Optimización de la red de Fulfillment en un Marketplace online

Fulfillment networks pose a significant challenge for online marketplaces due to their high investment and elevated fixed costs. As a result, they require a large volume of transactions to achieve profitability. At the same time, they must compete not only with other online marketplaces but also with traditional brick-and-mortar retail. In this competition, providing fast delivery times is crucial to outcompete other marketplaces and to bridge the gap with the immediacy of traditional retail. In this endeavor, the location of warehouses, the speed of operations, and the efficiency of the logistics network play important roles. However, it is also necessary to consider other strategies to win the race for speed. In this context, intelligent stock distribution and cost minimization become fundamental objectives to ensure both the success and sustainability of the e-commerce business. This thesis explores and discusses the optimization of the fulfillment network in an online marketplace. To achieve this, real data on sales, products, transportation, capacities, and costs are available. The analyzed datasets include millions of products and sales to ensure the robustness of the results. To meet the objectives, several data science models are developed in order to segment products, predict costs, and optimize the valuable resource: the sellers' stock. However, the mentioned components are not used in isolation; instead, they are integrated into the proposal of a complete optimization system for the strategic redistribution of inventory throughout the logistics network. The results of this process demonstrate a significant reduction in operational costs for the selected product set, an improvement in financial efficiency, and an improvement in user experience. Specifically, there is a 5.9% reduction in total logistics costs and a 7.1% reduction for the most relevant product set in the marketplace. Additionally, due to the application of this optimization system, product sales can be expected to increase by up to 50% through higher conversion rates.

Grantor dc:publisher

Universidad Torcuato Di Tella

Year dc:date.issued

2025

Author dc:creator

• Melnitzky, Eitán Nicolás

Advisor dc:contributor.advisor

• Marenco, Javier

• Red de distribución
• Comercio electrónico
• Inventario
• Programación lineal
• Análisis de costos
• Distribution network
• Electronic commerce
• Inventory
• Linear programming
• Cost analysis

Statement dc:rights

• info:eu-repo/semantics/openAccess

Language dc:language

spa, eng

Repository record dc:identifier.uri

https://repositorio.utdt.edu/handle/20.500.13098/13744

OAI identifier oai:identifier

oai:repositorio.utdt.edu:20.500.13098/13744