In this study we report on the students’ evaluation of a self-constructed constructivist e-learning environment for statistics, the Compendium Platform (CP). The system was built to endorse deeper learning with the incorporation of statistical reproducibility and peer review practices. Based on the Delone and McLean framework we tested an explanatory success model with a sample of 607 business students, collected during three consecutive academic years. The results indicate that system quality and teacher support are the most important success factors, directly or indirectly contributing to a higher degree of relative advantage and satisfaction, both of which strongly determine continuous intention to use. The findings ascertain the usability of the CP and promote a more radical constructivist approach to the teaching of statistics, but also other subjects.