The importance of a well-balanced cross-channel marketing strategy has increased over the past few years. The synergies caused by the interdependencies of different online channels, such as e-mail advertising, search engine and banner advertising, have also drawn the attention of many researchers. However, relatively little is known about the impact of offline marketing, such as TV advertising, on online user behavior. In this article, a model commonly used in clickstream analysis is extended by adding several TV advertising variables. Based on this model, a hierarchical Bayesian logistic model is developed to estimate the cross-channel effects of both offline and online channel contacts. By applying this model to a case study, it is shown which online channels are most supported by television ads. The findings of this paper have managerial implications for practitioners in the field, in particular because of the increasing use of a so-called "second screen" while watching TV.