Quick Summary: Event Correction removes the impact of abnormal sales or supplier activity from historical data to stabilize forecasts and replenishment plans. Replace disrupted sales months with normal history and exclude unreliable supply months when calculating lead times.
Why Event Correction Matters
When major disruptions affect trading, your recent history no longer reflects typical customer demand or supplier performance.
For example:
Sales may have dropped sharply due to temporary store closures.
Suppliers may have delivered late or inconsistently due to transport restrictions.
These effects can cause:
Artificially low computer-generated forecasts.
Safety stock that remains too high or too low long after trading normalizes.
Event Correction solves this by letting you adjust the data the forecasting engine uses.
How Event Correction Works
Event Correction provides two primary options to mitigate historical disruptions:
Sales History Replacement: You can specify which months of sales history were impacted and select an alternative "normal" month to use instead for the purposes of forecasting. This results in a new, stabilized forecast that disregards the abnormal sales data.
Purchase History Exclusion: You can specify which months of purchase history were impacted and elect to ignore orders receipted in those months when calculating lead times and safety stock.
Example
In the illustration below, sales in April, May, and June were unusually low. January’s “normal” sales month was used to replace those three months.
The dark blue line shows original sales history.
The dotted blue line shows where January’s data replaces the disrupted months.
The brown lines compare forecasts before and after correction:
Lower brown = forecast using disrupted data.
Upper brown = stabilized forecast after Event Correction.
This correction restores a realistic forecast trajectory aligned with normal demand patterns.
How To Use Event Correction
➜ Read: How To: Use Event Correction
⚠️ Watchouts
Feature Access: Only Admin users can enable and configure Event Correction.
Reforecast Required: Changes take effect only after forecasts are regenerated manually or at the next automatic forecast cycle.
Scope: The feature applies globally across all items; settings cannot be configured separately for specific product groups.
Do not correct repeatable events: Events that are genuinely repeatable (like Black Friday or an annual promotion) should not be corrected — this data is valuable for future seasonal forecasting.
💡 Tips
Choose Representative Months: When selecting a “normal” month to substitute for impacted sales history, choose the last good month of sales before the disruption. However, if your business is highly seasonal, select the corresponding month from the previous year instead.
Verify Results: After regenerating forecasts, check that forecast patterns and safety stock levels have stabilized.
Forget about these 👇 😞 😐 😃 Have your say here!