Decision Support Systems

Decision

In life, and more so in business we are constantly in the process of making decisions. We make good decisions and rejoice; make bad decisiosns and despair. Sometimes we agonise over making decisions anddelay to the point that it is no longer required. Other times we make snap decisions and wonder, "did we act too quickly?".

A decision is a choice that we make, from a number of competing options. We would like our decision to be rational and to be supportive of our objectives. Also, we would like make decisions quickly as time is of the essence.

A decision support system is a tool that helps you make good decisions and make them fast.

From Data to Wisdom

Businesses have a lot of data. Using that data a decision has to be made. Let us briefly look at the DIKW pyramid and understand it in the business context.

Data in a broad sense is a collection of facts, for example, bank transactions.

Information makes sense of the data, for example, summary of bank transactions by heads of expenditure.

Knowledge is combining of information to create a conceptual model that not only validates the information but also provides a means of extrapolation,for example, how is the expenditure related to production volumes.

Wisdom, shall we say is the discretion to seek knowledge and a discerning use of that knowledge.

Decision Support

A rational decision making process, of course, requires that data is converted into information. If it is sales data that we are talking about then the data must be processed to yield statistics that can inform decision making.

Generally, deriving of knowledge is domain-dependent and requires the application of heuristics to yield some answers. It may be the case that sales information is being used to make purchase decisions. The statistics (sales information) will then be used as inputs to a "resource planning model" to obtain outputs like phased requirements of goods.

If our models of the world were trustworthy then the output above could be treated as knowledge. However, business people know that models are imperfect. Also, they are prone to error if the underlying data is corrupt. It is safer, indeed essential, that the outputs are treated as somewhat plausible conclusions but a final decision is made by a human. That is, we do everything possible to convert data to information, to use information to feed an algorithmic model, so that a review can be undertaken quickly. Decision-making is supported but not taken over by machines. Thus, a decsison support system sits at the point where information is converted into knowledge.

Infrastructure

A general problem that organisations face in deployment of decision support systems is that they are typically proprietary software packages. These must be installed (if not cloud-based) and means of interfacing found to enable transfer of data from other business systems.

The cost of managing the information technology infrastructure is high, and, prohibitively high for small and medium enterprises. There is, however, an easier, affordable and better way of satisfying the need for decision support systems. It is to create them using the ubiquitous MS Office Suites - the functionality of Excel, Access and VBA is surprisingly wide and powerful to allow for creation of powerful decsison support tools. What this means is that you do not have to purchase new hardware and software. Moreover, you work with a familiar user interface - cutting down your learning phase to almost nothinhg.

Collaborative Systems

A great advantage of business-wide systems is that transactions are instantly recorded and shared across the system with relevant users. However, when we are using the data generated by business-wide systems we need to take a snapshot at a point in time and use that frozen data to inform our decision-making. The data extraction thus needs to on-demand (or at specified intervals) and this simplifies the interface between the the business system and decision support systems.

If a number of decision support systems are employed then they can be used to foster collaboration between employees. For exapmple, if a production planner uses a system to decide on a manufacturing plan then that plan can be transmitted across to a scheduling support system and the production planner can then decide on the exact production schedule.

Since all decison support systems are fetching data from the same source the personnel using the systems are working with the same data, adding value in terms of the decisions that they make and making this information available to others, without effort, so that others can build on the work already done.

The result is: good decisions, with minimal effort, at the right time. Therefore, better business.