Case Study: Leading Alcoholic Beverages Company achieves 5x lower forecasting costs with Sigmoid

ML-based demand forecasting solution to improve forecast accuracy while reducing costs by 5x

A leading alcoholic beverages company partnered with Sigmoid to improve market share estimation and demand forecasting across 180+ countries. The customer needed to model around 200,000 time series using limited business data and multiple external sources, but its existing monolithic demand radar code was slow, hard to scale, and could not model each country separately.

Sigmoid built an ML-based demand forecasting solution using Azure Data Lake, Databricks, Kedro, Prefect, AKS, and monitoring tools like Prometheus and Grafana, with hierarchical and scenario-based forecasting plus deep learning models. Sigmoid’s modular, country-wise pipeline improved forecast visibility and delivered a 20x faster data runtime, 5x lower forecasting cost, and 3x lower compliance and overhead cost.