Appearance
Forecast accuracy
Also known as
- MAPE
- mean absolute percentage error
- forecast error
- forecast bias
- accuracy %
- backtest
- forecast vs actual
- KPI
- demand planning
- accurracy
Forecast accuracy compares the final forecast with actual demand over the last months, per item × warehouse and per method. Demand planners use it to see where the forecast can be trusted and which method works best for your items.
Measure forecast accuracy
- Open Supply chainForecast accuracy.
- Choose the window: Last 3 months, Last 6 months or Last 12 months.
- Optionally choose a warehouse (All warehouses when empty).
- Read the KPIs: Accuracy, MAPE, Series measured and the Best method.
- Compare methods in Mean backtest MAPE by method (percent, lower is better).
- In By item and warehouse, sort by MAPE or Bias (forecast − actual) and select a row to open that forecast.
- Select Export CSV to download the table.
How it is calculated
- MAPE = mean of |actual − forecast| / actual × 100 over the months with actual demand. Months with zero demand are skipped.
- Accuracy = 100 − MAPE (never below 0).
- Bias = forecast − actual over the window. A positive bias means the forecast is too high (stock builds up); a negative bias means it is too low (stock-outs).
- Past months use the forecast made at the time; before the first run the backtest fills them, so accuracy is available from the first run.
The same numbers are in the Forecast accuracy (MAPE) report and the forecast accuracy KPI on the overview.
Related
- Demand forecasts
- Supply chain reports
- Forecasting settings: Accuracy window (months)
- Dashboards and KPIs
- Plan to produce
