Product portfolio configuration in supermarkets: A case study in Turkey

Official URL: https://risc01.sabanciuniv.edu/record=b2740429_(Table of contents)

Nowadays, the selection of products for a specific portfolio has become a crucial topic in stores with the enormous proliferation of supermarkets. Today, many different stores are present in the same neighborhoods and the environment has become highly competitive in terms of customer traction, obtaining loyal customers, and maintaining customer satisfaction. Supermarkets offer a broad array of products, and customer expectations are endless in terms of demanding different products. Stores have limited space in which to offer products on their shelves, but, limitless product options can be offered to customers. On this point, it is important to optimize the product portfolio to make the maximum money by using the minimum space. However, market managers are usually/often confused about determining their product portfolio because demand dynamic and accuracy in selecting the best product is problematic. Also, it is not possible to select/determine a single certain product portfolio for each supermarket in different locations because demand varies in each different location due to their differing customers, and thus, customer profile. This study approaches the issue for some specific stores and provides a solution method on determining product portfolio by using clustering method with respect to sales performance of products in portfolio, and combining these results with association rule mining. The proposed method is constructed on the data from real stores in Turkey. The findings are presented as evaluation of products in the portfolio and monthly portfolio recommendations are proposed in monthly basis. The created model can be preserved and the portfolio can be updated every year to keep pace with the changing demand of customers.