Following retail investors' increased accessibility to the financial sector, determining their risk propensity has become important to the industry. This research identifies the correlation between the existing customer risk aversion analysis method and the actual investment result. We use customer investment data and risk propensity survey results of the 370,000 customers of an investment company in South Korea. The results reveal that the two factors are insignificantly correlated. We suggest a new method of identifying actual customer risk aversion using the Xgboost model.