Title: An analysis of TV viewing behavior using a direct utility model
Authors: Shohei Hasegawa - Hosei University (Japan) [presenting]
Satoshi Nakano - INTAGE Inc (Japan)
Sang-gil Lee - INTAGE Inc (Japan)
Abstract: The purpose is to understand TV viewer behavior. In marketing literature, TV viewing data are usually used as viewing frequencies of TV advertisement and independent variables for measuring advertising effects. The authors use TV viewing data as dependent variables and analyze TV viewer's behavior using a direct utility model. The proposed model assumes that an individual allocates viewing time to different type of TV program genres by maximizing his or her utility subject to a time budget constraint. The model has two parameters, the first is baseline parameter which means a preference and the other is satiation parameter which means a rate of diminishing marginal utility. Heterogeneity into the model is introduced using a hierarchical Bayes model and estimate individual parameters. The empirical analysis using actual data shows that (1) viewing behaviors are different between weekdays and weekends, (2) news show genre has high preference on weekdays and variety show genre on weekends, (3) educational show genre is the most satiating both on weekdays and weekends. We apply the results to clustering viewers according to their estimates.