This paper is an empirical analyses of preference manipulation in recommender systems, comparing commercial platforms like amazon with non-commercial iterations