Copyright
Steven Lewis Alter.

Computer aided decision making in organizations : a decision support systems typology online

. (page 1 of 2)
Online LibrarySteven Lewis AlterComputer aided decision making in organizations : a decision support systems typology → online text (page 1 of 2)
Font size
QR-code for this ebook


LIBRARY

OF THE

MASSACHUSETTS INSTITUTE

OF TECHNOLOGY



k



'♦

■^



k



#



HD28

.m1+iU



/ Dewey






Center for Information Systems Research

Massachusetts Institute of Technology

Alfred P Sloan School of Ivlanagement

50 Memorial Drive

Cambridge, Massachusetts. 02139

617 253-1000



COMPUTER AIDED DECISION MAKING IN
ORGANIZATIONS: A DECISION SUPPORT
SYSTEMS TYPOLOGY



Steven Alter



REPORT CISR-11
SLOAN WP 855-76




May 1976



.mHi4




- Abstract

In an attempt to develop an understanding of the key Issues related
to the success or lack of success of decision support systems, interview
data concerning 56 systems was generated and analyzed. It was found
that the systems in the sample clustered into system types based on a
one-dimensional view of the kind of operation performed by the system
for the user. Differences among the t3rpe8 of systems are discussed
along with implications of this typology for practitioners and researchers.

I. Introduction

The phrase "decision support systems" denotes a computerized system
designed specifically to help people make decisions. This phrase
emerged in reaction to the frequently expressed feeling that the purpose
of most EDP applications was to increase efficiency in record keeping
and transaction processing rather than to Improve the quality of decision
making [3]. The emphasis on the word "support" is also Important since
decision support systems are usually viewed as tools which help people
make decisions but do not automate decision making per se. Although a
certain amount of conjucture has been generated concerning the "nature"
of decision support systems, the importance of interactive problem
solving, the relevance of certain system characteristics, the need for
special implementation skills and design processes, etc., there is
relatively little organized empirical data and the conjectures are often
contradictory.

In response to this situation, a largely exploratory study of
decision support systems was undertaken [1]. The purpose of this study

0727379



was to gain an understanding of the dynamics of decision support systems
and the key issues related to their success or lack of success. The
data consisted of 56 structured "mini-case studies" of decision support
systems. Effectively, each mini-case was a structured story of the
system in terms of interview responses to questions under the following
headings:

- general background

- system history and characteristics

- types of use and impact

- limitations and types of disuse or abuse

- factors in favor of or opposed to getting started

- factors in favor of or opposed to successful implementation

One of the major findings of the study was that "decision support system"
does not seem to constitute a homogeneous category. Rather, it was
found that the systems in the sample fell into seven reasonably distinct
categories which could be arranged in terms of the type of calculation
or operation performed by the system.

The purpose of this article is to present this typology and to
demonstrate that it differentiates among systems in significant ways.
As a result, it should be helpful in comparing and interpreting both
research studies and practical implementation experiences.

II. A Typology of Decision Support Systems
There are many ways to categorize computer-based systems. The most
commonly used taxonomic schemes include:

- functional area: marketing, production, finance

- decision perspective: operational control, management control,

strategic planning (see [3])

- problem type: structured vs. unstructured (see [6], [3], and [5])

- computer technology: interactive vs. batch

- modeling approach: simulation vs. optimization



3

Eight exploratory case studies [2] prior to the bulk of data
collection led to a certain amount of skepticism about the usefulness of
any one of the schemes in either motivating sampling or generating
hypotheses for the remainder of the research effort. For instance, a
financial projection system for operational planning seemed very similar
in concept and structure to a system for strategic planning. Likewise,
the expected significance of interactive computation seemed to be lost
when decision makers were not hands-on users of systems. Difficulties
in deciding whether one repetitive business problem was less structured
than another also diminished the usefulness of this distinction. As is
reported in [1], eventually a series of hypotheses were generated based
on a particular taxonomic viewpoint which did not work well because it
failed to focus the sample into convincing clusters. The taxonomy
described below was a post hoc attempt to sort out the patterns that
emerged from the 56 stories. Mason [4] describes a parallel, but more
abstract taxonomy suggested by Churchman.

The variable that forms the basis of the typology is the "degree of
action implication of system outputs," i.e., the degree to which the
system's output could directly determine the decision. This is related
to a spectrum of generic operations which can be performed by decision
support systems. These generic operations extend along a single dimension
ranging from extremely data oriented to extremely model oriented:

- retrieving a single item of information

- providing a mechanism for ad hoc data analysis

- providing pre-specified aggregations of data in the form

of reports

- estimating the consequences of proposed decisions

- proposing decisions

- making decisions

The idea here is that a decision support system can be categorized in
terms of the generic operations it performs, independent of the type of
problem, functional area, decision perspective, etc.



4

Clustered from this viewpoint, the 56 systems in the sample fell
into 7 reasonably distinct types which can be labeled as follows:

A. "File drawer systems" allow immediate access to data items.

B. "Data analysis systems" allow the manipulation of data by
means of operators tailored to the task and setting or
operators of a general nature.

C. "Analysis information systems" provide access to a series of
data bases and small models.

D. "Accounting models" calculate the consequences of planned
actions based on accounting definitions.

E. "Representational models" estimate the consequences of actions
based on models which are partially non-definitional.

F. "Optimization models" provide guidelines for action by generating
the optimal solution consistent with a series of constraints.

G. "Suggestion models" perform mechanical work leading to a specific
suggested decision for a fairly structured task.

Figure 1 illustrates that this typology can be collapsed into a simple

dichotomy between data oriented and model oriented systems. Such a

simplification loses a great deal of information, however, by grouping

systems which differ in many significant ways. This will be demonstrated

by discussing each of these system types in turn, with emphasis on some

of the key points which seemed especially relevant to systems of each

type. As a concluding overview. Figure 2 will summarize over 100 pages

of prose descriptions in [1] by presenting a tabular comparison of the

various types.

The main overall conclusion here is simply that "decision support

system" does not constitute a homogeneous category. Although the phrase

is very useful as a way of distinguishing decision-motivated computer

applications from efficiency-motivated applications, both Figure 2 and

the following descriptions by type imply that the differences among

types of systems are quite significant. As a result, "type of DSS"

should be viewed as a key contingency by people who are studying or



CO
UJ
Q-

>-



a

UJ

I-
z

UJ

•■«

o
I

<

I-
<



Z
UJ

»«
K

o



lU
Q
O






5



-1




<


CO


>


»«


UJ


CO


^4


>


OS


' -1


H


<


UJ


z


oe


<


<


<


K


1-


<


<


a


o



z
o

CO



z
o

^^

(O
UJ

o
o



1

Online LibrarySteven Lewis AlterComputer aided decision making in organizations : a decision support systems typology → online text (page 1 of 2)