Caching holds the potential of substantially reducing the peak rate. One of its key performance metrics is the delivery rate determined by the users' demand probability, which can be changed by recommendation systems. In other words, the appropriate content recommendation may reduce the delivery rate. Unfortunately, excessive recommendations may harm the quality of experience (QoE) since a user may not be interested in the recommended files. In this letter, we focus on minimizing the delivery rate while assuring a certain QoE. To this end, we formulate a cross-layer problem, the suboptimal solution of which is attained by a heuristic algorithm.