مشخصات سیستم کار DSS در دهه چهارم آن

The initially revolutionary DSS agenda is now ancient history. This paper argues that “decision support” provides a richer basis than “DSS” in both practice and research. Using a loan-processing example involving two banks, it shows how work system concepts might be applied to understand decision support in real world settings, and how decision support can come from many sources other than technical artifacts such as DSS. Shifting the focus from “DSS as artifact” to “decision support within a work system” reduces the chances of being misled by techno-hype, vendor sales pitches, and incomplete understanding of determinants of success in organizations.

Initially, DSS was revolutionary idea. It attempted to move beyond MIS (summarizing transaction and operational data for managers), which had attempted to advance beyond EDP (collection and processing of transaction data through computers and electronic media). Launched before PCs existed, the initial concept of DSS focused on using interactive computing in semi-structured decision making. The emphasis on semi-structured decision making seemed important (in academic politics if not in other ways) because that distinguished DSS from OR, especially from optimization models, which attempted to automate decision making, or so it seemed. The interactive use of computers seemed important because it was unclear whether more than a small minority of managers would be willing or able to use computers directly in management work. After 30 years, the original issues that led to the DSS movement have receded to ancient history. Computers are used interactively by managers, nonmanagers, and school children. Computerized data and models are used so commonly and for so many structured, semi-structured, and unstructured tasks that the non-use of computers in typical decision-oriented situations is sometimes a noteworthy exception. With today's widespread adoption of PCs and the Internet, we should simply declare victory on the original DSS agenda that included interactive computing, application of computing to semi-structured problems, use of computers by managers, and the ability to analyze data and models. However, doing this would leave us with a question of whether DSS retains any useful meaning today. With or without the DSS label, researchers and practitioners will continue to do research about sense making and decision making in organizations and will continue to build tools and methods that support those activities. With or without the DSS label, important progress continues in developing tools and methods related to OLAP, data warehousing, data mining, model building, expert systems, neural networks, intelligent agents, group support systems, and communication capabilities for virtual teams. New umbrella terms have emerged, such as business intelligence and decision support applications, but behind the new details and capabilities are many of the same issues and risks that existed in the past. Regardless of whether the new DSS capabilities emphasize better data availability, data analysis, modeling, or communication and coordination, those capabilities have little or no impact until they are incorporated into work systems within organizations. On the other hand, DSS does serve as an umbrella for convening groups of researchers interested in systematic and typically computer-based tools and systems related to sense making and decision making. The new SIGDSS within AIS is a prime example because it provides an institutional home base that supports what Keen [11] calls a self-defined community and what King [13, p. 293] calls an intellectual convocation. But is that all? Could we do equally well if we called the umbrella BWT or XSS or any other three letter acronym? This paper summarizes why my ideas about DSS have moved from enthusiasm to disillusionment to abandonment during the 20+ years since I finished one of the first PhD theses in the area. Next, it reconsiders the notion of decision support from a different viewpoint by exploring how work system concepts might be used to understand decision support in real world settings. Approaching the general area of DSS from a work system viewpoint shifts the perspective and may provide new insights. Decision support is not about tools per se, but rather, about making better decisions within work systems in organizations. The common emphasis on features and benefits of DSS as artifacts rather than on how to improve decisional aspects of work systems in organizations may contribute to the frequently cited (e.g., Ref. [9]) and occasionally questioned (e.g., Ref. [10]) failure rates of data warehousing, CRM, and other technology-based innovations.

The first part of this paper argued that DSS has little meaning other than as an umbrella covering a cluster of research interests related to using technology to support sense making and decision making. The second, more useful part of the paper suggested that decision support is a more useful concept than DSS for most purposes related to practice and research. This is based on the belief that most work systems of any significance include some form of computerized support for sense making and decision making. Possible sources of better decision support may include variations or modifications in any of the nine work system elements, not just the technology, and certainly not just technical artifacts that are sometimes called DSS. Furthermore, any DSS of genuine significance is usually an integral part of a work system and often cannot be separated out easily. For example, the loan approval process at Bank R simply would not operate without the algorithm, and the algorithm is designed specifically to serve this particular work system. Analyzing the algorithm might be interesting, but anyone trying to understand its implementation and success in the organization would need to look at the work system. In addition, the difference between automating and not automating the decision can describe strategy alternatives for a work system, but is less interesting for classifying DSS. For example, whether or not Bank R's highly automated approach fits under the DSS umbrella depends jointly on the definition of DSS that one chooses and on where Bank R sets the bar concerning the conditions and frequency of appeals. The above points do nothing to diminish the importance of creating new DSS theory or building new DSS tools. The development and experimentation required for these pursuits is necessary and significant source of innovations that later may be applied in real world situations. Although research on DSS per se certainly needs to continue, switching the perspective from “DSS as artifact” to “decision support within a work system” encourages consideration of research topics related to each of the work system elements, and this leads to a broader range of potentially valuable research: Business process: How changes in particular business process characteristics (such as degree of structure, range of involvement, complexity, and so on) affect the process efficiency and decision quality. Participants: Both for isolated individuals and for individuals working in teams, how individual characteristics such as personality type, risk aversion, gender, background, and status affect sense making and decision making; how to recognize and address significant differences concerning assumptions, goals, and understanding related to a particular decision. Information: How information and DSS tools can be used within work systems to minimize the effects of common flaws in decision making such as primacy effects, recency effects, overconfidence, poor probability estimation, and groupthink. Technology: How better tools can help people understand situations and deal with information overload; continued development of established tools and techniques such as mathematical modeling, statistical methods, OLAP, data mining, and group support. Product and services: How to evaluate the quality of decisions, especially as part of the decision process rather than after the fact. Customers: How to involve customers in the decision process and obtain greater clarity about their needs and goals. Infrastructure: More effective ways to exploit shared infrastructure within decision processes. Environment: How to visualize whether a possible decision might conflict with the surrounding environment, and how to adjust the decision accordingly. Strategy: How to assess and represent the extent to which a possible decision is aligned with the corporate, departmental, and individual strategies. The topics above are just a subset of the research topics that fit under a broadly construed view of decision support. Focusing on “decision support within a work system” rather than “DSS as artifact” helps align the interests of practitioners and researchers because both groups care about improving decisions within work systems even if only a few members of either group care about DSS as an artifact. Whether or not we retain DSS as an umbrella for getting together at conferences, a deeper look at the idea of decision support without the second S and from a work system viewpoint might help in exploring new directions for research with a high potential for real world application.