Dynamic segmentation fusion

Advertisers increasingly need to expand their brands’ franchises beyond traditional marketing channels to include multi-/crossmedia
platforms. At the same time, efficient utilization of such multi-platform media requires increased consumers insights.
Obtaining detailed information from a single research survey respondent, however, is impractical due to research quality and
economic factors.
As an alternative, media researchers and others are vigorously pursuing data integration techniques whereby independent
databases are integrated through rigorous, formal statistical procedures. The results are integrated data sources that can be used
to facilitate cross-media consumption and related consumer insights.
One prominent data integration technique is fusion. While the details of particular fusions are numerous, most are of one of two
forms, static/monolithic or analysis-time. With static/monolithic fusion respondents from different datasets are matched once
using a predetermined hierarchy of common variables. In contrast, analysis-time fusion reassesses and optimizes the unique
interrelations between the common variables each time a particular analysis is developed.

Download Paper (PDF)