First a discussion with the interviewer about my phD research, 35mins, he asked a lot of questions about the research top and the methodology; next, some questions related to basic econometrics, but requires deriving formula (e.g. to derive the variance of an OLS estimator from the OLS estimate itself); the next question is "how you evaluate the effectiveness of google mail with respect to persuading people using google more frequently than our competitors, say, yahoo, you can imagine you can get what ever data you could, - tell me what data you would like to collect and how you want to use them, and what potential insights would you expect to get out of them?"