Leading Indicators Forecast Method
 |
If you are not already using Leading Indicators, it is recommended that you use Causal Forecasting functionality.
|
This forecast method creates a forecast using projections of installed part quantities multiplied by part usage/failure rates. It can be more accurate than demand-based forecasting for new parts, aging parts, or moving products where history is not a good indicator of future use. The application has the ability to create an estimate from a similar part, assuming a similar part exists. Using this method is predicated upon knowing the following:
• Where (the install site) and how much product (the install base) is installed
• What parts are included in the product BOM (bill of materials)
• Predicted usage or failure rate of each part
• Additionally, if there is to be any strategic planning required, then you must know how the install base is covered and by whom
The number of parts installed in the field by install site are counted by using the BOM (bill of materials) to product relationship; then that value is multiplied by the number installed by the usage rate. This produces an estimate of the number of failures as a product of the number of parts installed. By identifying the location where the parts are installed, the appropriate number of service parts can be stocked in the location that supports the installed products. Additionally, by using forward estimates of changes in the install base, the forecast can adapt to changes in the install base and actively adjust service stocks accordingly.
There are several ways to determine the usage rate. If MTBF data exits in the host data system, then it can be converted to usage rate during the data transfer process. The user may also provide an initial usage rate from some other source. If there is demand history and the install base data is available, the application can suggest an initial usage rate. Once an initial usage rate is determined, the application can be configured to update the usage rate for every slice based upon actual demand.
Usage rate can be calculated using a weighted average method where previously calculated usage rate is weighted 80% and the current slice's usage rate is weighted 20% (by default). This provides a smooth transition from one usage rate to another and improves the accuracy of the forecast. The usage rate cannot be accurately calculated if there is insufficient demand history. Usage rate requires one full month of non-zero install base and non-zero central demand history. If you do not have this history, you will be prevented from calculating usage rate. You will be notified if there are problems with the data driving the leading indicators forecast.
|
|
Each forecast method has constraints, called best fit rules that may make it ineligible for forecasting.
|