Definition
Weighted absolute percentage error, or WAPE, is the sum of absolute forecast errors divided by the sum of absolute actual values, usually expressed as a percentage. It summarizes point-forecast error across a specified evaluation sample. It is undefined when that denominator is zero.
Why It Matters
- A planning dashboard needs an explicit score before teams can compare forecasting approaches on the same orders or units.
- For iKawn as a Commerce Intelligence OS, the useful question is whether the evaluation supports a replenishment decision for the affected assortment.
- An aggregate dashboard should retain the SKU and location context needed to investigate operational exceptions.
How It Works
- Freeze the eligible SKU-location periods, horizon, unit, and forecast issue time. Separate demand estimates from actual sales before evaluation.
- Calculate absolute errors for each observation, sum them, and divide by total absolute actuals. Do not average item percentage errors and label the result WAPE.
- Keep the denominator visible: growing demand can lower WAPE despite unchanged absolute error. Compare models on identical observations and consider scaled error measures when evaluating changing series.
- Alongside the portfolio score, list the products that missed their operational stock targets. Assign an owner to review those exceptions before an agent changes purchase quantities.
Ecommerce Example
Context: Illustrative example: three SKU-period actuals are 100, 50, and 0 units; forecasts are 90, 60, and 5. Absolute errors total 25 units.
Recommended move: WAPE is 25 divided by 150, or about 16.7%. The zero-sales SKU still contributes its five-unit error to the numerator.
Why it matters: The merchant should also inspect whether the ten-unit shortage forecast affects a key product. These hypothetical numbers are an evaluation example, not an iKawn customer result.
iKawn Framework
Specify
The iKawn framework ties the forecast score to a defined assortment and planning horizon.
Calculate
Retain observations, errors, and denominator in decision evidence.
Inspect
Connect aggregate performance with SKU-level commercial exceptions.
Review
Require a documented planning rationale before changing replenishment rules.
Concise Summary
WAPE summarizes aggregate absolute error relative to observed volume. Keep its denominator visible and connect evaluation results to the actual planning decision.