Encoding strategies vary in their duration of effectiveness, and individuals can best identify and modify strategies that yield effects of short duration on the basis of retrieval failures. Multiple study sessions with long inter-session intervals are better than massed training at providing discriminative feedback that identifies encoding strategies of short duration. We report two investigations in which long intervals between study sessions yield substantial benefits to long-term retention, at a cost of only moderately longer individual study sessions. When individuals monitor and control encoding over an extended period, targets yielding the largest number of retrieval failures contribute substantially to the spacing advantage. These findings are relevant to theory and to educators whose primary interest in memory pertains to long-term maintenance of knowledge.