Demonstrates how and why market-level scanner data and store-level scanner data reveal different promotional effects; argues that it is risky for managers and analysts to use market-level data, as currently provided, to estimate the effect of promotional activities.