The Fukuda-stepping-test (FST), i.e., repetitive walking on the spot while blindfolded, has been proposed as a means to assess the integrity of the vestibular pathways. While its sensitivity to detect abnormalities in patients is limited, it may be useful in studying the physiology of the subjective-straight-ahead (SSA). Considering reported systematic shifts in SSA in humans, we hypothesize that such asymmetries arise from individual differences in the orientation/configuration of the macular organs and in central processing of vestibular input. We hypothesize that such asymmetries are stable over time in individual subjects. Alternatively, such asymmetries may arise from random noise in the sensory/motor systems involved, demonstrating low reproducibility over time. Twenty-four subjects walked on the spot over 60 s while blindfolded (n = 6 trials per subject). Using an inertial measurement unit (IMU) placed at the chest, angular deviations were recorded and compared to manually-measured final positions. Both static (direction, magnitude) and dynamic (time-to-onset of deviation, pattern of deviations) parameters were retrieved from the yaw slopes. Significant deviations were found in 15/24 participants for the manual measurements (leftwards = 8; rightwards = 7), whereas when using the IMU-sensor 13/24 participants showed significant shifts (leftwards = 9; rightwards = 4). There was a high correlation (0.98) between manually measured rotation angles (average absolute deviations = 58.0 deg ± 48.6 deg; intra-individual variability = 39 deg ± 24 deg) and sensor-based yaw slopes (1.00 deg/s ± 0.88 deg/s; 0.67 deg/s ± 0.41 deg/s). Relevant yaw deviation was detected 22.1 s ± 12.3 s (range = 5.6 s-59.2 s) after the onset of marching (no relevant yaw-deviation in 15/139 measurements), showing a mostly linear behavior over time. We observed significant inter-individual variability in task performance in the FST, reproducing findings from previous studies. With test-re-test reliability being moderate only, but at the same time observing a preference in the side of shifts in most trials and subjects, we conclude that likely both individually varying estimates of straight-ahead and random noise contribute to the pattern of angular deviations observed. Using an IMU-sensory based approach, additional dynamic parameters could be retrieved, emphasizing the value of such a quantitative approach over manual measurements. Such an approach may provide useful additional information to distinguish patients from healthy controls.