Guide
Forecast accuracy metrics
How do you know if a forecast is good? These are the standard measures.
The main metrics
- MAPE (mean absolute percentage error)
- The average of absolute errors divided by actual sales. Easy to read, but unstable for slow-moving products with very low or zero sales.
- WAPE (weighted absolute percentage error)
- Total absolute error divided by total actual sales. More robust than MAPE across many products and widely used in retail.
- Bias
- Whether forecasts are systematically too high (driving waste) or too low (driving stockouts). A forecast can have low error but still be biased.
- RMSE (root mean squared error)
- Penalises large errors more heavily; expressed in units of the product.
Accuracy is a means, not the goal
The real test is in store: less waste, fewer stockouts, less time spent ordering. A forecast that is slightly less accurate on average but avoids big misses on peak days may be more valuable.
Related
Ready to cut away the waste?
Tell us about your store or chain, and we will show you how AI sales forecasting can help you order the right products in the right quantities.