Optimizing Product pricing and sales forecasting through advanced data science: a case study at Schneider Electric Iberia

Abstract

In the increasingly competitive global business milieu, product pricing optimization and accurate sales forecasting are paramount. This MSc thesis probes these critical areas in relation to Schneider Electric Iberia, a front-runner in the digital conversion of energy management and automation. Our emphasis is on the adoption of cutting-edge data science methodologies, including econometrics, Machine Learning, causality analysis, and Deep Learning, with a goal to both predict sales and optimize price-points considering the demand elasticity of diverse products across various markets. The thesis initiates with an in-depth analysis of Schneider Electric’s extant pricing and sales forecasting systems, proceeding to the selection of suitable data science techniques for enhancement. Utilizing these methods, we devise and deploy a pricing optimization model aimed at augmenting revenue or sales volume. This model’s potential is then harnessed for sales forecasting, measuring its influence on the company’s operations in aspects like efficiency, profitability, and strategic decision-making amplifications. Our methodology pivots on comprehensive data collection, meticulous preprocessing, and insightful exploratory data analysis. We leverage the benefits of Graph Causal Models for price optimization and the innovative Temporal Fusion Transformer (TFT) for sales forecasting, conjuring a formidable tool for strategic planning. The optimized prices and predictive sales model converge on an interactive Tableau dashboard, endowing Schneider Electric Iberia with a user-friendly, accessible platform for data-driven decision making. This study aims to empower Schneider Electric Iberia, while also making a noteworthy contribution to the wider field of industrial technology and the deployment of AI in product pricing and sales forecasting.