Statistical Forecast Model Selection
SyteLine supports multiple statistical forecasting methods that analyze historical demand patterns to project future requirements. Choosing the right model depends on your demand characteristics: stable demand suits moving average, trending demand needs linear regression, and seasonal demand requires seasonal decomposition. The Forecast Parameters form configures model selection and evaluation criteria.
- Moving Average model for stable-demand items with low variance (coefficient of variation <0.3)
- Exponential Smoothing with alpha parameter tuning for items with recent demand trend shifts
- Linear Regression for items with sustained upward or downward demand trends over 12+ months
- Seasonal decomposition model applied to items with repeatable quarterly or annual demand patterns
- Best-fit algorithm in SyteLine automatically selecting the model with lowest forecast error per item
Forecast Consumption and Demand Management
Forecast consumption determines how actual customer orders reduce (consume) the forecast quantity in each planning period. SyteLine's consumption logic uses backward and forward consumption windows to prevent double-counting demand when orders arrive earlier or later than forecasted. Misconfigured consumption creates artificial demand spikes that trigger unnecessary MRP actions.
- Backward consumption days allowing actual orders to consume forecast from prior periods
- Forward consumption days permitting orders to consume forecast from future periods
- Forecast consumption logic: Customer Order consumes forecast in the order's scheduled ship period first
- Unconsumed forecast treatment: remaining forecast becomes independent demand for MRP after consumption
- Demand fence configuration separating frozen schedule zone from forecast-driven planning zone
Forecast Accuracy Measurement and Improvement
Forecast accuracy should be measured and reported monthly using MAPE (Mean Absolute Percentage Error) or tracking signal metrics. SyteLine stores historical forecast vs actual data that enables accuracy analysis at the item, product family, and planner code level. Continuous improvement requires regular forecast review meetings between sales, planning, and production teams.
- MAPE calculation per item and product family using SyteLine forecast history vs actual shipments
- Tracking signal monitoring to detect systematic forecast bias (consistently over or under forecasting)
- ABC classification driving forecast method selection: A items get statistical + sales input, C items get statistical only
- Monthly forecast accuracy report by Planner Code identifying items needing parameter adjustment
- Forecast override audit trail in SyteLine tracking manual adjustments and their impact on accuracy
Try it here, free
SyteLine Change Request Form Template
A change control form for SyteLine configuration, customization, patch and integration changes with impact analysis and rollback plan.
Improve your forecast accuracy in SyteLine -- our demand planning consultants tune models that reduce inventory waste.