Causal Forecast Method
This forecast method is a BOM based forecasting of parts at all levels in the BOM. It also forecasts failure rates for parts using historic demand and installed base. Global failure rate is calculated based on installed base and causal factors, while higher weight is given to recent installed base and failures using exponential smoothing techniques.
 |
The Causal forecast method calculates the forecast for the top-most part only. The population and/or causal values of replaced parts are rolled over to the top most part. Alternates are treated as standalone parts.
Rollout amounts, failure rates, contracts, and causal values involved in causal forecasting must have start/end dates that fall on the first day of the slice. All date fields must fall on the first day of the slice. For example, in a monthly database, failure rates with Start Dates or End Dates cannot fall in the middle of a slice.
|
This forecast method uses the following calculations:
Current Effective Total Population * Sum(Causal Value * Failure Rate * Weight Factor)
(Product Rollout Amount * BOM Quantity * Attach Rate) * Sum(Causal Value * Failure Rate * Weight Factor)
 |
• Causal Value and Weight Factor are optional
• If there is no record in part_causal_type for a part, then Causal Forecast = (Product Rollout Amount * BOM Quantity * Attach Rate) * Failure Rate
• Product Rollout Amounts can be entered through the user interface (see Setting Product Rollouts).
• Product BOM quantities can be entered through the user interface (see Setting Product BOMs for Causal Forecasting).
• Causal Values can be entered through the user interface (see Setting Causal Values)
• Failure Rates/MTBF can be entered through the user interface (see Setting Failure Rates/MTBF)
• Data download and user interface import of contract types, contracts, causal forecast product BOMs, causal forecast product rollout, causal types, part causal types, and causal values is supported
|
|
|
Each forecast method has constraints, called best fit rules that may make it ineligible for forecasting.
|