Ecological restoration in its literal sense, i.e. bringing a system back to a former state, will be increasingly difficult in a globally changing world. The recovery of degraded systems, which still remains an urgent issue, will more and more often end up in systems that were never seen before at this particular site. The challenges of such "forward restoration" to "designer ecosystems" are tremendous, because it lacks any reference state to value the success of restoration projects. The only criteria left are "self-sustainability" and "ecosystem functions", and the aimed assemblage of such communities implies a thorough understanding of systems that we do not even know yet. So innovative conceptual approaches have to be developed, based on the fields of disturbance theory, succession theory, and assembly rules. Disturbed ecosystems should primarily be considered as "ecosystems" instead of "disturbed", because what is left is the ground from which the future state must be developed. Succession trajectories can hardly be predicted, so careful monitoring of the self‑running recovery process is needed. Most helpful will be a link to the currently flourishing biodiversity research. Experimental approaches in this field are dealing with the same criteria and questions, but look for general patterns in the opposite direction compared to ecological restoration, i.e. when systems are developing from complex to simple. In particular, biodiversity experiments often include "artificial ecosystems", i.e. communities that were assembled according to the experimental design and that have no natural correlate, which may serve as models for unknown designer ecosystems.