Lead-time forecasting should produce a probability distribution—not a single ERP parameter
A fixed lead time hides uncertainty, calendar effects and process-stage variability that directly affect inventory and service decisions.
Lead time is rarely one stable number. Treating it as a single ERP parameter creates false precision.
Observation
Recent planning commentary argues that Lead Time Forecasting should represent uncertainty through probability distributions rather than point estimates.
Operational implication
A fixed lead time can understate safety stock, distort replenishment timing and hide the tail risk of late supply.
Decision architecture
The organization should define which distribution is authoritative, who approves parameter changes and how lead-time uncertainty is combined with demand uncertainty.
Data requirements
Order dates, receipt dates, calendars, process stages, supplier lanes and exception history must be consistent and traceable.
What should remain lightweight
Model comparison, feature engineering and exploratory distributions can remain in a controlled analytical layer.
What should be integrated
Approved distributions, effective dates and fallback rules should be integrated into APS or ERP planning logic.
The planning parameter should express uncertainty—not conceal it.
Source discussion: Conor E. Doherty on LinkedIn.