What it measures
Systematic over or under forecasting tendency
Bias measures the outcome of a completed planning period. The financial consequences (write-offs from over-forecasting, lost sales from under-forecasting) are already locked in by the time the KPI is calculated.
Business impact
- Persistent bias creates a 'chaos tax' of unplanned operational costs: premium freight and external warehousing for under-forecasts, or disposal charges and rework for over-forecasts.
- A positive bias directly inflates the inventory position with stock that is aging past retailer shelf-life acceptance thresholds, creating phantom availability and guaranteed write-offs.
Formula
(Forecast - Actual) / Actual x 100Grain
One row per SKU per location per forecast version per period
Inside the Implementation Pack
A preview of what's actually built for this KPI — extracted from the real Implementation Pack, not a mockup.
This is the primary ratio the pack's other measures build on — trend, rank, and period-comparison variants, all on Fact_Demand_Forecast.
Fact_Demand_Forecast— 9 mandatory columns — plant key, forecast version key, ...Dim_Customer— key + nameDim_Forecast_Version— key + nameDim_Material— material_key + material_descriptionDim_Period— key + nameDim_Plant— key + nameDim_Product— key + nameDim_Date— Auto-generated calendar table — no customer data required
Full field-by-field data contract, in 4 SQL dialects (SQL Server, Snowflake, Databricks, BigQuery), ships in the pack.
| Target | 0.05 — reference value for this KPI |
| Warning | 0.08 — amber alert — triggers a review |
Both live in one Excel tab — change them, refresh the report, done. No DAX editing needed.
Everything you need to build Forecast Bias in Power BI — dimensional model, DAX measures, SQL views, and a food-industry demo dataset.