The environmental risk of chemicals is routinely assessed by comparing predicted exposure levels to predicted no-effect levels for ecosystems. Although process-based models are commonly used in exposure assessment, the assessment of effects usually comprises purely descriptive models and rules-of-thumb. The problems with this approach start with the analysis of laboratory ecotoxicity tests, because only a limited amount of information is extracted. Standard summary statistics (NOEC, ECx, LC50) are of limited use in part because they change with exposure duration in a manner that varies with the tested species and the toxicant. As an alternative, process-based models are available. These models allow for toxicity measures that are independent of exposure time, make efficient use of the available data from routine toxicity tests, and are better suited for educated extrapolations (e.g., from individual to population, and from continuous to pulse exposure). These capabilities can be used to improve regulatory decisions and allow for a more efficient assessment of effects, which ultimately will reduce the need for animal testing. Process-based modeling also can help to achieve the goals laid out in REACH, the new strategy of the European Commission in dealing with chemicals. This discussion is illustrated with effects data for Daphnia magna, analyzed by the DEBtox model.