Both cervical vestibular evoked myogenic potential (cVEMP) and ocular VEMP (oVEMP) are irreplaceable tools in the current vestibular science, but they have many drawbacks. For example, the test-retest reliability of the asymmetry ratio is inadequate for both tests. The American Academy of Neurology noted that there is insufficient evidence to determine whether cVEMP and oVEMP can accurately identify vestibular function specifically related to the saccule/utricle. One of the most important factors underlying the variability of cVEMP seems to be the control of, and compensation for, baseline muscle contraction power. Reasons for variability in oVEMP include a poor signal to noise ratio, the angle of superior gaze, effective and controlled presentation of stimuli, intracranial pressure, and electrode location. Many of these shortcomings could be improved by further development of recording methods and devices. This article examines the reasons for the insufficient reliability of VEMP and proposes avenues for improvement of VEMP recording systems.