The challenge
After the supply shocks that followed the pandemic, the company could not forecast demand or allocate supply well across more than 6,000 customer and product pairs.
What was done
- Segmented the portfolio at customer and product level.
- Measured forecast accuracy, completeness and bias against best in class benchmarks.
- Built forecasting models for the segments that performed worst.
- Created a revenue optimization model to guide supply allocation.
The result
Forecasting unlocked $150M of revenue, raised forecast accuracy by 10 points and reduced inventory by $25M.
Exhibits from the work
Exhibit 1Anonymized client work
Forecast measures: completeness, accuracy and bias
Exhibit 2Anonymized client work
Forecast accuracy against best in class benchmarks by time horizon