We proposed a competing risks approach to analyze customer behaviors in freemium products and services. The event of interest is when a customer starts to pay for additional features or functionalities. The observation of such an event may be preempted by an event where the customer quits using the product before paying and consuming the additional features or functionalities. One such freemium service is the online gaming category. The Fine-Gray regression model was implemented for an online games player activity data to study how covariates affect the paying hazard. Some covariates are hypothesized to have different discrete effects at multiple change points. We extended the Fine-Gray competing risk model to allow for change point in multiple covariates.