The “privacy versus personalization” dilemma refers to the situation in which it is necessary for users to disclose their sensitive personal data in order to benefit from collaborative personalized services. Solving this dilemma is a challenge because generating collaborative filtering recommendations requires access to the set of all user profiles in order to identify similar ones, and to compute the top-rated items. The privacy-preserving personalization (P3) paradigm builds on the idea of using locality-sensitive hashing (LSH) to find groups of similar users, while keeping their profiles local. In this work, we analyze the behavior of the adapted LSH algorithm from the perspective of the quality of final recommendations and the distribution of cluster sizes. We investigate the impact of different LSH parameter configurations on the basis of the MovieLens dataset, and empirically show a small, non-prohibitive cost ofprivacy protection on the recommendations' quality. © 2014 Alcatel-Lucent.