A distribution curve is a time series that specifies the proportion of a total forecast or total demand covered by a forecast period. This is always stated in percent.
Distribution curves are only used for manual forecasting and can be used mainly in two areas.
First, they help make manual forecasting more effective. Entering manual forecasts is simplified so only annual forecasts need to be entered, while period forecasts are done using a distribution curve. The annual forecast is automatically distributed per forecast period according to the distribution curve.
Second, they help to arrive at a specified variation of forecasted demand per period. For example, this can be used to change demand for special circumstances, such as phasing in new products or phasing out old. In addition, this can help take business cycle variations into account or predictable effects from specific marketing efforts.
Assume an item with estimated total forecast for the next year will be phased out and replaced by a new product over a period of six months from January 1. The distribution for the old product is set as follows.
|January||50 %||April||7 %|
|February||25 %||May||5 %|
|March||10 %||June||3 %|
An annual forecast of 200 will then be distributed so that the January forecast is 100, February is 50, March is 20, April is 14, May is 10, and 0 for June.