Calculating Audience Composition
The calculation of audience composition is the most ground-breaking aspect of Quantcast’s audience service. Bypassing the sampling error limitations of single-source panels, Quantcast’s visit graph inference model enables the characterization of digital media audiences of nearly any size.
Panel approaches are only able to provide estimates for properties that are of a certain minimum size due to the inherent limitations of sample based statistics, the result being that most services only provide data on 20-30k properties. Quantcast’s composition inference approach is built on the foundation of direct measurement and provides rich audience profiles for millions of digital media assets, with as few as 500 unique viewers.