Demand patterns for products that sell an average of ten times a month are different than those for products that sell an average of ten times a year. Items that are seasonal have a greater demand at certain times of the year than others.
The system calculates the demand history based on hits. The system defaults to the following conditions:
If a product has 25 or more hits in 91 days or less, then the system uses the last 91 days of sales history to determine demand.
If a products has less than 25 hits in 91 days, but more than 25 hits in the last 365 days, then the system uses the number of days back to the 25th hit to determine demand. For example, if the 25th hit occurred 180 days ago, the demand history uses the last 180 days.
Note: If more than one hit occurred on the date of the 25th hit, all hits on that date are included.
If a product has less than 25 hits in the last year, then the system uses 365 days for the demand history.
The more frequently sold items are statistically more predictable and demand calculations should, when possible, use shorter periods of time to reflect recent sales trends.
Seasonal products are those products that sell more often at certain times of the year. For example, heating products such as furnaces, sell more often during winter months. Seasonal forecasting lets you use the winter months only as the forecast period to determine appropriate demand.
When forecasting the demand of a seasonal product in September 2000, the system checks from September 1999 forward to gauge the winter month demand.
If there are 10 or more hits in a 45-day period, the system uses those 45 days of the product history for the calculation. For example, today is September 1, 2000. The system checks the history for Product A and finds 17 hits from September 1, 1999 to October 15, 1999. This 45-day period is used for the calculation.
If there were less than 10 hits in a 45-day period, but more than 10 hits in a 122-day period, the system uses the date of the 10th hit as the end date of the period for the calculation. For example, today is September 1, 2001. The system checks the history for Product B and finds that 10 hits occurred between September 1, 2000 and November 1, 2000. This 60-day period is used in the calculation.
If there are less than 10 hits in a 122-day period, the calculation is based on the full 122-day period. For example, today is September 1, 2000. The system checks the history for Product C and find that only 6 hits occurred between September 1, 1999 and January 1, 2001. The 6th hit occurred on December 15, 1999. The full 122-day period is used in the calculation.