In the challenging environment of the transparent electronic marketplace in which competitors are only a click away, Web retailers are particularly vulnerable to customer attrition. Central to business growth and survival, customer retention is an important issue that every business strives to understand and harness. While some studies have attempted to determine the factors that influence customer retention, few measure it quantitatively. However, businesses have long been eager to have quantitative information concerning their customer base: How many of their customers they can consider retained at any given time? What time lapse should trigger an alert that the customer may have defected? Based on real purchasing data from a Web retailer, and using 80 percentage of assurance as an example, this paper proposes a customer retention assessment method by calculating the aggregate 80th percentile of maximum inter-purchase times and confirms the validity of this method by showing that the assessment successfully sets apart valuable customers, in terms of number of orders, average spending per order, and total spending. This research not only enables researchers to undertake longitudinal studies of customer re-patronage behavior, but also helps practitioners monitor customer retention effectively.