In 2020, we introduced an open science policy in our space research institute SRON in the Netherlands. One of the main goals of the policy is that each refereed publication from an SRON first author should be accompanied by a publicly available reproduction package. Sharing the data and software used in a publication should make the reproduction of its scientific results much easier. Since there are no detailed guidelines (yet) about how to make such a package in astronomy, we developed a simple template for researchers to create a package that follows the FAIR principles as much as possible. In this talk, we will discuss the details of our open data policy and our experiences with implementing it. What steps appear to be essential to take to move scientists toward open science? And what aspects of open data policies make this journey difficult? We will review the lessons that we have learned since we first introduced the policy in our institute.