Relating Products to TV Program Clusters
A key issue in media selection is selecting combinations of programs most likely to be viewed by the targeted product user or owner group. Media models might provide weights by criteria of cost, reach and frequency for targeted groups relative to program ratings in selecting such a combination of programs. But the question remains whether more direct selection is possible by linking combinations of programs to product usage or ownership.
Two joint-space models are used to examine this question. One, utilizing multidimensional scaling, uses viewer overlap to form program clusters and then positions products based on the association of product usage or ownership to these clusters. Such a joint-space approach forms program clusters a priori and then determines product-to-program linkages.