This paper presents evidence from a natural-experiment which evaluates the effectiveness of a student mentoring program. The mentoring includes several compulsory, scheduled, faceto- face appointments between a mentor and a student in the first study year. All mentors are graduated and employed by the institution. For the evaluation, I use the fact that the mentoring is only offered to students in an economics and management program, whereas it is not offered to students in an industrial engineering program. However, students in both programs take the same classes and write the same exams in their first study year. I find that the mentoring program significantly decreases the failure rates in the first semester exams.