This report presents an analysis of patient setup errors resulting from inaccurately positioned wall lasers. It suggests that laser beams should agree within 0.2 degree or better with the machine axes that they are delineating. For typical simulator and treatment rooms having wall-to-isocenter distances of 3 m, this requirement is satisfied when the beam-emitting aperture is mounted within about 1.0 cm from the intersection of the respective machine axis with the wall. To achieve the required precision, we developed and clinically tested a simple, inexpensive tool, the Laser Placer (LP). The essential component of the LP is a cube with mirror surfaces that is aligned with the machine axes using built-in spirit levels and the light field and crosshairs of the collimator. Wall, ceiling, and sagittal lasers are installed and aligned according to reflections of their beams by the cube, and reference lines provided by the LP. Measurements showed that, even in new accelerator installations performed by highly experienced technicians, wall lasers are often mounted off target by more than 1.5 cm. Such inaccuracies can contribute systematic errors of 2 mm or more to the random setup errors attributable to interfraction movement in patient anatomy. To keep setup errors to a minimum, medical physicists should check beam orthogonality in addition to beam congruence at isocenter as recommended by the TG-40 quality assurance protocol from the American Association of Physicists in Medicine.