Shoulder arthroplasty is technically demanding and relies heavily on precise surgical technique and preoperative planning. Proper glenoid component sizing plays a crucial role for successful shoulder arthroplasty. In this study, we measured the glenoid size together with penetration depth using three-dimensional computed tomography (3D-CT). From January 2010 to January 2011, 38 patients, including males and females, without evidence of a pathological glenoid, were enrolled for this study. CT images were taken and subsequently reconstructed in 3D images. The height of the glenoid was measured and the width was measured at five different levels (H1–H5). Axial images were taken at each level, with the anteroposterior (AP) glenoid diameter divided into eight areas (W1–W7). The penetration depth between the near and far cortices (thickness) at points W1–W7 was also measured. The overall mean height of the glenoid was 37.67 ± 4.09 mm. The width of the glenoid was the greatest at the distal 4/5th point and it was the least at the proximal 1/5th point. The penetration depth of the glenoid increased as the reference point progressed in the posterior direction, which was at the 5/7th point from the anterior margin. The measurement was greatest at the W4 point at the H1 level, but the W5 point was greatest at all other levels. On the basis of this study, the posterior and inferior parts of the glenoid are thinner than the anterior and superior parts. Thus, caution must be taken when inserting screws into the posteroinferior parts, where the glenoid is thinner than 15 mm, especially in females, to avoid penetration of the far cortex. Clin. Anat. 25:1043–1050, 2012. © 2012 Wiley Periodicals, Inc.