This article examines the implementation of a set of compulsory IT-based, self-scored, and adaptive nationwide tests in a low-stakes accountability system. We exploit exogenous variation resulting from students voluntarily retaking the nationwide test after the IT system was down for ten days. We find beneficial effects of testing across the student population. Disadvantaged schools were more likely to (re)take the test despite the crash, and their students gained more from being tested. Our results indicate that the core component in accountability systems, i.e. student testing, is beneficial even without the high stakes that are part of many accountability systems.