To improve the overall quality of a service experience, managers often focus on enhancing particular attributes of the service. Yet identifying the attributes to improve—those with the greatest impact on customer satisfaction with the overall service—is often challenging. Generally, improving a poorly performing attribute provides a notable boost to the customer’s perceived overall experience, but further improvements have a diminishing positive effect. The task is further complicated by the fact that the point of diminishing returns, or more generally, the shape of attribute utility curves, likely differs by attribute. Thus, the priorities for attribute improvement may change as service performance changes. Marginal utility analysis (MUA), a new method for understanding the priorities for attribute improvement, is introduced. The method is demonstrated on a sample data set and compared to alternative techniques for understanding priorities including direct measures of customer priorities for improvement, importance–performance analysis (IPA), and multiple regression. MUA was found to provide a superior understanding of priorities for attribute improvement by more accurately modeling the shapes of attribute utility curves. MUA is not hindered by multicollinearity, a common phenomenon that reduces the reliability of some other methods of estimating utility curves. MUA provides diagrams of each utility curve where managers can readily identify which attributes are the highest priorities for improvement in the customer’s mind and how those priorities change as each attribute’s performance changes. The method is easy to implement with the necessary data obtained via minor modifications to a typical customer satisfaction survey.