In this study, we investigate what factors are influential to customer satisfaction of paid knowledge, especially among different customer segments, by integrating user activities on both free and paid platforms. Considering the complexity of knowledge acquisition, we first propose a novel measurement of “customer expertise” based on text mining, as a criterion for customer segmentation. Drawing upon the value-percept diversity theory, we then postulate a conceptual model proposing that customers with different expertise would react differently to the price of knowledge and historical knowledge-consuming transactions, in terms of customer satisfaction. We test the model empirically through the hierarchical OLS regression with data collected from Zhihu and Zhihu Live. Distinguishing expert and novice customers, we have findings that (1) expert customers are less sensitive to price; (2) historical price positively influences the satisfaction of novice customers, but negatively for expert customers; (3) expert customers are less influenced by historical satisfaction, which have important implications for market targeting and knowledge pricing strategy.