Data & Calculations
Part 1. Forecast Accuracy
Forecast accuracy is measured by tracking errors between actual sale quantity against forecast quantity on monthly basis.
- For each forecast unit, for instance product family, or planning BOM,
- Accuracy = 1 – [Absolute value (Forecast_Qty – Sales_Qty)]/ Sales_Qty
- For all forecast units,
- Total Accuracy = average of all forecast units’ forecast accuracy
Part 2. Forecast Bias: There are two types of bias:
+ve bias, positive bias or over forecast, if (Forecast_Qty – Sales_Qty)>0
-ve bias, negative bias or under forecast, if (Forecast_Qty – Sales_Qty)<0
The forecast bias are calculated in terms of dollar amount and percentage against total sales:
- Over forecast amount $ = sum of Over Forecast Qty x Unit Cost
- Under forecast amount $ = sum of Under Forecast Qty x Unit Cost
- Over forecast % = (sum of Over Forecast Qty x Unit Cost)/Total Sales
- Under forecast % = sum of Under Forecast Qty x Unit Cost/Total Sales
The forecast bias enables one to drill down the data to identify possible sources of forecast errors.
Aberdeen Group did survey on 151 organizations and found that Average Forecast Accuracy at product family level was 86% for the Best-in-class vs. 68% for the Industry Average and 25% for the Laggards.