The authorised IBM version of what decision support systems, activities are all about With the launch of the Data Interpretation System from Metaphor (CI No 1,180) and new versions of Application System and Personal Application System, IBM thought it was time to clarify where its decision support products fit in with each other – explaining that executive support systems are separate from, albeit closely related to, office and decision support systems. Executive support systems, it says, provide access for senior staff to computer-based systems for any business function. Their distinguishing characteristic is that they give access to information or tools in a way that matches the executive’s distinctive needs. Hence Executive Decisions/VM (CI No 1,182), which is also closely related to OfficeVision Family and to AS 2. But before getting on to IBM’s perception of its own products, some general (mainly IBM) words on what decision support is all about. Necessary characteristics A key element of any decision support system, says IBM, is its ability to retrieve data. Customers require access to a broad range of internal and external business data. Data may be personal, departmental, or corporate level. Data formats include both IBM and non-IBM data as well as relational and nonrelational structures. Non-IBM data includes OEM databases and public data banks, such as Dow-Jones. Users need to be able to select only the required data, in the desired sequence, and to have that data logically combined from multiple tables or databases to create the desired data view. Most decision support activities are performed upon existing data, but the user must have the capability to define, create and maintain personal data that is used in decision-making processes. In some cases, it may be desirable for the decision support user to modify departmental and other databases that go beyond personal use. In addition, the decision support system must provide data validation, audit and security functions to ensure integrity of the data. To begin to make effective use of the data, users need to be able to create logically new information from existing data. The system must allow new columns of data to be created by the application of simple calculations and summary functions. For example, says IBM, a user may want to know the average number of hours worked by an employee within a department. Another user may want to calculate a Salary by adding Base Salary to Commission. Simple arithmetic operations such as addition, multiplication as well as basic summary functions such as sums, average, and running totals, are supplied by decision support data manipulation functions. Once the data is in usable form, there are more specialised tools and processes to perform analysis to assist decision-making such things as calculations and defined formulae; spreadsheet; modelling with support for What if scenarios; statistics for applications such as forecasting; project management with the tools to plan and schedule resources, apply risk analysis to calculate probable completion dates, and analyse the impact upon enterprise productivity and profitability; linear programming to find the best solution that both satisfies all the constraints in a problem while meeting the user’s objectives; graphical presentation tools; communications; and facilities to define, save, and run user-created tasks on a repetitive basis. And it needs to be accessible and easy to use, other necessary characteristics including flexibility, a fast response time, availability to those who need to use it, and immediacy of results. The target users of decision support At whom are decision support systems aimed? Anyone in an organisation that makes decisions that affect their jobs – a brand manager uses decision support the better to understand the competition, the effect of promotional schemes upon sales, and so forth. A school administrator uses decision support to help plan and project teachers’ salaries for the next term. IBM divides such users into two major groups – primary user
s and supportive users. Primary users are the individuals who use the decision support and derive direct benefit from it – typically the decision makers in an organisation and the staff that assist them. They understand the problem to be solved, have the tools to provide a solution, and use the tools themselves to solve the problem. The brand manager and school administrator in the examples are primary users – but so are secretaries and clerks that operate the decision support on behalf of others. Supportive users simply enable primary users to make more effective use of the decision support system – helping them to use it, creating applications, organising the database in an appropriate manner, and providing the means for the qualified people to access it. The distinguishing features of decision support systems What distinguishes decision support from other business applications? A priori that it is personal. Decision support is an activity that implies personal business use, says IBM, in other words the operator of the decision support system is the one who receives benefit personally – in contrast to a batch or production system, which is generally operated by a third party who may have no personal interest in nor benefit from the results. Data is central to any decision support system says IBM rather obviously, because it is the data transformed to usable information that assists the user in making decisions. The data itself is characterised by being extracted from any number or kinds of sources, whether it be local or remote, departmental or corporate. And decision support is used primarily for non-operational applications. Much decision support activity involves unanticipated results and use of the system in a flexible, non-structured manner. Decision support is not used for transaction-processing systems, nor is it generally suited for multiple concurrent data update. A decision support is also characterised by supporting two modes of operation, ad hoc and repetitive, IBM reckons, adding that most users will operate a system in both modes at different times – the manufacturing floor manager queries the parts database to determine whether a required item is in stock – an ad-hoc task; the business analyst has to run the sales analysis and procedure each month to present to management or the board – a repetitive task.
