DSS Design and Development Conclusions
In 1985 Jack Hogue and Hugh Watson surveyed managers in organizations with DSS. Each participant was an active DSS user. Two-thirds of the organizations had built their DSS using an evolutionary, prototyping approach and the remaining organizations had used more of an SDLC approach. It appeared that if the DSS supported managers throughout the company or that it required company-wide data, then the SDLC approach was used. The evolutionary approach was used for smaller-scale systems where a DSS development tool was available. Nine of the eighteen companies used DSS generators to develop their systems. This finding is probably descriptive of current practice.
When managers could specify information requirements in advance, then the systems development life cycle approach was more likely to be used. Hogue and Watson also found that when IS Specialists developed the DSS then SDLC steps were more likely to be followed. Senior managers reported they were most involved in the idea, information requirements and acceptance steps associated with building a DSS. Middle managers reported they were somewhat involved in all of the steps involved in building the DSS that they were using. When prototyping and evolutionary design was used, managers reported more involvement in the design and development process. The IS group was usually involved in building the DSS, but staff from an Information Systems department were rarely in a leadership role. Potential users of the DSS usually assumed the leadership role.
The DSS design and development approach that is used for a new DSS project should depend on the amount of data needed and its sources, the number of planned users, any models and analytical tools used, and the amount of anticipated use. Many small, specialized DSS are built quickly using end-user development or rapid prototyping. Large, Enterprise-Wide DSS are built using sophisticated tools and systematic and structured systems analysis and development approaches. Creating Enterprise-Wide DSS environments remains a complex and evolutionary task. An Enterprise-Wide DSS inevitably becomes a major part of a company's overall information systems infrastructure. Despite the significant development differences created by the scope of a DSS, all DSS have similar technical components and share a common purpose, supporting decision-making.
A number of authors suggest the perceived usefulness and the perceived ease of use of an Information System or Decision Support System is a major determinant of its use. MIS managers can influence both the perceived usefulness and the perceived ease of use of a new system by using a participative development process. MIS staff need to establish a meaningful "social exchange" with potential users and DSS developers must be responsive to user requests, questions and needs.
More research is needed on the effectiveness of approaches for designing and developing DSS. But, in general, MIS professionals should use a decision-oriented design process and then either a rapid prototyping or SDLC development process. End-user DSS can be satisfactory and inexpensive and MIS staff should support such development rather than discourage it. Rapid prototyping will be useful in building many types of DSS, but SDLC has a role in developing complex, networked, Enterprise-Wide, Data-Driven DSS. DSS analysts and managers need to be familiar with all of the approaches for building DSS.
One can state some generalizations about Design and Development of Decision Support Systems. Fist, when a project idea is proposed, focus on description and diagnosis of decision-making and an analysis of the decision and processes involved. We call this Decision-Oriented Diagnosis.
Second, following diagnosis, one should conduct a feasibility study and in many situations prepare a feasibility report. Third, if the project seems feasible, then managers and IS staff need to decide to build or buy the proposed DSS. In many situations, a solution will be customized for the DSS.
Fourth, in general, Model-Driven and Knowledge-Driven DSS are built using rapid prototyping. Data-Driven DSS are built using rapid prototyping or a Systems Development Life Cycle approach. Communications-Driven and Group DSS are usually purchased and installed on company computers.