How can DSS impact steps in a general decision process model?
by Dan Power
There are many types of computerized decision support systems that support a diverse set of tasks. Experts do not however have a theory of decision support tools that explains what person/group and DSS combinations improve decision making task performance. Discussing and examining a general decision process model can identify how DSS may impact specific steps or general decision tasks. This analysis may assist in formulating a theory of decision support tools. A simple, general, sequential, repeating seven step decision process model (Power, 2002) is reviewed and examined in the context of computerized decision support.
Theory building is difficult and there is always a tradeoff (cf., Weick, 1982) along 3 dimensions: 1) simplicity of the model, 2) generality of the model, and 3) accuracy of the model. The following seven step model emphasizes simplicity and generality, but it lacks some accuracy. Organizational decision making is a complex human activity.
Pondy (1982) identified six characteristics of real decisions that should be kept in mind, but that are not incorporated explicitly in the general process model. First, real decision processes have a shifting group of participants; second, decisions are dependent on past commitments and choices; third, decisions at a given point in time are often intertwined and interdependent with other decision processes; fourth, justifying a decision to others is often more difficult than choosing a course of action; fifth, the definition of a decision problem, its causes and frame of reference may shift during the decision process; and sixth, there is often conflict among decision process participants about facts, assumptions, preferences and outcomes. This complexity complicates the inclusion of DSS in decision processes, but also suggests reasons why DSS are often needed and useful.
So what is the general decision process model? Organizational decision making is more than choosing or deciding. Each of the seven general steps commonly identified in a decision process is important; each step can cause errors and each can potentially be supported by some type of computerized decision support. The next few paragraphs review the seven steps in a general decision process model: 1. Define the problem or decision question, 2. Decide who should decide, 3. Collect information, 4. Identify and evaluate alternative solutions; 5. Decide, 6. Implement, and 7. Follow-up Assessment.
1. Define the problem or decision question
Many managers feel that a well-defined problem is much easier to solve and problem identification reduces the chances of having a good answer but to the wrong problem. When the wrong problem is defined it is impossible to making a successful decision. How a problem is "framed" and defined influences how it is solved and the type of decision support, if any, that is used. Recognizing problems and identifying decision questions can be difficult. The complexity of today's organizations makes it hard in many cases to identify "real" problems and causes and to get beyond problem symptoms. A number of tools and actions can assist in problem identification including a good information system, well thought out standards, and clear and regular communication with key people in an organization. An annual plan which summarizes progress and establishes specific plans for the next year, awareness of new developments in technology, and regular contact and interaction with managers in other organizations also helps managers in identifying decision problems.
2. Decide who should decide
In decision situations, an individual makes some decisions with available information. An individual manager makes other decisions after consulting with colleagues to gather information and opinions. Finally, some decisions should be made by groups using a participative decision making process. Vroom and Yetton (1971) developed a decision tree to help managers decide who should decide in a given decision situation. Their criteria for choosing an autocratic, consultative or group decision process included: need for acceptance of the decision; adequacy of available information; subordinate acceptance of organizational goals; and likelihood of conflict among subordinates about a preferred solution. A computerized tool can remind a manager of the criteria suggested by Vroom and Yetton.
3. Collect information
Once a problem is defined, one can proceed to determine the factors that affect the problem and the information needed about viable alternatives. Without information, decision making is based upon hunch and intuition. On the other hand, too much time and money can be spent gathering data. Formal search and data gathering has a cost in terms of both money and time. The additional search costs need to be weighed against the benefits of additional data. MIS and data-driven DSS can provide information for decision making, but a cost is incurred in development and in use of DSS. The entire realm of business intelligence and query and reporting tools is focused on analyzing data, collecting and creating decision relevant information. Decision makers may and usually do loop back to this step from subsequent steps.
4. Identify and evaluate alternative solutions
The most creative part of decision making is the identification of alternatives and the determination of which ones should receive serious consideration and analysis. Computer supported brainstorming can help generate useful ideas in many situations. A long list of ideas is generally more useful than a short list. A large number of ideas is more likely to lead to some high quality ideas than focusing on one or a few readily available ideas. Early in the brainstorming process the objective is increasing the number of ideas. How good, unique or impractical an idea may be is of very little concern in brainstorming. A commonly used group brainstorming and idea evaluation tool is the Nominal Group Technique (NGT). Some GDSS have tools based on NGT, including silent idea generation, idea sharing, rating or ranking of alternatives (cf., Delbecq, Van de Ven and Gustafson, 1975). Also, identifying and using relevant criteria with computerized support can help evaluate alternatives.
To make a decision is to commit to a course of action or to intentionally postpone action. It is common in real decision situations for some decisions to be delayed or postponed. In some situations however, a decision must be made. The specific situation or circumstances require or demand that a choice is made. Decisions are then sometimes made very quickly and with less information than one would like and some feasible alternatives may not be evaluated or even considered. DSS are not usually as helpful in these "crisis" decision situations, but access to contingency plans can be helpful. In other less time sensitive situations, computerized decision support can assist in reaching a group consensus or help in justifying a choice to stakeholders. DSS can potentially reduce procrastination and indecision by helping structure a decision situation and gather information. DSS can also help weight and structure decision criteria on "soft" criteria like company impact of alternatives or reaction of competitors.
A choice or decision is a culmination of one process. The specific decision process may be long and convoluted or rapid and simple. But for any problem and set of alternatives, made with or without decision support, once a decision is made, something usually happens. Decisions often trigger actions and information technology can focus and direct those actions and complete a broader process of action and change. Computerized DSS can help communicate decisions, monitor plans and actions, and track performance. Computerized performance management can help with implementation.
7. Follow-up and assessment
Measuring and evaluating the consequences of a decision that has been implemented calls for the decision maker to accept responsibility for the decision. During follow up, new problems may or may not be discovered. In some cases, minor adjustments and corrective actions are necessary. Because situations do not remain the same for very long, managers are often dealing with problems that grew out of the solutions chosen to previous problems. So the decision loop or cycle is repeated. Again definition of a problem leads to assessment of the decision that was implemented which leads to consciousness of new problems. DSS can help in monitoring, follow up and assessment.
Computerized decision support is more appropriate and useful for some steps and tasks than for others. Multiple decision support tools may assist in a complex decision process. Decision support analysts who understand the decision process can do a better job of identifying appropriate computerized decision support.
As always, your comments and feedback are welcomed.
Delbecq, A.L., A.H. Van de Ven, and D. H. Gustafson. Group Techniques for Program Planning. Glenview, IL: Scott, Foresman, Inc., 1975.
Mintzberg, H., D. Raisinghani, and A. Theoret. "The Structure of 'Unstructured' Decision Processes." Administrative Science Quarterly, Vol. 21, June 1976, pp. 246-275.
Pondy, L. "On Real Decisions." In Ungson, G.R. and D. N. Braunstein. Decision Making: An Interdisciplinary Inquiry. Boston, MA: Kent Publishing, 1982, pp. 309-311.
Power, D. J., Decision Support Systems: Concepts and Resources for Managers, Westport, CT: Greenwood/Quorum Books, 2002.
Vroom, V. and P. Yetton. Leadership and Decision-Making. Pittsburgh: University of Pittsburgh Press, 1973.
Weick, K. E. "rethinking Research on Decision Making." In Ungson, G.R. and D. N. Braunstein. Decision Making: An Interdisciplinary Inquiry. Boston, MA: Kent Publishing, 1982, pp. 325-333.
Last update: 2008-05-14 09:51
Author: Daniel Power
You cannot comment on this entry