Maintenance aims to ensure the availability and reliability of equipment so that production processes keep operations, and products sustain with good qualities. However, maintenance decisions rely essentially on the distribution of product’s lifetime. For instance, professional users may use machine tools intensively, whereas as hobby users, the usage of same tools is much less frequent. This variability poses a challenge for companies when adopting product-service offerings, as the maintenance policies should differ to minimise the number of unexpected failures as well as the product life cycle usage costs. This study investigates the identification of customer usage behaviour groups in mixed Weibull distributed product failure datasets. The study proposes an algorithmic model using the Weibull moments to distinguish the underlying Weibull distribution in a mixed distribution. The developed knowledge defines whether customer usage data needs to be distinguished between groups.