Scanner-panel data have long been important for understanding advertising’s effects. Although the nature of advertising often has been investigated with scanner-panel data, the nature and value of scanner-panel data itself have rarely been considered. Although scanner-panel data are revered - even worshiped - by many, they have measurement issues like any other complex data set. While early estimates of advertising effectiveness from scanner-panel data may have appeared too low, some estimates are not vastly different from other data bases as can be ascertained from recent meta-analyses. Learning about the situations when, where, and how advertising does have large effects, however, is critical, and the future development of scanner-panel data does have a way to go to help answer these key questions. To make scanner-panel data more powerful, we recommend that choice data sets be augmented to correct for their inherent weaknesses.