be dispersed from the central or distributed system. The business school computer in a university, the researchand development computer in a laboratory, and the environmental control computer in a factory are examples of dispersed facilities in an otherwise centralized or distributed MIS.
6. &ga&ational Considerations The choice between centralized, distributed, or dispersed systems is guided by the interactions of a number of organizational characteristics. There is no definitive list of these characteristics, but Walsh [7] suggests that th.e dominant forces include computer technology, pec**nnel, users, the existing organizational structure, and budgetary ConsideraGons. To ~ ---
-\
Guontitotiw
\
Distributed hardware, programming and analysis; centralized planning
\
\
\
\
\
\
2;Fried \
\
\
\
\ \
Distributed hordwore and analysis; cen!rdlized planning and pro&amming
\ \
\
\
\
\
\
\
\
DMributod hqrdwore; c~ntralix~’ planrrinlp, plairniir@ atid analyrir
\ \.
Tradition01
\
Hommeous ’ OPEdAJlONS Fig. 5. Comkiizations of characteristics.
I
\
D. W. Kroeber, H.J. Watson/MIS Department Location
dominant characteristics of the organization. These characteristics can then be compared to those described in Table 3. This comparison can provide insights as to what placements is best for the organization. It should also be kept in mind that hybrid locations are sometimes possible and desirable. Davis [l] explores a number of combinations, including one where hardware and planning are centralized while systems analysis and/or programrning are decentralized. For example, a university academic computing facility is likely to have centralized hardware, but applications may well be developed by departmental faculty to support unique classroom work or research. It is not possible to describe here all of the combinations which exist and serve various organizations well.
7. Conclusion Which MIS department location is best? It should be clear by now that there is no single best location for all organizations. There are a variety of organizational characteristics which argue for different placements. However, Table 3 should provide assistance in searching through the characteristics to find the op-
173
timal placement for a particular organization. Of course, any of the three locations described here centralized, distributed, or dispersed - can be tailored to the needs of the organization by subtle variations in the placement of MHScomponents.
References [ 1] G.B. Davis, Management Information Systems: Conceptual Foundations, Structure, and Development, (New York: McGraw-Hill, 1974). [2] A.L. Kelsch, “Dispersed and Distributed Data Processing,” Journal of Systems Management, Vol. 29, March, 1978. [3] D.W. Kroeber, An Empirical Study of the Current State of Information Systems Evolution, doctoral dissertation, (University of Ceoigia, Athens, 1976). [4] W.E. Reif and R.M. Monczka, “Locating the Systems Department,” Journal of Systems Management, Vol. 24, December, 1973. [5] D.H. Sanders, Computers in Business: An Introduction New York: McGraw-Hill, 1975). [6] R. Turn, Computers in the 1980’s (New York: Columbia University Press, 1974). [7] M.E. Walsh, “MIS - Where Are We, How Did We Get Here and Where Are We Going?” Journal of Systems Management, Vol. 29, November, 1978.