Improved Forecasting by applying Analytics & AI/ML

Background

Our client is a US-based multi-store retailer with multi-product manufacturing capabilities targeted towards the needs of different segments. They were experiencing unprecedented volatility in the demand of some specific SKUs from a single product category. They wanted to build a forecasting model to predict the demand of certain products for the next 1-12 months.

---

Our Solution

Developed forecast simulator models (based on an ensemble of Long Shot-term Memory & Autoregressive Time-Series Models) to predict demand for the next 1-12 months, and correct appropriately for serial correlation (the correlation over time of the impact of unobserved variables on the variable being predicted). The ensemble technique combined forecasts from multiple models to forecast reliability.

---

Results

Our solution optimized the costs of business operations and improved fulfillment for various products across different segments. This led to a jump of 12% in revenues.