Our classical Data Archive focus was on the documentation of already conducted projects within the social science area. For that, we use structured metadata, especially the DDI-Codebook standard, and tools supporting this approach. Over the recent years, a DDI-Lifecycle approach was added to that, for better re-use options in the context of some projects which actively do data collection or have longitudinal designs. The idea was to implement the “Tornado” approach (from the Generic Longitudinal Business Process Model: many cycles of design, use and re-use) to metadata. Currently, we follow a metadata capture and processing pipeline at GESIS with the ExploreData project, based on the DDI-Lifecycle standard. The pipeline addresses technical challenges like XML transformation, database import/export, and search index creation, including the underlaying different technology stacks. Our goal is that all metadata can be published for secondary data use or for further re-use in other data collection activities. Our presentation will describe our experiences with the implementation, show the lessons learned from automated versus human testing of metadata publication in retrieval systems, and will highlight particular challenges of managing study level and variable level metadata in the workflow pipeline.