Retail markdown optimization

This report relates some of the work I did during my internship, as a Data Scientist in a major consulting firm. During this internship, I had the opportunity to work on a markdown optimization project for a big cloth retailer, i.e. building an algorithm able to fix optimal prices for the sale period. The goal of the project was to solve a double problem, first maximizing the revenue of the company and second ensuring we liquidate the inventories. This master thesis will be focused on the model we used to solve this case. For confidentiality reason, I cannot show the code of the model itself but we will detail the model behind it and each of its steps. After a quick introduction of the markdown business, the data and the model, we will look at the demand forecast module in details, in its first and original version. Then we will have a closer look on how we improved this simple model. Finally, a last chapter will deal with the optimization module, transforming our demand forecast into a sale forecast and more importantly the prices we will attach to each item.

Questa tesi tratta di un algoritmo di Markdown Optimization per una firma di Retail, fato durante uno stage de sei mesi. L’obiettivo era di costruire un algoritmo per fissare prezzi per la prossima sessione di ribasso. Dopo di una presentazione del Markdown Business, dei data e del modello, vedremmo il modulo di demand forecast originale del nostro algoritmo. Puoi, come si può migliorare questo modulo. Finalmente, andiamo a analizzare il modulo di ottimizzazione, trasformando le previsioni di domanda en previsioni di vendite, ed fissando i prezzi per ogni articolo.