Among academic institutions data repositories are a trending addition to the ecosystem of open science, and more often than not, there is a push towards self hosting. Dataverse is one of the most popular software in such cases, as it offers out of the box much desirable features, focusing on explorability, interoperability and a promise for flexibility. From a software development point of view, the open model driven by community feedback is a great asset. Since 2018 at the Center for Socio-Political Data, we opted for migrating our DDI metadata bank from Nesstar to Dataverse. We were eager to customise Dataverse metadata schema to follow as closely as possible the Cessda Metadata Model main elements defined by our metadata experts (controlled vocabularies). We had to compose with a quite relaxed implementation of DDI 2.5 standard by Dataverse. We will share strategies, tools and workarounds we set up to approach our goal. We have also strived to make the most out of the tool, using templating mechanisms or enhancing the user interface. We will conclude with a selective list of features we want to promote in order to better align Dataverse with standards as DDI or OAI-PMH. You can find a recorded version of the whole session here: https://youtu.be/v1rUcJTvfaQ