Even though B2B EC is becoming popular, there have been not so much studies about performance evaluation methodology for B2B systems. In this paper, after analyzing buyer-carts systematically focusing on the buyer's interactional efforts on the typical buying processes of each buyer-cart, we propose a quantitative performance evaluation model. For this, we categorize buyer-carts in B2B EC as s-cart, i-cart, and b-cart depending upon its residing sites: seller, intermediary, and buyer sites. And after proposing the desired features of buyer-carts in B2B EC as identification, collection, trashing, ordering, payment, tracking, recording, purchasing decision support, and transmission of records to e-procurement systems, we derive a performance evaluation model by calculating detail sub-processes from the desired features' viewpoints. By setting variables from a survey on the actual condition of using buyer-carts in companies in Korea, we try to evaluate the performance of buyer-carts in B2B EC. In this paper, we suggest a new methodology of performance evaluation for B2B systems, and show that the b-cart platform is more efficient than other buyer-carts especially in B2B EC.