This qualitative comparative study aims to investigate differential effects of low-stakes mandatory testing systems within a multi level structured school system at classroom level. We first analyzed the mandatory testing systems in two German states on the basis of a model on school performance feedback systems (Visscher & Coe, 2003; Verhaeghe, Vanhoof, Valcke & van Petegen, 2010). Although both states implemented low-stakes mandatory testing, the "Kompetenztests" in Thuringia provide more elaborated and reliable feedback information and a more coherent support strategy for schools compared to the "Vergleichsarbeiten" in Baden-Wurttemberg. Therefore teachers in Thuringia are supposed to use performance feedback more intensively. This study aims at verifying this theoretical assumption and at identifying favorable terms of mandatory testing systems for acceptance and usage at classroom level. We therefore conducted a qualitative interview study with secondary school teachers in Thuringia (N=27) and Baden-Wurttemberg (N=38) on how they interpret and use school performance feedback data at the classroom level. Results show that a majority of teachers in both German states rejects the idea of test data based instructional improvement. We only found some teachers in Thuringia who integrated demanding test assignments into future teaching.