IRIS.TV takes a 3 pillar approach to making personalized recommendations:
1) Contextual Analysis: analyzes similarity of content utilizing meta-data (titles, description, tags, category, keywords) associated with the video.
2) Consumption Analysis: tracks and assesses anonymous individual user consumption/behavior building a user consumption profile which is paired to the publishers content library. Consumption data includes: videos consumed, interactivity, geographic location, device type etc
3) Audience Analysis: by utilizing pillars 1 and 2, IRIS.TV creates an audience consumption tree helping the system understand consumption trends in context to the content and audience.