Much of the existing deterrence research showed that increasing penalty severity does often not increase compliance. The deterrence hypothesis and the effectiveness of penalties were cast into doubt altogether by one particular study, which argued that “a fine is a price” (Gneezy & Rustichini, 2000, p. 14). Descriptive norms, on the other hand, have effectively increased rule compliance in a wide variety of cases but they sometimes backfire, and it is difficult to predict its effectiveness a priori. Hence, tools for policymakers to increase compliance seem rather limited. Therefore, a natural field experiment was conducted at Erasmus University Rotterdam’s (EUR), aiming to broaden policymakers’ tools for increasing compliance. The descriptive norms treatment entailed presenting late borrowers with the borrowing compliance among the EUR library’s borrowers. The treatment was based both on literature and a preliminary survey and it was hypothesized that it would increase compliance. This means that late borrowers would hand in their overdue books faster if presented with descriptive norms than when not. In order to control for policy changes, the data also allowed for conducting a natural experiment, which consisted of the EUR library’s change in borrowing policy. It was hypothesized that eliminating fines for being one and two weeks due would decrease compliance proportions. Based on the results three main conclusions were drawn. Firstly, borrowing compliance increased significantly when two, lower fines were eliminated, which might have been affected by a higher fine at a later stage. Secondly, borrowing compliance further increased when borrowers were presented with descriptive norms. Finally, the effectiveness of the norm suggests that an online survey is a practical way to implement norms, since it can predict a norm’s (in)effectiveness a priori. These results have relevant implications for policy measures and make a compelling case for further research on the (in)effectiveness of penalties as well as new potential applications of descriptive norms.