Modelling customers’ product preferences towards an optimized store assortment

The growing variety of products provided by suppliers, the limited store space, and the ease of customers accessing competitors' offers lead to the necessity of an effective assortment planning process to remain competitive.

For this reason, our client, a retailer in the health and wellness industry, planned to revise its assortment planning process concerning glasses, highly dependent on informal knowledge and simplistic analyses.

The main goals were to take advantage of customer preferences insights hidden on its sales history, be capable of reacting to changes in demand patterns and be able to differentiate assortment according to customer profiles.

The new process encompasses three analytical modules in a user-friendly and flexible interface. The first module aims to create store clusters by analyzing the preferences of their customers, combined with demographic and economic data.

The second stage estimates the marginal value of each brand, within the store cluster, and evaluates the space and equipment restrictions in order to define how many products of each brand will be available in each store.

The final step estimates the attribute preferences, based on sales history, such as color or material, forecasts products’ sales based on their attributes and ranks products by maximizing the value criteria, which can be sales, diversity and/or profit.

In addition, the designed process will ensure that assortment decisions are effective, even if the demand patterns and product characteristics are continuously evolving. It will allow our client to meet customers’ expectations at any time and to identify opportunities that until now could go unnoticed.