Types of Forecasting
Forecasting techniques can be classified into three broad categories qualitative time-series analysis and causal relationship forecasting. Qualitative techniques are subjective or judgmental in nature and are based on estimates and opinions. Such techniques are used primarily when there are no data available. Time-series analysis the main focus of this chapter is based on the idea that data describing past demand can be used to predict future demand In other words the time-related trends that generated demand in the past will continue to generate demand in the future. Causal relationship forecasting on the other hand assumes that demand is related to some underlying factor or factors in the environment and that cause and effect relationships are at work. Time series analysis typically u ed in short range situations, such as forecasting worker requirement for the new week. Causal relationship forecasting is usually used for longer term issues u h as electing a site for a retail operation. Exhibit 9.2 briefly describes some of the different varieties of the three basic types of forecasting models. In the chapter we present the three time senescence analytic methods listed in the exhibit and the first of the causal relation hip force sting technique.
Exhibit 9.3 show comparison of the strengths and weaknesses of these different forecasting methods. The moving average and exponential smoothing method tend to be the best and easiest techniques to use for short-term forecasting with little data required. The long-term models are more complex and require much more data. In general the short-term models compensate for random variation and adjust for short term change (such as consumers' response to a new product). Medium term forecasts are useful for seasonal effects and long term models identify general trends and are especial useful identifying major turning points. Which forecasting model or model a firm hold n on several factor including (a) forecasting time horizon, (b) data avail ability (c) accuracy required (d) size of the forecasting budget, and (e) availability of qualified person
capacity in the form of front-line workers must be available when and where customers require it. However while forecasting can provide managers with future information that will allow them to run their operations more effectively and efficiently, managers also must recognize that forecasts are not perfect. Inaccuracies in forecasting occur because there are too many factors in the business environment that cannot be predicted or controlled with certainty. Rather than search for the perfect forecast, it is-tar more important for managers to 'establish the practice of continually reviewing these forecasts and to learn to live with their inaccuracies. This is not to say that we should not try to improve the forecasting model or methodology, but that we should try to find and use the best forecasting method available within reason.
In this respect. it is important to note that the cost of obtaining small improvements in forecasting accuracy is very high after reasonable forecasts have been developed as illustrated in Exhibit 9.1. The goal of this chapter is to present an introduction to several different forecasting technologist and models (both qualitative and quantitative) that are commonly used in business, recognizing that additional and more sophisticated forecasting techniques and models are available for people seeking more in depth knowledge in this area. We address primarily time series techniques and causal relationships including a discussion of the sources of errors and their measurement.