Systems and methods for providing content recommendations have been described above. As described above, content recommendations may be generated based on the proximity of a time difference to a multiple of a predetermined time period (e.g., 24 hours) where that time difference is based on a consumption time of consumed content and a reference time associated with unconsumed content available to be recommended. It will be apparent to those skilled in the art that various modifications and variations can be made without departing from the scope or spirit. Other embodiments will be apparent to those skilled in the art from consideration of the specification and practice disclosed herein. For example, instead of applying a weight to a similarity score, the similarity score for content pairings may be adjusted higher or lower based on the proximity of the time difference to a multiple of the predetermined time period in order to obtain an adjusted similarity score. The similarity score, for example, may be adjusted to be relatively higher when the time difference is relatively closer to a multiple of the predetermined time period and adjusted to be relatively lower when the time difference is relatively further from a multiple of the predetermined time period. Content recommendations may then be generated based, at least in part, on the adjusted similarity score. It is intended that the specification and examples be considered as illustrative only, with a true scope and spirit being indicated by the following claims.