Despite previous literature strongly criticising the use of student evaluations of teaching (SETs), it is unlikely that the frequency of SETs will be decreased for university instructors. However, little attention has been given to how we can make SETs work as a tool for improving teaching effectiveness. This research explores how faculty members of a large professional master's programme in Brazil interpret the results of their SETs. The findings show that qualitative questions combined with quantitative scores are effective in improving classes. It also shows that faculty critically assess students' feedback by disregarding emotional positive and negative comments and looking for suggestions in comments. Instructors with lower scores have difficulty interpreting feedback and tend to attribute their performance to nonresponse bias. This research contributes to the SET literature by showing that a combination of qualitative and quantitative questions can be effective in improving teaching quality but that some instructors may need support in interpreting SET results to change their teaching.