This capstone develops a data-driven forecasting and simulation framework to improve market share prediction for a leading consumer packaged goods (CPG) company in the U.S. cookies category. The current approach relies on historical trends and internal assumptions, limiting its ability to capture competitive dynamics and evolving consumer behavior. To address this gap, the project integrates SKU-level sales data, demographic information, macroeconomic indicators, and consumer search trends to identify key drivers of market share. A causal modeling framework is applied to estimate the impact of pricing, promotions, and distribution while accounting for endogeneity, using Double Machine Learning (DML) and Causal Forest models to capture both average and heterogeneous effects. The estimated relationships are translated into a scenario-based simulation tool that enables decision-makers to evaluate alternative strategies and assess their projected impact on future market share, supporting more informed and forward-looking business decisions.
