4 Forecasting Methods for Supply Chain Management

Listed below is a brief description of four forecasting methods that may be used to generate a demand forecast for a product. They are listed in order of effort required to product the forecast.

These forecasts are typically come from an expert’s ‘gut feel’. Being familiar with both products, the industry and customer preferences, an expert’s opinion is usually useful when new products are being introduced into the market where there is no historical performance data to fall back on.

Time series forecasts rely on historical demand in order to predict the future demand. There are a variety of computational methods that can be used to facilitate this process. Usually, this method is ideal for “staple stock” items, or items that have a generally defined historical pattern that does not change radically from one year to the next.

Causal forecasting attempts to correlate one or more variables to the demand forecast. For example, pricing warranty levels, disposable income indicators, etc. may be used to determine the demand for new plasma TV’s. This method requires a higher level of sophistication in computational effort and may require special computer programs.

This method requires special purpose computer programming. Its main benefit is that it permits the ability to answer multiple ‘what-if’ scenarios. For example, what would the impact on the forecast be if prices are increased or if disposable income decreases? By evaluating the sensitivities of the model’s inputs it is possible to develop a more robust forecast.