Background: An important reason for failure of anatomic total shoulder arthroplasty is glenoid component loosening. We investigated the effect of backside bone support on the risk of failure of a glenoid component. Methods: A finite element model was developed. Virtual surgery was performed for 2 types of glenoid components (cemented all polyethylene [PE] vs. metal backed [MB]), both with gradually decreasing backside bone support. Both bone failure and fixation failure were analyzed. The percentages of bone failure and fixation failure in terms of the critical cement volume (CCV) and micromotion-threshold percentage ratio (MTPR) for the PE and MB components, respectively, were defined and compared. Results: For the reference PE model, the percentages of bone failure and fixation failure (CCV) were 17% and 34%, respectively. With eccentric loading for the MB component, the percentages of bone failure and fixation failure (MTPR) were 6% and 3%, respectively. A global increase in failure was observed with decreasing bone support. The increase in fixation failure, starting from the reference values (MTPR vs. CCV), was relatively more pronounced for the MB component (136% vs. 128%). Discussion: Decreasing backside bone support for an anatomic glenoid component leads to an increased risk of fixation and bone failure. For PE components, decreasing backside support to 95% bone support had only a limited effect. In the case of an MB component, we noticed an increase in micromotion and bone failure already starting from 97% bone support. We conclude that an anatomic glenoid component should always be implanted while maximizing backside bone support.