Understanding Demand Spreading
Demand spreading refers to the process of spreading the demand forecast quantity for a given date or date range across a specified time period. In this process, the planned dates and quantities in the Demand Detail records are consolidated, spread across a week in daily buckets, and then distributed into aggregate weekly or monthly buckets that are required for forecasting and planning.
This step is necessary because the demand information transmitted by customers to the supplier can vary. Customers can define demand in daily, weekly or monthly quantities. Customers may also consider different days as starting days for a week, or they might define a month in different ways. This table indicates how the demand information can vary from customer to customer:
Customer |
Month |
Actual Time Period |
Quantity |
|---|---|---|---|
Customer A |
June 2001 |
June 4 - July 1 |
18,000 |
July 2001 |
July 2 - August 5 |
19,000 |
|
August 2001 |
August 6 - September 2 |
20,000 |
|
Customer B |
June 2001 |
June 1 - June 30 |
10,000 |
July 2001 |
July 1 - July 28 |
12,000 |
|
August 2001 |
August 1 - August 31 |
14,000 |
|
Customer C |
June 2001 |
June 4 - July 1 |
50,000 |
July 2001 |
July 2 - July 29 |
50,000 |
|
August 2001 |
July 30 - September 2 |
55,000 |
To accommodate the production schedule at a given branch/plant, the supplier has to accept these varying planning schedules and transform them into a common forecast.
When you process a batch of demand records through the Create Demand Schedule program (R40R010) to transfer planned demand into the Forecasting system, the system deletes any existing detail and summary forecast records (F3400 and F3460) for any combination of item, branch, customer and forecast type that occur in the batch. After the deletion, the system processes the batch one record at a time to load a new forecast. If a time period includes multiple demand records, the final forecast for this period contains the sum of the forecasts from all the overlapping records. The resulting forecast is considered the latest forecast for the combination of item, branch, customer and forecast type.
After removing the previous forecast records, the system automatically spreads and consolidates the demand quantities across the specified time period. In the process, the system invokes a demand spreading template, if one is available, and divides the demand into daily buckets, based on the percentages specified in the template. If no template is available, the demand is spread evenly across the workdays of the date range.
After the demand is broken down into daily values, the system rolls up these daily values into a forecast based on the date ranges defined for the forecast buckets and stores the records in the F3400 table and the F3460 table. The forecast buckets can be set up for weekly or monthly forecasts.