This study analyzed usage data from 11 university libraries to compare ebook collections selected by librarians versus patrons. It found that patron-selected ebooks were used 2-5 times more often and had audiences that were 2-3 times larger than librarian-selected collections. Additionally, patron-selected collections were similarly balanced across subjects and Library of Congress classifications as librarian-selected collections. The results suggest that fears about patron-driven purchasing leading to unused or unbalanced collections appear to be unfounded, at least for the Ebook Library model analyzed in this study.