There are basically two broad categories of forecasting techniques. These are:
1. Qualitative Techniques:
These techniques are primarily based upon judgement and intuition about the environment and are used especially when sufficient quantitative information and data is not available so that complex quantitative techniques cannot be used.
Under certain situations, qualitative judgments about future are more reliable than quantitative conclusions because the quantitative methods are based on the analysis of the past data and its trends which may or may not remain the same. Secondly, quantitative techniques follow a certain pattern and do not provide for accommodating any unexpected occurrences.
Some of the widely used qualitative approaches to forecasting are:
This is the method by which the relevant opinions of experts are taken, combined and averaged. For example, the managers of various divisions could be asked about the profitability of introducing a new product and a judgement made on the basis of their opinions.
These opinions could be taken on an individual basis or there could be brainstorming group sessions in which all members participate in generating new ideas that can later be evaluated for their feasibility and profitability.
This method is fast, less expensive and does not depend upon any elaborate statistical data or methodology and brings in specialized viewpoints. Its main disadvantage is that it is based upon subjective opinions which may be overly optimistic if the business conditions are good and may be overly pessimistic when economic conditions are not favourable.
This approach involves the opinions of the sales force and these opinions are primarily taken into consideration for forecasting future sales. The sales people, being closer to the consumers, can estimate future sales in their own territories more accurately. Based on these and the opinions of the sales managers, a reasonable trend of the future sales can be calculated.
These forecasts are good for short-range planning, since sales people are not sufficiently sophisticated to predict long-term trends. This method, also known as the “grass roots” approach lends itself to easy breakdowns of product, territory, and customer and so on, which makes forecasting more elaborate and comprehensive.
This method involves a survey of the customers as to their future needs. This method is especially useful where the industry serves a limited market. Based on the future needs of the customers, a general overall forecast for the demand can be made. The major problem with this method is that the future “needs” do not necessarily mean future “commitments” to buy, since needs may change depending upon the circumstances. However, the method is fairly reliable where the target market is small such as buyers of industrial products and where the customers are knowledgeable and cooperative.
The Delphi method, originally developed by Rand Corporation in 1969 to forecast military events, has become a useful tool for other areas also. It is basically a more formal version of “jury of executive opinion” method. A panel of experts is given a situation and asked to make initial predictions about it. On the basis of a prescribed questionnaire, these experts develop written opinions. These responses are analyzed and summarized by a central coordinator and submitted back to the panel for further consideration, evaluation and refinement.
All these responses are anonymous so that no member is influenced by other’s opinions. This process is repeated until a consensus in obtained. This method is very useful where either the past patterns are not available or where the past data is not indicative of future events and the issues are general in nature such as future energy needs, possible after-effects of a nuclear war and so on.
2. Quantitative Techniques:
These techniques involve mathematical and statistical analyses of data banks, which is primarily the information related to past activities. These techniques are fairly sophisticated and require experts in the field to use them. “Time Series Analysis” is a popular statistical forecasting technique. It extrapolates the past trends into the future. “Econometric Models” for forecasting are more complex in nature and involve inter-relationships of many variables tied together in a mathematical model.
The sales, for example, are not only a function of time but also depend upon many other variables such as changes in personal disposable income, credit availability, and so on. Complex computer models simulate future events based on probabilities and multiple assumptions. “Statistical Surveys” use statistical analysis of opinion polls and attitude surveys to predict such variables as future consumer tastes, employee preferences, and political choices and so on.
The major disadvantage of using quantitative techniques is that the conclusions derived from quantitative models are only as good as the assumptions and judgments made about the variables that are put into the model. Faulty assumptions will yield inaccurate results.
No matter what model or method is used, forecasting basically rests on human judgement. Accordingly, forecasts should serve as guidelines and not as indicators of certainty. Henry Albershas summarized his assessment of forecasting:
“A successful forecast is something of a miracle and often occurs for the wrong reasons. However, it should not lead to the assumption that nothing has been accomplished. There are some good “rule of thumb” forecasts. A part of the problem is that too much is expected from forecasting. People want more precise answers than are possible in an environment characterized by uncertainty.”