Most biodiversity data are collected by volunteers organised in natural history societies or citizen science projects, often closely aligned with (sub )national agencies and local authorities. Data may be heterogeneous in space, time and quality. Here, we present first results of trend analyses of joint work with German natural history societies and state and national conservation agencies through the sMon synthesis project within iDiv. We combine and harmonize exemplary datasets of different taxa and habitats to evaluate the potentials and limits for analysing changes in the state of biodiversity in Germany. We show trend analyses of occupancy frequency data for 60 dragonfly, 42 grasshopper species and amphibia across 3 federal states 1980-2015, using Bayesian hierarchical trend analyses that build on occupancy detection models. Based on these insights and evaluation of citizen science programmes globally, we derive principles for good practice citizen science project design, data collection and archiving and explore methodologies that can deal with fragmented data of different spatio-temporal resolution and quality. This includes harnessing the potentials offered by modern technology. Combined with experiences of joint working of volunteer experts, agencies and academic scientists, this informs perspectives for future biodiversity monitoring programmes in Germany.