A system selects meals according to a user's tastes and presents media files instructing execution of the meal. The media file is presented as a set of segments, each segment having a thumbnail image and metadata associated therewith. User interactions with the segments are monitored and evaluated to determine whether the segments are viewed, viewed for inspiration, and/or viewed while executing the meal. Actions monitored and evaluated may include movement of a user device during a viewing, pausing of playback, appropriateness of pausing locations to meal preparation steps, and orientation of the device. The metadata of segments and the viewing status thereof is used to determine a user's tastes and skill level to facilitate selection of subsequent meals for the user.