Restoring degraded ecosystems is crucial for human wellbeing and biodiversity conservation. Nowadays, ecological restoration goes far beyond recovering a degraded ecosystem according to a historic reference. Instead, more specific restoration goals are set, following global environmental strategies that are individually highly important, but often conflicting. Furthermore, various pitfalls exist regarding the evaluation of ecosystem degradation and, directly related to this, the question what is the most desirable ecological state of an ecosystem. Ignoring such issues can lead to a failure of restoration projects and do more ecological (and social) harm than good. A crucial aspect in tackling conflicting goals and circumnavigating restoration pitfalls is the considerate choice of the indicators to assess ecosystem degradation and restoration capacity. In this Perspective, grasslands and rangelands are used exemplarily for ecosystems with globally high restoration demand. I discuss potential restoration pitfalls related to enhancing carbon sequestration, soil fertility, and ecosystem service multifunctionality. For all three goals, strong trade-offs and unwanted side effects exist. For example, while increasing carbon storage and restoring soil fertility are widely acknowledged environmental goals, both can compromise other restoration targets such as grassland biodiversity and further ecosystem services, depending on the specific context. Thus, there are no universally applicable indicators for ecosystem degradation and restoration. Instead, indicator systems have to account not only for strong trade-offs among restoration goals but also for a number of environmental and socio-ecological misconceptions, such as presented for the case of grassland ecosystems. I argue that one-sided goal setting and an imprudent choice of indicators can misguide the science-policy dialogue and related restoration efforts. To avoid this, restoration programs must integrate holistic assessments of their objectives across spatial scales and with all stakeholders concerned. The associated ecological indicator system for restoration success and program performance must therefore also be based on multidisciplinary and participatory approaches. Restoration and degradation indicators have to further ensure the target ecosystem is correctly and comprehensively identified, and the manifold conflicting land management objectives associated with heterogeneous human societies are taken into account. Researchers can assist this process by bydefault considering the socio-ecological context of a restoration target and by identifying trade-offs arising from potential solutions, before these are suggested to the public. Only when all these aspects are considered, restoration projects at the local to global scale will result in long-term sustainable outcomes.