<p>Customer attrition risk assessment is important in formulating proactive marketing and service strategies throughout a customer’s journey. Previous work has shown that the customer characteristics that indicate attrition risk change over the customer tenure (i.e., the time a customer is with the company), and the same tenure length may represent varying levels of attrition risk for different customers. This work proposes a predictive analytics approach for tenure-aware customer attrition risk modeling. We extract effective customer features and develop an estimator to predict "high-risk" and "low-risk" customer tenure distributions for new (unseen) customers. Given the proposed estimator, we derive new features to train a customer attrition risk model. The proposed approach effectively predicts if a customer is at a high or low attrition risk and the timing of these events. Extensive experiments on data obtained from an insurance company validate these findings using various machine learning algorithms and performance metrics. The proposed approach addresses the gaps in traditional customer attrition modeling and contributes to effective data-driven customer retention strategies.</p>