It is commonly recognized that online transactions are “one-shot” transactions. However, a contemporary dataset from a dominant online marketplace in China reveals that averagely 24.3% transactions are repurchase transactions. Given that consumers already have purchase experience with a specific seller, their repurchase behavior may be influenced by both the seller’s reputation and their perceived seller performance. A consequent research question is: Whether and how do these two streams of information jointly affect consumers’ repurchase behavior? We adopt a belief update model, and also collect actual transaction data to examine this research question. Our findings include: (1) both seller reputation and consumers’ perceived seller performance have positive effects on consumers’ repurchase probability; (2) the effect of seller reputation is positively moderated by performance ambiguity; (3) consumers’ perceived seller performance has stronger effects on their repurchase probability when the seller has low reputation (vs. high reputation).