Cast Demographics, Unobserved Segments, and Heterogeneous Switching Costs in a Television Viewing Choice Model

Abstract

An accurate television viewing choice model is an important tool for television executives and advertisers. The authors present a new viewing choice model, which differs from the traditional Rust-Alpert model in three ways: (1) It introduces a new show characteristic—the demographic characteristics of a show's cast; (2) it allows the preferences over the traditional show categories to be a function of both observable and unobservable individual characteristics; and (3) it allows the cost of switching among the viewing alternatives to vary across show types and individual characteristics. Cross-validated predictive testing shows that this model fits the data better than the Rust–Alpert model does. Furthermore, the authors demonstrate that network executives can improve ratings by using this model. The authors predict that by following their scheduling strategies (in particular, by broadcasting comedies after 10:00 P.M.), ABC, CBS, and NBC could have increased their average ratings by 12.9%, 7.5%, and 6.7%, respectively, on selected evenings.