Regulation mostly relies upon the results of standard toxicity tests that have been validated for a reduced set of species and endpoints. This pragmatic strategy ensures the reliability of toxicological data. Yet, current practices might not appropriately cover the broad range of toxicant mode of action (e.g., endocrine disruptors - EDCs). In particular, the standard approach may fail to provide realistic risk forecasts when the test design is not sensitive enough to highlight toxicity and when the various standard tests provide conflicting results. Through two case studies, we explain why the standard approach may not allow assessing properly the risk of reprotoxic pesticides in invertebrates. The first case study illustrates current issues in risk assessment of EDCs. In this context, we show how non-standard partial life-cycle (PLC) tests allow identifying sensitive development phases/endpoints to EDCs (e.g., anti-androgens) along the life-cycle of Lymnaea stagnalis, a candidate species for the development of forthcoming OECD test guidelines. We also show how non-standard data analysis methods, such as mixed effect models, can be used to avoid underestimation of EDCs toxicity. The second case study illustrates how a non-standard approach can contribute to provide complementary data when standard approaches led to conflicting risk forecasts. In this context, we repeatedly exposed L. stagnalis to pulses of diquat and monitored life-cycle traits. We show how non-standard full life-cycle (FLC) tests allow assessing effective diquat concentrations more realistically and why non-standard models, such as TK-TD models, are relevant to characterise the time-course of biological responses to complex exposure patterns. For each case study, we show how to derive relevant effects criteria and include them in the regulatory risk assessment process. We then investigate whether including this additional data might modify current risk forecasts for the studied compounds. This study illustrates how non-standard test procedures of different nature may contribute to ecological risk assessment of chemicals, when standard procedure are lacking or provide conflicting results. Most importantly, the use of some simple but relevant PLC and FLC tests methods with L. stagnalis should be promoted when standard approaches are suspected to provide unrealistic risk forecasts.