Topic: Decision Support Systems
Randall E. Louw
University of Missouri St. Louis
Prof. Vicky Sauter
Fall 2002 ©
Decision Support Systems have evolved over the past three decades from simple model-oriented systems to advanced multi-function entities. During the 1960’s, most Decision Support Systems were fairly based on powerful (and expensive) mainframe computers which provided managers with structured, periodic reports. MIS theory developments during the 1970’s saw Decision Support Systems evolve into more elaborate computer-based systems that supported production, promotion, pricing, marketing and some logistical functions. By early 1980’s Decision Support Systems enjoyed more interests from academics and the framework for Decision Support Systems was greatly expanded by the end of the decade. It was only during the 1990’s that a paradigm shift occurred in Decision Support Systems and more complex systems, which incorporated, advanced database technology and client/server capabilities, were emerging from many areas in business processes. As many organizations started to upgrade their network infrastructure, object oriented technology and data warehousing started to make its mark on Decision Support Systems. The rapid expansion of the Internet provided additional opportunities for the scope of Decision Support Systems and consequently many new innovative systems such as OLAP and other web-drive systems were developed.13
According to Sprague and Watson (1996) conceptual models or frameworks are crucial to understanding a new and/or complex system. They define DSS broadly as an interactive computer based system that help decision-makers use data and models to solve ill-structured, unstructured or semi-structured problems.3
DSS provides varying analysis without much programming effort and is usually directed towards non-technical users/managers. Managers main uses for a DSS includes searching, retrieving and analyzing decision relevant data to allow them to summarize main points which assist them in making more informed and educated decisions. Users often search for correlations between data without rewriting the underlying MIS or software application and most DSS allows graphic capabilities, which not only allows trend analysis and reporting for top executives, but also assists managers in mapping out conjoint analysis and alternative scenarios to answer “what if” queries. Consequently, DSS supports both tactical and strategic decisions and are employed to leverage manager’s expertise in a certain field.
DSS varies in scope – some are intended for multiple users (more common nowadays) and other are “stand-alone” units (common in the past). In addition to that, DSS can take on many different forms and can be used in many different ways, i.e. some DSS focus on models, others on data and others on communications.5 The better the manager understands the different categories, scope and uses of DSS, the better he will be able to specify requirements for a DSS that he wants to implement or buy.
Thus, in order to comprehend the intricate complexities of what services a DSS can provide, we first have to look at the stand-alone units that supports the DSS. Although DSS can be dissected into many different components, I will mainly concentrate on a few important aspects of its design.
Having summarized the most important functions of a DSS, it should be remembered that a DSS is only as good as the individual components that it consist of: DSS is built on top of a transaction system, a database and a data model, all of which provides the DSS with data and information that is processed and presented to the user in a simplified form.
The fist important aspect of DSS is that they provide information which are used in the decision making process. The emphasis here is not on the quantity of information, but rather the quality. There are multiple factors that qualify information as having good quality (such as timeliness, relevant, accurateness, consistency, unbiased, etc.) but the important consideration factor is how information is used in order to attain a certain goal.
It is a common notion that information can be (and often is…) misinterpreted, leading to inaccurate conclusions which adversely affects the quality of the decision making process inside an organization (as is humoristically depicted in figure 1…).
Managers often postpone difficult decisions and request “more information” without really attempting to comprehend the information that is already available to them. They falsely rely on the assumption that information should give them guidance, in stead of realizing that their understanding of relevant information will help them formulate their own ideas, which should be based on both a intricate understanding of the information, the necessary knowledge available around them and ultimately their intuition that is develop based on experience. This curvilinear relationship between information and comprehending information is depicted in figure 2*. The graph suggests that the quantity of information is only useful up to a certain point, afterwhich more information becomes redundant/obsolete.
Thus, the user should try to digest the current information available to him/her first and resist the temptation to seek more information without first understanding the current information. Only as the user gains “new insights” and “breakthrough thinking”, additional information should be sought where relevant. To make a good decision, one needs not only information about the specific instance, but also an understanding of the domain. In other words, one needs a set of principles, models, templates or other abstractions. Better understanding enables the identification of what information is relevant and consequently, less information is required because the irrelevant components can be ignored. This not only decreases the complexity of the decision process, but also decreases the processing load on the manager and leaves him with more time to focus on critical and situation-relevant information segments.
A successful business makes good decisions, implements them well - and then learns from the experience and abstractions in order to do better next time. These abstractions are then re-usable for making new decisions with different information, facilitating the process of knowledge management and ultimately enhancing the overall quality of decision making inside the organization.
Unlike information, which often relates only to specific instances, knowledge is contend-rich and re-usable and should thus be captured whenever possible to provide a point of reference for future similar scenarios.
As mentioned earlier, DSS can be classified in many different categories. Amongst the common ones are the following:
2.1 Data driven DSS
These DSS has file drawer systems, data analysis systems, analysis information systems, data warehousing and emphasizes access to and manipulation of large databases of structured data
2.2 Model driven
The underlying model that drives the DSS can come from various disciplines or areas of specialty and might include accounting models, financial models, representation models, optimization models, etc. With model drive DSS the emphasize is on access to and manipulation of a model, rather than data, i.e. it uses data and parameters to aid decision makers in analyzing a situation. These systems usually are not data intensive and consequently are not linked to very large databases.
2.3 Knowledge driven
These systems provide recommendation and/or suggestion schemes which aids the user in selecting an appropriate alternative to a problem at hand. Knowledge driven DSS are often referred to as management expert systems or intelligent decision support systems. They focuses on knowledge and recommends actions to managers based on an analysis of a certain knowledge base. Moreover, it has special problem solving expertise and are closely related to data mining i.e. sifting through large amounts of data to produce contend relationships.
2.4 Document driven
These systems help managers retrieve and mange unstructured documents and web pages by integrating a variety of storage and processing technologies to provide complete document retrieval and analysis. It also access documents such as company policies and procedures, product specification, catalogs, corporate historical documents, minutes of meetings, important correspondence, corporate records, etc. and are usually driven by a task-specific search engine.2
2.5 Communication driven
This breed of DSS is often called group decision support systems (GDSS). They are a special type of hybrid DSS that emphasizes the use of communications and decision models intended to facilitate the solution of problems by decision makers working together as a group. GDSS supports electronic communication, scheduling, document sharing and other group productivity and decision enhancing activities and involves technologies such as two-way interactive video, bulletin boards, e-mail, etc.
(See table 1 in appendix section for a overview of an expanded Decision Support Systems framework)
2.6 Inter- and Intra-organization DSS
These systems are driven by the rapid growth of Internet and other networking technologies such as broadband WAN’s, LAN’s, WIP, etc. Inter-organization DSS are used to serve companies stakeholders (customers, suppliers, etc.), whereas intra-organization DSS are more directed towards individuals inside the company and specific user groups. The latter, because of their stricter control, are often stand-alone units inside the firm.
2.7 New breeds of DSS
Designed for managers looking to make sense of their information, OLAP tools structure data hierarchically – the way managers think of their enterprises, but also allows business analysts to rotate that data, changing the relationships to get more detailed insight into corporate information.17
· OLAP and web-based Decision Support Systems are by far the more popular Decision Support Systems these days. Their definition and functionality extends far beyond the scope of this paper and consequently we will not explore these systems in detail. Many other Decision Support Systems are on the market today, but to explore all of them would, for now, be a farcical objective.
Keeping the various distinctions and classifications of DSS in mind, a DSS should be described in terms of:
An example: a management team might want a document-driven, inter-organizational, product design, web-based DSS
Managerial Decision Systems4 are only as good as the underlying DB supporting it. Users can manipulate data temporarily and needs special access and/or authorization to manipulate data in underlying DBS
In order to provide the DSS with the right data, there should be a structure in place for storage, access, manipulation and retrieval of large amounts of data. The database is the “feeder” of the DSS and good database design is thus crucial in improving the overall performance of the DSS. Database design and management is a complex and often ill-conceived topic and extends well beyond the scope of this paper. Consequently our discussion on databases will parsimonious at best. A good DBMS ensures data integrity, reduces data redundancy, follows a logical sequence and are consistent in its performance.
Relational databases are in most cases the system of choice when it comes to designing a DSS. This is primarily because of the flexibility associated with a relational DB but also because it allows normalization (reduction in data duplication), which helps with maintenance of a large database.7 The ability of these databases to identify relationships between entities makes information analyzes an important feature to the success of the DSS. Some hierarchical and network databases are still being used today, but merely so because of the costs involved with migrating to a totally new DB platform overshadows the maintenance costs of legacy systems – even this is gradually changing as more organizations are putting additional requirements on their DSS, underscoring the need for a more suitable DB design.
Because the database constitutes such a vital link in the capabilities of any DSS its structure and design should be carefully evaluated and implemented with due concern of applications build on top of it. There seems to be a general trend in recent years, to either migrate to web-based DSS, using a task specific search-engine, or to build the DSS around a thin client fat server environment, using network and web-based technologies. This is perhaps a further justification of the need for an advanced database system since the Internet is the most comprehensive network of thousands of interconnected databases and web pages.
4. The Data Warehouse
DB generally provides current information about the organization relating to the underlying transactional processes, but it fails to provide historical, contend rich information that are often more important to the decision making process than stand alone islands of information. The data warehouse fills this gap by capturing operational data and presenting it in a more meaningful format, using a relational database, and ultimately complimenting the functions of the DB used in the DSS. Thus the data warehouse and the DB coexists to provide synergistic outcomes which supports information requirement of the DSS superimposed on the systems platform. Development in the areas of DB has brought about the use of “intelligent agents” which assists in the speed and simplicity of ad hoc queries in large DB and data warehouses.
Information uses and requirements differ at each managerial level.6 Higher managerial levels have a greater reliance on external environmental information and to some degree, less reliance on internally generated information. Top management commonly use information to make decisions about long term planning and thus analyze long term trend information to make their decisions (Gore et al, 1984).3 Also at top-level decision-making, conjoint analysis are crucial and managers need more information about risk and uncertainty, potential outcomes (and their likelihood/probability, etc.) and possible actions to take. At lower management levels, information emphasis is internally generated and relies on short-term goals. However, reports and analyses generated by lower management teams are often used in decision-making efforts of top management, and it is thus crucial that the DSS supports the lower management just as much as it lends itself to top management. In short, the information needs for different levels of management are directed towards supervisory functions for lower management, tactical decision making for middle management and strategic decision making for top management. See appendix 1 for a more elaborate outline of information requirements of managers.
Moreover, because the DSS are employed to improve management control, it should address the primary tasks of management control. These are:
i. Allocating resources to specific activities
ii. Preparing budgets covering both the expenditure of resources and the expected achievement goals and objectives
iii. Observation of results achieved in return for resources used
iv. Evaluation of these results
v. Modification of activities and resource allocation in accordance with the evaluation of results
6. Intangible payoffs of Decision Support Systems
Amongst the many benefits of using a Decision Support Systems, some of the intangible advantages include:
i. Improved internal control
ii. Better management awareness of internal strength and weaknesses and external threats and opportunities.
iii. Enhances long term profitability by improving quality of decision making
iv. Faster response to changes in the business environment
v. Allows managers’ to record experience and use it in future for similar scenarios that may arise within the organization. Recorded experience becomes information once it is used in the managerial decision making process7 and it is part of a bigger sub-category of MIS called knowledge management.
Keeping these benefits in mind, it is not surprising that Decision Support Systems has enjoyed widespread recognition and support over the last decade and are increasingly being used in companies to enhance their strategic competitiveness within various industries.
The DSS design and development process involves input from every aspect of the business process that is going to be affected within the firm. Business executives and managers should familiarize themselves with DSS so that they can more accurately communicate their needs and requirements, making it easier for the formulation of deliverables, outline of capabilities, outcomes, needs, and what decisions should be supported by the proposed DSS.3 The following steps can be used as a methodology to evaluate, design and implement a DSS:
1. Firm’s strategic focus
First and foremost, the organization wishing to deploy a DSS needs to clearly outline its strategic focus within the industry, with relevant emphasis on future objectives and the direction in which the firm is moving. This will in turn provide the basis for information requirements and scalability of the DSS and will give a framework for the project team charged with designing and implementing the system.
2. Creating an evaluation team
Using a multi-disciplinary team approach with individuals from database, networking, communications, accounting, etc. a project team should be selected based on backgrounds and expertise. This team will perform an initial analysis that will provide the blueprint for the rest of the project. Team formation (forming, norming, storming and performing) and group dynamics are key aspects, which ensures optimal performance of the team as a whole. Consequently, the following issues should be considered during this important phase of the project:
· Specify roles and delegate responsibilities
· Anticipate and manage time commitment
· Choose appropriate leadership
· Create a charge/direction for the team with a clearly defined methodology
3. Defining requirements
The more the team know about DSS categories, the better they will be able to fine-tune the specifications of the system and select the appropriate type of the Decision Support Systems. A multi-disciplinary team approach is necessary with feedback from all relevant stakeholders who are involved in the implementation of the DSS. Team members should be made responsible to communicate with other members of the organization who will be affected by the selected DSS and they should try to express all the relevant needs from different players in the firm. It is crucial to elicit “buy-in” from all individuals who will use or will be affected by the Decision Support Systems since non-compliance by some parties might dilute the strength of the system.
· Issues to consider:
o What functions will managers perform, now and in the future, with the DSS?
o What controls and security are needed?
o What are the operational performance requirements?
o What is the design concept
o How does it fit into the current architecture and network layout
4. System design and evaluating products/alternatives
· Clearly specify deliverables
· Include a preliminary evaluation of both products and vendors available
· “off-the-shelf” DSS are hard to find and most organizations have to adopt a specific DSS to suite their individual business needs
· Functional screening and detailed review based on predetermined specifications and criteria:
o Cost, design, support and installation issues should be addressed separately and as a whole
o Is the “package” compatible with the needs of the company?
o Hardware and software requirements
· Includes performance evaluation:
o Can the system handle heavier loads?
o What is the scalability (horizontal and vertical)?
o What are the direct and indirect costs?
o Stability, reliability and consistency of system?
o Quality of information provided: timely, accurate, reliable, unbiased?
o Restrictions of systems
o Capabilities of system
o After implementation support? Costs involved?
5. Negotiating with vendors
Review vendors according to product offerings, reputation in industry, financial stability, experience in similar projects/systems. Also clarify issues such as
o Service contract
o Price of package
o Time frame of deliverables
o Review vendors offerings and compare with other vendors
o Vendors involvement with implementation/deployment of system
o Training of users and after implementation support
DSS can be extremely beneficial to any organization’s overall performance. However, DSS can also be the cause of great confusion, misperception and even inaccurate analysis – these systems are not designed to eliminate “bad” decisions. DSS are there to facilitate a manager in making operational decisions, but the ultimate burden of responsibility lies with the manger. Managers can sometimes be over-optimistic in their expectations of a DSS and develop a unrealistic reliance on the system (Power, C.J; Caveat Emperor).13 Also, if managers continue to ask the “wrong” questions (queries), the benefit of the systems is already partially lost. When managers have preconceived notions and misconceptions about a certain operational function, a DSS can magnify the harm by justifying the manger’s position simply because of the logic of the managers’ queries or because the manager draw the wrong conclusion about a certain output. Users of the DSS should be “critical” consumers and never replace the systems functions with their own “clinical reasoning”
A holistic understanding of all the models, templates and abstractions that the Decision Support Systems are built on will give managers a better understanding of how to approach opportunities, challenges and threats within the organization (see appendix 2 for an outlay of the processing model for Decision Support Systems).
The design methodology for Decision Support Systems provides the foundation on how the systems fits into the operations of the firm and how it can enhance quality decisions for managers. This area of analysis, design and implementation should be viewed with almost a superstitious awe, since failure to produce the right system can lead to more damage than good if managers are unaware of the implicit logical layout of their system.
However, Decision Support Systems remains a tool that can provide firms with a sustainable competitive advantage and in many industries a robust Decision Support Systems are the rule rather than the exception.
As technologies evolve, Decision Support Systems will change and adapt to the challenges of the new economy and firms should thus continuously be aware of new landscapes that are being formed and streamline their Information Systems applications with current industry trends and technological advancements.
1. Goal setting
2. Long-range plans
1. External Information, e.g. competitor actions, government regulation, economic factors, resource availability
2. Internal Information e.g. financial reports, key exception reports, etc.
3. Long term trends
4. Conjoint analysis (“what if’ analysis)
1. Definition of objectives
2. Medium ramge plans
3. Tactical decisions
1. Internal information
2. Short term trends
3. Some conjoint analysis
1. Attainment of objectives
2. Short range plans
1. Internal information e.g. recent historical information, detailed operational reports, appropriate exception reports