Challenge

A Fortune 500 Food & Beverage Manufacturer with 1000s of brands across 100 plus categories were in the process of integrating a demand planning platform to help in end-to-end process automation of workflows and collaboration between marketing, merchandising, and supply chain. The goal was to improve demand forecasts for every SKU (item) in every category for each retailer in each warehouse at a weekly level for the next 26 months. These forecasts are to be used to plan daily shipping quantities at an item-retailer-warehouse-location level.

Approach

We built forecasting models at the item-retailer-warehouse-location level using 3 years of historical data; integrated the models to run on the cloud-based platform by developing appropriate python plug-ins; dynamically check forecast accuracies at multiple levels – brand/category/retailer/warehouse on a weekly basis; and enabled automation of typical Root Cause Analysis rules that generate exception reports and potential root causes of forecast deviations beyond pre-defined thresholds.

Outcome

Estimated to reduce $200MM in costs over the next 3 years by implementing integrated demand planning at scale. This will be achieved by improving planning and execution (given higher forecast accuracy), reduction in procurement costs, and increase in production efficiencies.