Software engineering has moved towards being a more evidence-based discipline. For example, systematic literature studies have become a standard practice in our field. However, to be evidence-based, we, as authors of both primary and systematic literature studies, still need to ensure that we publish credible evidence that researchers and practitioners alike can build upon. The keynote will discuss evidence and illustrate that we need to take action to become more evidence-based and that we are all responsible for improving as authors, reviewers, program chairs and editors. Some of the problems are illustrated generally, and then the findings from analysing a sample of requirements engineering papers are presented before concluding the presentation.