We investigated whether increased variability leads to better retention of a target face. Previous research suggests that increasing variability had no effect on memory for a target presented a few minutes later. In the real world, however, the time between when we first encounter a face and when we need to be able to identify it can be much longer. We hypothesized that increasing variability, then, might help to establish a more durable representation of a face. To test this, participants were randomly assigned to one of four conditions: a) no variability (still image), b) low variability video (little to no change in target’s appearance and context), c) high variability video (several changes in the target’s appearance and context), d) control (no video). Participants completed a sorting task 15 minutes and 5 days after viewing the target identity. Hit rates were similar in the no variability and high variability conditions, and higher than in the low variability condition. False alarms were rare, and similar across conditions. Recognition accuracy did not decrease over 5 days. The results suggest two factors influenced face learning in this task: 1) variability in context; and, 2) opportunity to attend to the target’s invariant facial features.