In this study we conducted behavioral analyses to gain insights into patterns of user interaction in a video discussion platform, "Vialogues." Vialogues provides an asynchronous online discussion environment around video. Using a hierarchical clustering analysis on users' clickstream data, we identified four different behavior patterns: (1) video watchers with no discussion activity, (2) opinion seekers and active repliers with little to no video watching activity, (3) users who watched and discussed videos, and (4) users focused on viewing and/or creating metadata. Despite being the largest group, Cluster (3) had the least classifiable characteristics. Consequently we conducted additional analyses to examine finer-grained user segments. For each segment we created a transition network using weighted directed networks in order to understand the transition pattern between two consecutive click activities. [For the full proceedings, see ED593090.]