The effectiveness of advertisements with respect to a group of panelists is measured. Based on the results of such analysis, advertisements and/or variants thereof are selected for presentation to consumers. Effectiveness of advertisements is measured, in one embodiment, by detecting exposure to advertisements, and then monitoring panelist behavior following exposure to an advertisement. In one embodiment, the group of panelists is a representative sample of a larger population, so that observations of panelist behavior can be used as a basis for making decisions regarding presentation of advertisements to a larger audience having characteristics similar to those of the panelists.
Once the relative effectiveness for various audiences has been determined, advertisements can be selected for presentation to individual consumers or to groups of consumers, so as to maximize effectiveness. In one embodiment, such analysis and selection is performed substantially in real-time.