Our client is a global CPG major. The client wanted to create omnichannel forecasts essential to meet the ever-changing demands online (Delivery and Pick Up in-store). The key step of the journey is to understand the root cause of common inventory-related issues in store and enable decision-making by identifying gaps.

The main objective was to-

• Perform robust forecasting & safety stock measurement to optimize the minimum stock levels
• Identify phantom issues & minimize loss
• Overlay service-level issues to identify the additional root causes of OOS

• Sparseness of data for low selling UPCs X Stores makes it hard to estimate the future sales
• Adjusting for Impact of Covid in the training period data by using external data like Mobility Index
• Scale of Training & data processing was extremely high (300K+ time series) which required creative ways to optimize & execute the experiments