Stability and Responsiveness:

Generally the actual demand for a product varies randomly from period to period. Such variations may hide an underlying pattern because of a long-term growth trend or seasonal effects. A good forecasting system should smooth out such random variations so that subsequent forecasts are not unduly affected by such irregular changes. In doing so the system displays stability.

At the same time, the forecasting system is also expected to be sensitive to the fluctuations observed to reveal any genuine changes that generate demand for the product. In other words, it is desirable that the forecasting system should be responsive so as to sense a rise in the average level of demand due to a recent market expansion rather than smooth it out.

We must be clear that stability and responsiveness are two opposite criteria, and a compromise is necessary. It should be possible to operate the forecasting system in two modes - in stable mode when the changes in environment are slow, and in responsive mode when there are significant developments that affect the generation of demand.