A value-weighted forecasting metric measures forecast performance across a portfolio by giving greater influence to products with greater business importance. It answers a practical question: are forecasting efforts improving the items that matter most to the supply chain, rather than merely the items with the largest volumes? section 11.0 section 14.4.1
A single unweighted portfolio metric can mix fundamentally different products. An error of one anvil may matter far more operationally than an error of one nail, even when nails dominate unit volume and total error. Value weighting makes such differences explicit so unlike products can be assessed together. section 11.0 section 11.1
Choose weights from the business objective rather than treating one weighting rule as universal:
- Weight errors by cash or a value measure when the goal is to maximize profit. section 11.3
- Weight products by cost when the goal is to minimize capital invested. section 11.3
- Track cost-weighted versions of MAE and bias to keep both accuracy and systematic over- or under-forecasting visible. section 11.3 section 14.4.1
Nails, hammers, and anvils
Consider forecasts of 10 anvils versus demand of 12, 150 hammers versus demand of 100, and 1,000 nails versus demand of 1,500. Nails create the largest unit error, but a portfolio decision should not assume that this makes them the most important error. A value-weighted measure applies the chosen business importance to each product before judging overall performance. section 11.1 section 11.3
Use the metric alongside Forecast Value Added when reviewing teams, channels, regions, or forecast adjustments. This focuses process improvement on critical items and helps reveal whether changes improve accuracy and bias where business value is concentrated. Select success metrics before pursuing forecasting improvements; misaligned management metrics can undermine a project. section 12.2.3 section 12.3 section 14.4.1
