Disclosed are systems and methods for generating ratings for content items based on a user's consumption history. The content items may comprise various forms of media content, including, video, audio, Internet webpages, etc. When a user or consumption device accesses content items, a computing device may monitor the amount of the content items consumed by a user over one or more consumption sessions. In one embodiment, threshold values may be identified for one or more rating levels of a rating scale associated with the content items, and a rating for a content item may be generated based on the amount of the content item consumed by the user or the amount paid for the content item by the user. The computing device may calibrate or update generated ratings by measuring accuracy of the ratings and adjusting one or more threshold values associated with the content items.