Survival Analysis Applied to the Return of Containers in a Circular Economy System

Abstract

 Within the framework of the circular economy, the reuse of containers is essential to achieve both sustainability and profitability. Circulo S.A. markets cleaning products using returnable containers, whose return by customers is variable and uncertain. This creates challenges in stock management and forces the company to purchase new containers, impacting operating costs. This study applies survival analysis to model the return time of containers, using data from the company's internal traceability system (TRAC). Each record was transformed into usage cycles, considering assignments to customers, returns to the warehouse, or losses (censored cycles). Tools such as the Kaplan-Meier curve and the Lognormal distribution were employed. The results show that 90% of the containers are returned 116 days, according to the Lognormal distribution. However, the Kaplan-Meier curve, which includes censored data, reveals that 45% of the containers had not been returned by day 125. Differences by container type were also observed: 5-liter drums have a higher return rate, while bicarbonate jars are the least likely to be returned. Based on these findings, it is recommended to establish differentiated penalty timeframes, segment customers according to their return history, and include information on the type of product transported in each container. This study demonstrates the value of survival analysis as a key statistical tool to optimize operational management and strengthen sustainable practices in circular economy businesses.