Features described herein relate to providing content recommendations for content-consuming users by adjusting a contribution factor associated with previously consumed content for determining a recommendation score for candidate content recommended to an individual. As described in further detail below, the value of the contribution factor is based, at least in part, on a time difference between the consumption time of the previously consumed content and a reference time associated with the candidate content. In some example implementations, the value of the contribution factor may be relatively higher where the time difference is relatively closer to a multiple of 24 hours and relatively lower where the time difference is relatively further from a multiple of 24 hours. Stated differently, the contribution factor may weight the contribution that the previously consumed content makes to a recommendation score for candidate content based on the proximity of the time difference to a multiple of 24 hours. In other example implementations, the contribution factor may be adjusted based on alternative periodicities or observed viewing patterns. Various ways of using this time difference are described further below, following the description of an example network and computing device on which features herein may be used.