Forecast Family
A forecast family is a grouping of items that are generally similar, considering only demand. For this reason, a forecast family can be used to forecast demand for specific items and be used as a forecast object for them.
Description
Forecast families are used to make manual forecasting more effective. Forecasts only need to be made for a limited number of forecast families instead of for every item within each family. Item forecasts are then automatically calculated from the forecast for the family using the family structure. The family structure describes the relation of each member item to its proportion of the total forecast. The family structure is expressed in percent, which is determined from historic demand.
Forecast families can also be used to distribute the total forecast for items to different warehouses. Forecasted demand per warehouse can be calculated using the family structure since it specifies the percentage demand for each item per warehouse. These percentages are also determined from historic demand.
The proportions for different items and the proportions for different warehouses per item can also be combined, as well as be varied from one period to the next. This allows different percentages to be used for the same item and warehouse for different forecast periods. The percentages per period can be changed when products in a forecast family are phased in or out, or when sales are started or ended from a warehouse.
Example 1
Assume items A, B, and C are included in forecast family FF. Mean demand for the three items over the last two years is 750, 230 and 560, respectively for a total of 1,540. The proportion of demand for each item against the total demand for the family is 750 / 1.540 = 49%, 230 / 1.540 = 15%, 560 / 1.540 = 36%. These proportions would then be entered in a structure for use in forecasting demand for each item.
For January 1996, a total demand of 140 is forecasted for forecast family FF. The family structure can then be used to forecast January demand for each item as follows: item A - .049 x 140 = 69; item B - 0.15 x 140 = 21; and item C - 0.36 x 140 = 50.
Example 2
Item P is sold from three warehouses, L1, L2 and L3. The previous year’s demand for the product was 2,370 total, with 650 from L1, 970 from L2 and 750 from L3. The proportion of the total demand for the product at each warehouse is as follows: 650 / 2370 = 27%, 970 / 2370 = 41%, 750 / 2370 = 32%, respectively. These proportions would then be entered in a structure for use in forecasting demand for each warehouse.
For January 1996, a total demand of 260 is forecasted for item P. The family structure can then be used to forecast January demand from each warehouse as follows: L1 - 0.27 x 260 = 70; L2 0.41 x 260 = 107; 0.32 x 260 = 83.