This article compares three customer satisfaction models: expectation disconfirmation, ideal point, and Kano. It depicts the details of the three types of satisfaction stochastic models. A beauty shop's e-commerce database is used as empirical data for parameter estimates and model comparisons. The model calibration uses both root-mean-square deviation (RMSD) and the Chi-square test. The results demonstrate that the expectation disconfirmation model has the maximum fitness in the RMSD index, but the lowest goodness of fit in the Chi-square test. In contrast, the ideal point model produces opposite findings on the RMSD index and the Chi-square test. The expectation model, which has a larger number of parameters than the other two models, can be used for elastic changes to explain varied situational elements of pleasure, but it also requires more data to be stable. However, the ideal point model has a simpler structure than the other two models. There is only one parameter to estimate, which makes it easy to apply. However, it is less accurate than the other two models when measuring dynamic satisfaction.