This paper introduces a computational model of online collaborative learning system and investigates how the heterogeneity of opportunity cost among users initiates a vicious cycle of accumulation of unnecessary duplicate postings and eventually the system shuts down. Further examination explores how deterring newcomers improves collaborative learning system efficiency by reducing unnecessary duplicate postings resulting from opportunistic behavior and explains why extrinsic rewards may fail to improve use of a collaborative learning system due to an unfavorable alignment between opportunity cost and cognitive structure of knowledge. This paper suggests a new approach for explaining the relationship between opportunity cost and collaborative learning system success as well as for identifying and aligning the opportunity cost of using the system, which is suggested as a critical factor to system success.