In the field of retail product sales, a retail store would like to identify the most popular products in the market to make sure the product assortment of the store carries those popular products. By focusing on the most popular varieties of products and decreasing the exposure of the least popular varieties of products, a retail store can optimize its assortment selection and increase its profits and sales. The present invention supports the optimization of a retail product selection in a computerized system for categorizing products in inventory, determining market demand for products in the market, and optimizing the variety of products in inventory based upon retail sales data for a geographic region and a chain of stores. With the present method and system, a significant amount of retail sales data is manipulated into an easy to understand format for product category management and variety optimization.