The proliferation of online auctions has attracted significant research interest in understanding real-life bidding behavior. However, most of the empirical work has focused on business-to-consumer (B2C) auctions. A natural question is whether the findings obtained from B2C auctions are applicable to business-to-business (B2B) auctions, which often involve much higher stakes. In this paper, we examine how professional bidders choose their bidding strategies in multichannel, sequential B2B auctions. Using an extensive data set from the world's largest B2B market for cut flowers, we find a stable taxonomy of bidding behavior and identify five distinctive bidding strategies. In addition, we demonstrate that bidders' choice of strategies is associated with their demand, budget constraint, and transaction cost. These findings challenge the conventional view that bidders' bidding strategies will converge as they gain experience. We also analyze the economic impacts of different strategies. Our results provide useful implications for practical design of B2B auctions.