Ch. 3
Analyzing Business Decision Processes

How do individuals and groups make decisions? What steps should be completed? A sequential model of decision-making can help analyze how decisions are being made and how they should be made (cf., Mintzberg, Raisinghani, and Theoret, 1976).

Simon (1965) identifies 3 stages in a sequential decision-making process: 1) intelligence -- finding occasions for making a decision; 2) design -- finding, inventing, developing, and analyzing alternative courses of action; and 3) choice -- selecting a course of action. A fourth stage called implementation is also often discussed even though Simon considers implementation as a separate decision process of intelligence, design and choice. Prior to implementation a major decision has is made and implementation then involves many supporting actions and hence choices. Managing these stages and how they interact can be a major challenge in complex, rapidly changing and ambiguous or uncertain decision situations. Each of the above stages can be supported by a variety of Decision Support Systems.

According to Hammer and Champy (1993, p. 36-39), IBM Credit Corporation, a wholly owned subsidiary of IBM, had a business decision process that evaluated customer's requests for financing that included the following five steps:

Step 1. A salesperson called in a request for financing, which was recorded on paper by 1 of 14 clerical staff members "sitting around a conference room table in Old Greenwich, Connecticut". This step initiated the process.

Step 2. Someone physically walked the paper request to the credit department, where a specialist entered the request into a computer and checked the credit status of the customer. The result was written on the credit repot. Then, the paper-based credit report was delivered to the business practices department.

Step 3. The business practices department used a different computer system to modify a standard loan agreement according to any special requests made by the customer. The document was attached to the original request and delivered to the pricer.

Step 4. The pricer keyed all the information into a PC spreadsheet and determined the appropriate interest rate. This figure was written onto the other forms and delivered to the clerical group.

Step 5. The clerical group converted all paper documents into a quote letter and delivered it to the sales representative using FedEx.

The entire process took six days on average, although it sometimes took as long as 2 weeks. Some people would say a Model-Driven DSS is needed to support Step 4, but the entire process can be redesigned and automated. What would you do? Can you redesign the process and then recommend appropriate DSS for each step? Would a Communications-Driven DSS help?

To redesign the process two senior managers at IBM Credit took a financing request and walked it themselves through all five steps asking personnel at each step "to put aside whatever they were doing" and process the request as they normally would. They learned the actual work took 90 minutes. The problem was in the structure of the process and the lack of integrated computer support. IBM Credit developed a new computerized systems for a deal structurer who handled all of the steps! Difficult decisions could be referred to a small group of specialists. The new Decision Support System and process resulted in a 90-percent reduction in cycle time and an enormous improvement in productivity.