Demand Forecasting For Wines and Spirits
  • GFAIVE LinkedIn page
  • GFAIVE Facebook page
  • clutch
  • YouTube - White Circle

© 2019 by GFAIVE - ML demand forecasting for retail and e-commerce. Privacy Policy

Demand Forecasting For
Predicting Out-Of-Stock Situations
And Optimizing Stock Levels

Challenges and Industry

A retailer from the wines and spirits sector, whose business largely depended on existing customers, was struggling with frequent but seemingly random out of stocks in key locations. The retailer realized that he was losing revenue and disappointing his customers when the products they looked for were unavailable. Hence, he approached GFAIVE with the goal of increasing revenue by reducing out of stocks across all store locations.


The challenge for GFAIVE was to build a demand forecasting tool that would take into account customer preferences as well as external factors that could be affecting alcohol demand. That way, the tool would provide a high degree of accuracy and allow the retailer to continuously adjust stock levels across locations and leverage every sales opportunity.

Solution

 

The solution development process started with a proof of concept that demonstrated the abilities and accuracy of the sample demand forecasting tool. Based on data gathered from the client's internal data sources, the tool forecasted customer demand by store and by product.


Once the client was satisfied with the results and the accuracy of the demand forecasting tool during the testing, the team proceeded to build a custom predictive engine where various machine learning algorithms were tested and configured to ensure the best suitability and results.


Next, the GFAIVE team included data from external data sources into the model (market and trend data, holidays, etc.) This was done in order to boost the accuracy of the prediction engine and ensure that all known factors affecting demand were considered.

Results

 

Following the integration of GFAIVE’s demand forecasting tool, the wines and spirits retailer was able to receive weekly, monthly and quarterly demand predictions by product category for each of their stores. Within 3 months their revenue increased by 16%  and their marketing ROI rose by 23% as they were now able to leverage predictions to better plan marketing campaigns to drive demand for slower-selling products.

Technologies Used