Managers must make decisions in an increasingly complex, rapidly changing, volatile, and ambiguous environment. This environmental turbulence increases risks for managers and organizations. To help reduce and manage risk now is an opportune time to implement more and better computerized decision support. Managers should implement or update systems to provide better business intelligence (BI), analytics, and other types of computerized decision support. This turbulent environment should motivate managers to evaluate computerized decision support projects. What has changed? Modern decision support is more useful and more sophisticated.
In the past 25 years, software vendors have regularly used new terms for capabilities associated with decision support. For some vendors, legacy terms such as decision support system (DSS) were rejected as too general, while for others legacy terms reminded potential customers of failed projects or unrealistic expectations. A new term such as analytics provided a new start for selling a decision support capability. Despite the changing terminology, managers continue to want and need computerized information systems to support their decision-making.
Decision support does not insure correct decisions. One hopes vendors have realized it is important to identify and better manage customer expectations. Decision support applications differ widely depending upon the purpose of the system and perceived need. Current technologies can support a wide range of decision-making tasks. Decision support consultants, designers, and researchers have learned much about using information technology (IT) solutions to support decision making and that knowledge can benefit managers and their organizations.
Prior research and experience supports two fundamental premises associated with computerized decision support. First, computers and IT can help people make important decisions. Second, computerized decision support assists and supports managers, and keeps them as part of the decision-making process. The overriding goal of computerized decision support developers is unchanged—improve human decision-making effectiveness and efficiency with IT solutions.
Many organizations have integrated computerized decision support into day-to-day operating activities and use systems for performance monitoring. Frequently, managers download and analyze sales data, create reports, and analyze and evaluate forecasting results. DSS can help managers perform tasks, such as allocating resources, comparing budget to actual results, drilling down in a database to analyze operating results, projecting revenues, and evaluating scenarios. Data warehouses can create a single version of the truth for advanced analytics and reporting. More managers are using business dashboards and scorecards to track operations and support decision-making.
Decision support research began in the 1960s and the concepts of decision support, decision support systems, and the acronym DSS remain understandable, intuitively descriptive, and even obvious in meaning. Related terms such as analytics, BI, and knowledge management are of more recent origin and are interpreted in different ways by vendors and consultants. Decision support is a broad concept that prescribes using computerized systems and other tools to assist in individual, group, and organization decision-making. One goal in this and succeeding chapters is to make some sense out of the decision support jargon.
The seven questions included in this chapter discuss the need for decision support, the technology skills and knowledge needed by managers, targeted users, a historical perspective on decision support, a theory of decision support, and adopting decision technologies. The final question identifies characteristics of modern decision support applications.
What Is the Need for Decision Support?
Today decision-making is more difficult: the need for decision-making speed has increased, overload of information is common, and there is more distortion of information. On the positive side, there is a greater emphasis on fact-based decision-making. A complex decision-making environment creates a need for computerized decision support. Research and case studies provide evidence that a well-designed and appropriate computerized DSS can encourage fact-based decisions, improve decision quality, and improve the efficiency and effectiveness of decision processes.
Most managers want more and better analyses and decision-relevant reports quickly. Managers do have many and increasing information needs. The goal of many DSS is to create and provide decision-relevant information. There is a pressing need to use technology to help make decisions better. Decision makers perform better with the right information at the right time. In general, computerized decision support can help transfer and organize knowledge. Effective decision support provides managers more independence to retrieve and analyze data and documents, as they need them.
From a different perspective, we need decision support because we have decision-making biases. Biases distort decisions. Reducing bias has been a secondary motivation for decision support, but it is an important one. Most managers accept that some people are biased when making decisions, but doubt a computerized solution will significantly reduce bias. Evidence shows information presentation and information availability influence and bias a decision maker’s thinking both positively and negatively. Evidence shows system designers can reduce the negative bias. Also, evidence shows decision makers “anchor” on the initial information they receive and that influences how they interpret subsequent information. In addition, decision makers tend to place the greatest attention on more recent information and either ignore or forget historical information.1 Good decision support software can reduce these and other biases.
Managerial requests for more and better information, today’s fast paced, technology-oriented decision environments, and significant decision-maker limitations create the need for more and better computerized decision support. Managers should strive to provide computerized decision support when two conditions associated with a decision situation are met: (a) good information is likely to improve the quality of a decision, and (b) potential users recognize a need for and want to use computerized support in that situation.
Introducing more and better decision support in an organization does create changes and challenges for managers. For example, using a smart phone with decision support applications or a tablet PC connected to the Internet and corporate databases requires new skills and new knowledge of managers.
What DSS Skills and Knowledge Do Managers Need?
Technology skills quickly become obsolete. Concepts and theoretical knowledge have a much longer “half-life.” Managers need to master the what, when, who, and why of computerized decision support. Managers need less knowledge about the how-to of computerized decision support, analytics, and BI systems. The concept of decision support has broadened over the past 50 years to encompass a wide variety of information technologies that support decision-making. The basic philosophy of decision support is that technology and software positively impact decision-making. Managers must know much more about IT solutions than when they began their careers.
Analytics, BI, and DSS use sophisticated information hardware and software technologies, and therefore, managers need computing and software knowledge to understand such systems. In addition, there is an increasing need for managers to provide input to hardware and software choices. At a minimum, in today’s business environment, a manager needs to be able to operate the software environment of personal computing devices (e.g., a workstation, a portable computer, and a smart phone).
Our hardware and software environment is rapidly changing (i.e., new versions of Microsoft Office, new Google products, new hardware devices, and new intracompany web-based applications are introduced). In addition, managers often need to master software products relevant to the job. In some situations, it may be necessary to develop small-scale budgeting or cost-estimating applications in Excel or a product such as Crystal Reports. There is a growing need for “end user” development of small-scale DSS and preparation of special decision support and analytic studies.
Networks and enterprise-wide systems are expanding globally. Because managers and knowledge workers are the primary users of enterprise-wide DSS, managers must understand the possibilities and be involved in designing the systems. Managers need to develop the skills and knowledge to think about IT solutions, including defining a problem, formulating a solution, and implementing a solution.
Managers need to understand the benefits, costs, and risks of building a specific IT decision support capability. Decision support, analytics, and BI systems can solve problems and create new problems. Managers need broad knowledge of technology to help them make informed decision support implementation choices.
Computing and IT knowledge needs and skills are constantly evolving. We all need to learn continuously about new concepts and new skills. Some new decision support requirements build on previously learned materials; others force us to change our thinking dramatically.
Who Uses Computerized Decision Support?
Many people use computerized decision support for work and in recent years to aid in personal decision-making. Identifying the targeted or intended users for computerized decision support helps to differentiate the specific system. Knowing who does or will use a capability provides useful information about how the content and design of the application might or should differ. Let us review examples of job titles and occupations of targeted users for decision support, BI, and analytic systems.
In 1978, Keen and Scott Morton described six diverse systems and targeted user groups, including: (1) a DSS to help investment managers with a stock portfolio, (2) a DSS used by the president of a small manufacturing company to evaluate an acquisition prospect, (3) an interactive DSS used by product planners for capacity planning, (4) a model-driven DSS used by a brand marketing manager for making marketing allocations, (5) the geodata analysis and display system (GADS) used to redesign police beats, and (6) a DSS to explore and define alternative school district boundaries.
By 1996, Holsapple and Whinston identified many management users of decision support applications. For example, the management staff of the distribution department at Monsanto used a DSS for ship-scheduling decisions; a DSS helped managers with vehicle fleet-planning decisions; cargo planners used a DSS for scheduling ship unloading in Rotterdam; plant supervisors at Dairyman’s Cooperative used a PC-based DSS to optimize daily production planning; maintenance planners at American Airlines used a decision support application; and analysts and executives in the US Coast Guard used a document-driven DSS to help make procurement decisions.
Turban and Aronson also identified DSS used by staff for special studies. Staff at Group Health Cooperative used a data warehouse and statistical analysis tool to generate periodic reports and for monitoring key performance indicators, and staff at Siemens Solar Industries constructed a simulation model DSS of a “cleanroom” to explore alternative design options.
DSSResources.com has 52 decision support case studies that identify users, including managers, staff, customers, the general public, and workers in business, government, and not-for-profit organizations. Job titles of users include engineers, loan officers, salesmen, fire department commanders, examiners in the Pennsylvania Department of Labor and Industry, business and financial analysts, and emergency management professionals.
A web search identifies even more uses and users. For example, some medical doctors are using a web-based clinical DSS. DSS are used by judges, lawyers and mediators, farmers, and agricultural policy makers.
The US Marine Corps needed an application that allowed Marine Command staff to import, manipulate, and analyze terrain data relative to their operations. Road maintenance supervisors evaluated a maintenance decision support system during the winter of 2003 in Central Iowa. DSS are used for air traffic monitoring. Also, a DSS is used by staff to facilitate manpower planning for the US Marines. Military analysts use a financial data mart at the Military Sealift Command at the Navy Yard in Washington, DC. TIAA-CREF portfolio managers use a DSS for more than 160 billion US dollars of daily equity investment.
Fico.com cites many uses of predictive analytics by companies. The company’s website claims “Predictive analytics is widely used to solve real-world problems in business, government, economics and even science—from meteorology to genetics.” Managers and staff implement and use analytics and especially predictive analytics in credit scoring, underwriting, collecting past due accounts, increasing customer retention and up-selling, and fraud detection.
So who uses computerized decision support including analytics and BI systems? Managers, knowledge workers, and staff specialists in a wide variety of professions, occupations, industries, and disciplines. Decision support users include internal and external stakeholders of an organization. Ultimately, anyone who makes decisions and has access to a computer is a potential user of a computer-based decision aiding application.
What Is the History of Computerized Decision Support?
Some knowledge of the history of computerized decision support should help managers understand decision support and make better adoption decisions. This brief review of the evolution of decision support technology primarily touches on decision support innovations and successful projects (Figure 1.1). More details about decision support history are available at DSSResources.com.2
Figure 1.1. Decision support history time line.
First Generation Decision Support
We can trace the origins of computerized decision support to 1951 and the Lyons Tea Shops business use of the LEO I (Lyons Electronic Office I) digital computer. LEO handled accounts and logistics. Software factored in the weather forecast to help determine the goods carried by “fresh produce” delivery vans to Lyons’ United Kingdom shops.3 Lyons innovated, but the new DSS was not sufficient to help managers adapt to changing customer needs.
On election day November 4, 1952, a computer application was used to predict the US Presidential voting results from exit interviews, but news reporters were skeptical of the prediction and did not use it. A few years later, work started on the Semi-Automatic Ground Environment (SAGE), a control system for tracking aircraft used by North American Aerospace Defense Command.
The name SAGE, a wise mentor, indicated the decision support nature of the system. SAGE was a high cost, innovative real-time control, communication, and management information system (MIS).4 The technology was quickly obsolete, but continued to function until the early 1980s.
The pioneering work of George Dantzig, Douglas Engelbart, and Jay Forrester established the feasibility of building a wide range of computerized decision support applications. In 1952, Dantzig, a research mathematician at the Rand Corporation, implemented linear programming on an experimental computer to solve a variety of analytical problems. In the mid-1960s, Engelbart and colleagues developed the first groupware system, called NLS (oNLine System). NLS had on-screen video teleconferencing and was a forerunner to group DSS. Forrester was involved in building the SAGE data-driven system. In addition, Forrester started the System Dynamics Group that built complex computerized, quantitative models for decision support at the Massachusetts Institute of Technology Sloan School.
Prior to about 1965, it was very expensive to build large-scale information systems. From 1965 onward, the new mainframe technology like the IBM System 360 and timesharing technology made it practical and cost-effective to develop MIS in large companies. MIS focused on providing managers with structured, periodic reports derived from accounting and transaction systems.5 Technology developments stimulated decision support innovation and adoption.
Moving to the Next Generation
In the late 1960s, a new type of information system became practical, so-called model-oriented DSS or management decision systems. For his Harvard Business School doctoral research, Michael Scott Morton devised and studied a computerized management decision system. He studied how computers and analytical models could help managers make a key decision. Scott Morton conducted an experiment where marketing and production managers used a management decision system to coordinate production planning for laundry equipment. The decision system ran on a 21-inch cathode ray tube monitor with a light pen connected using a 2400 bits per second modem to a pair of UNIVAC 494 computer systems.6 The computer support improved decision-making and planning.
In 1971, Gorry and Scott Morton argued that MIS primarily focused on structured decisions and suggested that the information systems for semi-structured and unstructured decisions should be termed as DSS.7 The article initiated an academic subfield.
By the late 1970s, researchers were discussing both practice and theory issues related to DSS, and companies were implementing a variety of systems. In 1979, John Rockart published an article in the Harvard Business Review8 that led to the development of executive information systems (EISs). In 1980, Steven Alter published a framework for categorizing DSS based on studying 58 DSS. He identified both data-oriented and model-oriented DSS.9
Ralph Sprague and Eric Carlson’s book, Building Effective Decision Support Systems,10 explained in detail a DSS architecture framework: database, model base, network, and dialog generator. In addition, they provided a practical, understandable overview of how organizations could and should build DSS. By 1982, many researchers considered DSS as an established class of information systems.
In the early 1980s, financial planning and modeling systems became especially popular decision support tools. The software idea was to create a “language” that would “allow executives to build models without intermediaries.”11 Analytical models gained acceptance in business decision-making.
By 1982, managers and researchers recognized that DSS could support decision makers at any level in an organization. DSS could support operations, financial management, management control and strategic decision-making. Technology and conceptual developments had expanded the scope, purpose, and targeted users for computerized decision support.
Generation 3: Expanding Decision Support Technologies
The scope and capabilities of computing technology expanded tremendously with the advent of personal computers. Spreadsheets made analysis of data and model building easier and faster. Researchers developed software to support group decision-making using local networks.12 In 1985, Procter & Gamble built a DSS that linked sales information and retail scanner data. BI was embryonic as the decade changed, analysts were advancing a set of concepts and methods to improve business decision-making by using fact-based support systems. Companies were implementing briefing books, report and query tools, and EIS.13
In the early 1990s, Bill Inmon and Ralph Kimball actively promoted using relational database technologies to build DSS. Kimball was known as “The Doctor of DSS” and he founded Red Brick Systems. Inmon became known as the “father of the data warehouse” and founded Pine Cone Systems. From Inmon’s perspective, a DSS involved “data used in a free form fashion to support managerial decisions.”14 The DSS environment of the 1990s contained only archival, time variant data. Both data warehousing and On-Line Analytical Processing (OLAP) technologies improved data-driven DSS.15
A major technology shift had occurred from mainframe and time-sharing DSS to client/server-based DSS at about this time. Vendors introduced desktop, personal computer–based OLAP tools. Database vendors recognized that decision support was different from online transaction processing and “started implementing real OLAP capabilities into their databases.”16 By 1995, large-scale data warehousing, a convergence of technologies and systems, and the possibilities for distributed DSS created by the World Wide Web stimulated innovation and created a renewed interest in computerized decision support and BI.
By 2000 possibilities for real-time decision support became more realistic. In 2012, developments in analytics, operational, and mobile BI continue to stimulate decision support innovation in organizations. Computer support for decision makers continues to expand and improve.
What Is the Theory of Computerized Decision Support?
Past practice and experience often guide computerized decision support development more than theory and general principles. Some developers have concluded each decision situation is different so no theory is possible. Some academics argue that we have conducted insufficient research to develop theories. For these spurious reasons, a theory of decision support has received limited discussion in the literature.
Nobel Laureate Economist Herbert Simon’s writings provide a starting point for a theory of decision support. From his classic book, Administrative Behavior,17 are derived three propositions:
Proposition 1: If information stored in computers is accessible when needed for making a decision, it can increase human rationality.
Proposition 2: Specialization of decision-making functions is largely dependent upon developing adequate channels of communication to and from decision centers.
Proposition 3: When a particular item of knowledge is needed repeatedly in decision-making, an organization can anticipate this need and, by providing the individual with this knowledge prior to decision-making, can extend his or her area of rationality. Providing this knowledge is particularly important when there are time limits on decisions.
From Simon’s article18 on “Applying Information Technology to Organization Design” are three additional propositions:
Proposition 4: In a postindustrial society, the central problem is not how to organize to produce efficiently, but how to organize to make decisions—that is, to process information. Improving efficiency will always remain as an important consideration.
Proposition 5: From the information processing point of view, division of labor means factoring in the total system of decisions that need to be made into relatively independent subsystems, each one of which can be designed with only minimal concern for its interactions with the others.
Proposition 6: The key to the successful design of information systems lies in matching the technology to the limits of the attention of users. In general, an additional component, person, or machine, for an information-processing system will improve the system’s performance when it:
1.has a small output in comparison with its input, so that it conserves attention instead of making additional demands on attention.
2.incorporates effective indexes of both passive and active kinds. Active indexes automatically select and filter information.
3.incorporates analytic and synthetic models that are capable of solving problems, evaluating solutions, and making decisions.
In summary, computerized decision support is potentially desirable and useful when there is a high likelihood of providing relevant, high-quality information to decision makers when they need it and want it.
What Influences Adoption of Decision Support?
Change and innovation continue to be related to computerized decision support. Some managers seem quick to purchase new technologies and try out new capabilities, others are slower to adopt an innovation. Adopting a new technology is the first step in building a new capability and gaining technology acceptance in an organization. Managers can adopt an innovation, but intended users may not accept the new technology. So why are some managers quick to adopt and others delay? Leaders and laggards are common among individuals and organizations in the adoption of technology. In the late 1950s, sociology researchers19 proposed a technology adoption lifecycle model. Moore in his 1991 book Crossing the Chasm20 proposed a variation of the lifecycle that identified a significant gap in time from when early adopters and more pragmatic managers buy technology. This gap is caused by the perceived disruption from the innovation.
Decision support technologies are often disruptive innovations. Also, some decision support innovations are quickly obsolete or “faddish.” Some decision support applications are purchased and adopted, but quickly become dated or revised. Also, underused and poorly accepted software known as “shelfware” was bought and now sits on shelves in IT departments.
Buying or building a new decision support capability often is a significant decision. The new system may be a one-time purchase or an ongoing commitment of resources. Some general reasons why one company is often an early adopter of significant decision support innovation and another is often slow can be identified. Reasons include the availability of resources, risk propensity, knowledge of the technology, the culture, and senior management characteristics. The interaction of these factors is complex and that hinders a researcher or consultant trying to understand a particular company’s situation.
Some more detailed reasons for slow adoption of innovative technologies are (1) mistrust between IT and business executives, (2) lack of data quality and too many data sources, (3) delayed infrastructure projects, (4) IT staff are poorly trained, and (5) new technologies are confusing and poorly understood. This list21 of reasons suggests changes managers could make to encourage faster adoption of new technologies.
Adoption of decision support technologies should be a pragmatic, rational decision. Practical considerations should be the most important factor when managers adopt decision support technologies for their organizations. In some situations, organizations have benefited from early adoption of technology, but examples of waste and negative disruptions are also common. In general, managers should cautiously adopt a potentially disruptive decision support technology.
What Is Typical of Modern Decision Support?
Modern decision support started to evolve in about 1995 with the specification of HTML, expansion of the World Wide Web in society, and the introduction of handheld computing. Web 2.0 technologies, mobile integrated communication and computing devices, and improved software development tools have revolutionized decision support user interfaces. Additionally, company decision support data stores are growing in size and contain varied and extensive “big data.” The fourth generation of computerized decision support is maturing.
We need to recognize that analytical systems often have multiple capabilities. Attributes of contemporary analytical and decision support systems typically include the following:
1.Access capabilities from any location at anytime.
2.Access very large historical data sets almost instantaneously.
3.Collaborate with multiple, remote users in real-time using rich media.
4.Receive real-time structured and unstructured data when needed.
5.View data and results visually with excellent graphs and charts.
Change is accelerating. Where is the current leading edge and what technologies are on the horizon that can be exploited to build more advanced decision support? Grid computing and parallelism seem particularly interesting. Both speech generation and recognition can be exploited; and stereographic displays and wearable computing technologies are improving.
Modern BI and DSS have more functionality than systems built prior to the widespread use of the Internet and World Wide Web. Managers are choosing to implement more decision automation with business rules and more sophisticated knowledge-driven decision support. Current systems are changing the mix of computing and decision-making skills needed by managers in organizations. There is a shortage of managers and analysts with the expertise to conduct analyses and use decision support capabilities.22 Managers are realizing that better computerized decision support is crucial for competing in a global business environment.
Summary
Decision support and BI systems serve varied purposes and targeted users and are implemented with a variety of technologies. Analysis and interpretation of data from computerized sources is becoming a very important skill for managers. With current technologies, computerized applications can and do support a wide range of decision-making tasks. Contemporary decision-making environments create a need for more, and better, computerized decision support.
A general understanding of decision support history provides a context for understanding modern decision support. The first generation systems were on mainframe computers, but the SAGE defense system provided sophisticated real-time decision support. By the 1980s, decision support technologies broadened the possibilities for computerized decision support to include collaboration and BI. Since approximately 1995, computer support for decision makers has significantly expanded and improved.
Nobel Laureate Herbert Simon’s ideas provide a theoretical rationale for building computerized DSS and using analytics. Computerized decision support and analytics can expand the rationality of decision makers. Managers should adopt decision support capabilities when it is likely that significant benefits will be realized. Modern decision support assists decision makers and helps them make better decisions by exploiting new technologies and expanding capabilities. Modern decision support helps managers.