A Supervised Learning Algorithms for Consumer Product Returns Case Study for FLO Offline Stores

Abstract

Abstract. One of the key dimensions of customer loyalty is consumer product returns, which are critical for success and quality of services. For this reason, manufacturers and retailers need to consider operational challenges for management of product returns. A consumer product return in “shopping” is the procedure of a customer bringing previously bought products to the retailer and getting a payment in the original payment option, an exchange for a similar or different item, or a shop credit in retail. Defective products brought to the stores for return by the customers are received by the stores and put into the review process, which takes several weeks. As a result of this examination, it is understood whether the malfunction is a user error or not. Machine learning algorithms serve to ease the burden in return operations and increase efficiency. Intelligent decision-making mechanisms, organizations will decide whether the product return should be accepted, or not by comparing attributes such as historical return data of the products, supplier information, quality of raw materials like leather or artificial leather, seasonal conditions, consumer behaviors. Boosting algorithms are commonly used for resolving binary classification issues. This study aims to present a real case study that is conducted in FLO, which is one of the biggest shoe producer and retailer in Turkey to improve decision making processes of shoe returns using customer and product data via machine learning algorithms. Keywords: Customer loyalty, Product return, Artificial intelligence, Machine learning