One of the major challenges faced in Wikidata, like in any crowdsourcing project, is to attain a high level of editor engagement. There is a clear need for attracting new editors and growing the community, but at the same time, retaining existing editors is also crucial for the success of Wikidata. While Wikidata has currently thousands of active editors, only a few of them are power users who perform large amounts of edits, and/or stay long in the project. We conducted a longitudinal study to compare the editing behaviour that editors with different levels of engagement exhibit over time. Our analysis covers Wikidata edits over almost 4 years and monitors editing behaviour in a session-by-session- and monthly-basis. We observed multiple dimensions, such as the way the participation, the volume and the diversity of edits done by Wikidata editors change. Using the findings in our exploratory analysis, we define and implement prediction models that help us predict whether (i) an editor will contribute with a high volume of edits or not, and whether (ii) an editor will stay long in Wikidata or not. The goal of the talk is twofold: first, I would like to present and discuss the results of our data analysis (30 minutes). This work was a joint work with Alessandro Checco (University of Sheffield), and Gianluca Demartini (University of Queensland), in collaboration with Djellel E. Difallah (New York University), Michael Feldman (University of Zurich) and Lydia Pintscher (Wikimedia Deutschland). Second, I would like to start a discussion on how we, as a community, could intervene in order to try to convince standard users to become power users (15 minutes). Examples of topics that would be interesting to discuss in this direction are: Understanding power users What are the good and bad habits of power editors? And what are tips and tricks that they recommend? How and when do power users decide what to edit? Defining actions towards an increase in the number of power users What should be strengthen in terms of dissemination and training? What kind of assistive tools could we implement for standard users?