This paper presents the findings of a study on the use of synthetic data for the rapid implementation of an online store user profiling system. The model in question [3] is trained on data derived from monitoring user behaviour on websites. In order to obtain the optimal results, it is essential to collect representative data, which vary for each e-shop. However, data collection is a time-consuming process, particularly in small shops where the number of purchases is low. Therefore, we investigated the possibility of using synthetic data generation methods for training the profiling model, which would allow for a shorter implementation time of this model in an online store.