Essay One: Modeling Advertisers’ Willingness to Pay (WTP) in TV Commercial Slot Selling Auctions

This paper uncovers advertisers’ underlying willingness to pay (WTP) in HK television commercial slot selling auctions which are overlapped discriminatory multi-unit auctions with discrete semi-sealed bids. We specify advertisers’ WTP as a parametric function of their intrinsic valuations of pure slot attributes, as well as extrinsic valuations which depend on the context of the focal auction and competition from other auctions selling similar commercial slots. The two “no-regret” bidding principles (first proposed by Haile and Tamer, 2003) are extended to obtain informative boundary conditions for asymptotical identification. Results show that the WTP largely depends on TV rating....[ Read more ]

Essay One: Modeling Advertisers’ Willingness to Pay (WTP) in TV Commercial Slot Selling Auctions

This paper uncovers advertisers’ underlying willingness to pay (WTP) in HK television commercial slot selling auctions which are overlapped discriminatory multi-unit auctions with discrete semi-sealed bids. We specify advertisers’ WTP as a parametric function of their intrinsic valuations of pure slot attributes, as well as extrinsic valuations which depend on the context of the focal auction and competition from other auctions selling similar commercial slots. The two “no-regret” bidding principles (first proposed by Haile and Tamer, 2003) are extended to obtain informative boundary conditions for asymptotical identification. Results show that the WTP largely depends on TV rating. Advertisers’ WTP decreases with tighter budget constraints, more bidding experience, or more available similar slots. The model is useful for the TV station to predict revenue. In order to protect advertisers’ profits, two ads in the same product category are not allowed to be broadcast in the same commercial break. We find that advertisers’ WTP increases significantly when the TV station merges a group of smaller categories into one bigger category.

Essay Two: Are Advertisers Getting Their Money’s Worth? – Evidence from A Comprehensive Study of TV Viewers’ Commercial Viewing Behavior

In traditional broadcast networks, commercials are usually priced based on the expected audience size of associated TV programs. However, the actual audience size of commercials varies since viewers may zap during commercial breaks. This paper conducts a comprehensive study of TV viewers’ zapping behavior and examines the impact of the resulting audience loss on advertising cost-effectiveness.

Specifically, we model individual viewer’s binary viewing decisions in each minute of a commercial break and adopt a latent class approach to account for viewer heterogeneity. Unlike previous research in which data are limited in one way or another, this research employs a unique, large-scale and rich data set (more than 2 million observations across consumer and minute) to include all important variables collectively identified by prior literature and their interactions. Thus, we can obtain more accurate predictions. To the best of our knowledge, this is the first empirical study which includes individual viewer’s interactions with environmental factors as predictors of zapping, in particular viewers’ immediate and cumulative involvement levels in associated programs. We find that the zapping propensity is negatively correlated with both immediate and cumulative involvement levels. Moreover, the magnitudes of correlations with cumulative involvement level vary across program genres and viewers in zap-prone segments are more/less sensitive to immediate/cumulative involvement level than those who are less zap-prone.

Using actual price paid for each ad, we further examine advertising cost-effectiveness by studying firms’ best advertising purchase strategies in different scenarios. When advertisers are allowed to choose specific ad positions within a commercial break, one major healthcare product company in HK saves 8.8% amount of money to achieve the same number of ad exposures by accounting for the audience loss during commercial breaks in the advertising purchase strategy, compared with purchasing ad solely based on program audience sizes.