In a recent article in this Journal, S. Vargo, the service-dominant logic theorist, makes the unarguable claim that “What is needed is a true science of service (Vargo and Akaka, Service Science, 2009, p. 39, emphasis in original). To constitute a true science of service, services research must use valid measures. The sine qua non requirement for a valid measure is content validity, which comprises item-content validity and answer-scale validity (see Rossiter, International Journal of Research in Marketing, 2002, and C-OAR-SE Construct Measurement for the Social Sciences, Springer, forthcoming). The present article begins by criticizing the content validity of E-S-QUAL (Parasuraman, Zeithaml, and Malhotra 2005), the principal academic measure of e-retailer service quality, which is probably the most important construct in contemporary services research. Rossiter’s (2002) C-OAR-SE construct measurement theory is then applied to develop a new and more valid measure, called ER-SERVCOMPSQUAL. This new measure seeks potential customers’ and current customers’ judgments about the quality of the components of e-retail service (thus ensuring high item-content validity), which are rated in terms of behavioral answer categories (which have high answer-scale validity). The new measure clearly shows e-retail managers where to make improvements in their service quality.