Given that roughly 40% of academic library book holdings go unread over a 5-10 year period after publication, there is a need to reengineer current acquisitions strategies. The large number of books that remain unused represents a financial drain on acquisitions budgets, which make libraries more vulnerable in an age of greater usage data transparency. No one has yet succeeded in devising a strategy that predicts which books will circulate and which will not. We propose a vision whereby libraries will, in the future, provide MARC (or similar) records in their online catalogs of books that they do not yet own. Patrons will encounter these records—that will look like any other records in the online catalog—but instead of finding a call number for these titles, they will find a link to order the book. The title will then be printed on demand or supplied to the library patron within a short time, and added to the permanent collection thereafter. This paper will explore what such a patron-driven collection development model might look like in practice, and touch on various approaches to how such a model might work. Included is discussion about implications of such a model on library holdings, book vendors and distributors, and publishers. [ABSTRACT FROM AUTHOR]