New sensing technologies like Microsoft’s Kinect provide a low-cost way to add interactivity to large display surfaces, such as TVs. In this paper, we interview 25 participants to learn about scenarios in which they would like to use a web browser on their living room TV. We then conduct an interaction-elicitation study in which users suggested speech and gesture interactions for fifteen common web browser functions. We present the most popular suggested interactions, and supplement these findings with observational analyses of common gesture and speech conventions adopted by our participants. We also reflect on the design of multimodal, multi-user interaction-elicitation studies, and introduce new metrics for interpreting user-elicitation study findings.