My project with the Wharton Interactive Media Initiative (WIMI) aims to model individual consumption behavior of 2010 World Cup content across ESPN's media platforms. We specified and tested two Hidden Markov models on simulated data to forecast mutli-platform reach. Currently we are developing a Bayesian multivariate regression with time-varying covariates. Communicating with ESPN executives and analyzing data from the 2010 NCAA March Madness Basketball championship allowed us to identify expected reach patterns for digital platforms. The March Madness dataset also helped us understand how the World Cup data will be measured. Our next steps are to improve the specifications of the Bayesian regression model to better accommodate the correlations, heterogeneity, and non-stationarity of the expected data. Once the first two weeks of the World Cup data become available, we will apply and refine our models to forecast the digital cross-platform reach during the last three weeks of the championship. A predictive model from multi-platform reach will help ESPN better serve customers and more accurately value advertisements.