The recent growth of business-to-consumer (B2C) Internet auctions challenges researchers to develop empirically-sound explanations of critical factors that allow merchants to earn price premiums in these auctions. The absence of a comprehensive model of Internet auctions leads us to conduct an exploratory study to elucidate and rank critical factors that lead to price premiums in Internet auctions. We employ Classification and Regression Trees (CART), a decision-tree induction technique, to analyze data collected in a field study of eBay auctions. Our analysis yields decision trees that visually depict noteworthy factors that may lead to price premiums and that indicate the relative importance of these factors. We find shipping cost, reputation, initial bid price, and auction ending time as the factors most predictive of price premiums in B2C Internet auctions.