The document discusses using external data sources and machine learning models to improve customer experience when searching for vehicles online. It analyzes factors like household income, gas prices, vehicle fuel efficiency, and state incentive laws to predict which customers are more likely to buy fuel efficient cars. Random forest and KNN models are tested, with random forest achieving 79% accuracy in predicting fuel efficiency preferences based on political demographics and other variables. The results recommend using state laws/incentives as well as web behavior data to optimize search rankings and suggestions for more relevant vehicle options.