Product Segmentation

Retail • Artificial Intelligence


Varied inventory levels, with high and low product turnovers made management complex.

Products were not prioritized by high velocity and profitability.

Out-of-stock events couldn’t be avoided due to lack of data.


Built a clustering algorithm to categorize products by recency of purchase, purchase frequency, and profit margin, for each store.

Categorized products into four classes (A, B, C and D) to focus on best performers


Increased overall margin by 2% by focusing on high-margin product categories.

Reduced slow velocity product inventory and increased best-selling inventory.

Improved inventory management by treating different product categories differently.

Avoided more out-of-stock events using product categories.