Creating a Supply Chain System with a Forecasting System & Tool
Challenge
The client demand planning team uses an in-house stat model (secondary sales forecasts, i.e., distributor to retailer for the future 3 to 5 months), which then undergoes the S&OP process with the CMM & TMM team to land at a consensus forecast. This is then used to create a distribution and production plan, which drives improved availability and inventory reduction.
The client demand planning team uses an in-house stat model (secondary sales forecasts, i.e., distributor to retailer for the future 3 to 5 months), which then undergoes the S&OP process with the CMM & TMM team to land at a consensus forecast. This is then used to create a distribution and production plan, which drives improved availability and inventory reduction.
Solution
1. Creating a dynamic forecast engine tool that can create forecasts at multiple granularities (brand or parent SKU, depot or national level, and combinations of these) and ensemble/stack multiple models to forecast the best possible point estimate.
2. Creating a sensitivity/simulation module in the tool to dynamically update the demand
basis new quantified drivers and understand the impact of changes in these drivers.
3. Creating a range-based forecast tool that can forecast ranges for a specified confidence interval. Also, the range forecast should tie back to the point forecast.
1. Creating a dynamic forecast engine tool that can create forecasts at multiple granularities (brand or parent SKU, depot or national level, and combinations of these) and ensemble/stack multiple models to forecast the best possible point estimate.
2. Creating a sensitivity/simulation module in the tool to dynamically update the demand
basis new quantified drivers and understand the impact of changes in these drivers.
3. Creating a range-based forecast tool that can forecast ranges for a specified confidence interval. Also, the range forecast should tie back to the point forecast.
Results
Additionally, as the sales team always has an aspirational number vs. the forecasted demand leading to high inventory in the system, we need a solution that can estimate a range (lower and upper limit at specific confidence level) so that during the S&OP process we can push the sales team, leading to reduced inventory, and experience better accuracy with the ranges.
Additionally, as the sales team always has an aspirational number vs. the forecasted demand leading to high inventory in the system, we need a solution that can estimate a range (lower and upper limit at specific confidence level) so that during the S&OP process we can push the sales team, leading to reduced inventory, and experience better accuracy with the ranges.