Leading beverage company reduced stockouts by 20% with AI-based recommendations.
01. Business Challenge
Stock Outs or Out-of-Stocks (OOS) generally refer to products unavailable for purchase at retail, as opposed to elsewhere in the supply chain. A substantial loss from out-of-stock affects the bottom line as well as the brand’s image.
The client wanted to implement a solution that would crunch millions of customer’s data and provide customers with a demand-based recommendation. In the end, this solution would reduce OOS by providing recommendation lines and quantities to be sold during a customer visit.
02. Approach
The client summarized the problem as ‘we want to reduce Out-of-Stock’. After a series of discussions and understanding of the data gathered at the customer level, a seven-layer approach was developed. In which we
- Understood the buying behavior of products for all customers.
- Develop ML models to suggest the quantity of the products required at the POS.
- Optimization of invoice amount based on the SKU prioritization and other inputs for every customer.
The system would fetch the data, generate the optimized recommendation for every customer every night with the Machine learning models. These recommendations are fed to their existing SFA system through ETL pipelines within the client environment.
03. Impact
The Predictive Solution helped the client:
- Forecast the demand better at the most granular level by Customer x Visit x SKU,
- Enabled in driving better execution efficiency with their Sales Representative.
- Ultimately, resulting in an increase in bottom-line revenue by reducing stockouts by 20%.
Key Highlights
- Accurate demand forecast at Customer x Visit x SKU
- Seamless integration of solution with existing SFA
- Increased Salesforce productivity
