Definition
Forecast Value Added (FVA) is a measure that seeks to compare different forecasts and see how much improvement is made at each step along the way.
This is done by evaluating the forecast error of two different forecasts to see if one has improved upon the other, or not.
FVA can be positive or negative. A positive value means that changes made the forecast more accurate, whereas negative means the forecast error has been increased.
In StockIQ, we use this to compare the benchmark forecast to our statistical model to ensure we are achieving good baseline performance.
Additionally, FVA is commonly used during a SIOP process to evaluate how a forecast is being improved (or not) at each step of the SIOP process, so that each snapshot/checkpoint can be evaluated along the SIOP pipeline.
Applies To
All items in StockIQ