Business Challenge
This large retailer realised that 'shelf-outs' - when a product is missing from the store shelf yet inventory says that it is in store - were costing millions in lost sales. In partnering with our supplier, they wanted to find an effective solution to predict demand and inventory needs accurately.

This added functionality layered over existing SKU management systems and directed human direction where it was predicted to be needed. The alternative is a time consuming, systematic approach to check each SKU individually. If predicted, shelf out's could be mitigated by a manager correcting system inventories and ordering ahead of time.

Approach
Our supplier leveraged machine learning to perform analytics on historical data, which served as the foundation for their demand forecasting solution.

The fully developed solution helped the company to predict inventory needs to better serve its customers: from gathering and processing data, building and training the ML model with fresh data, and predicting inventory needs accurately (all at scale).

Outcome
Implemented solution helped to reduce margin of error to ~2000 SKU’s. Additionally, internal merchandising processes was improved. Incremental revenue increased from improvements in shelf availability.