This study quantified the effectiveness of tactile guidance in indicating a direction to turn to and measured its benefits compared to spatial language. The device (CAYLAR), which was composed of 8 vibrators, specified the requested direction by a vibration at the corresponding location around the waist. Twelve participants were tested in normal light and in total darkness with 3 guidance conditions: spatial language, a long tactile rhythm (1 s on/4 s off vibrations) providing a single stimulation before movement, and a short rhythm (200 ms on/200 ms off vibrations) allowing information-movement coupling during body rotation. We measured response time, heading error, and asked participants to rate task easiness, intuitiveness and perceived accuracy for each guidance mode. Accuracy was higher and participants' ratings were more positive with the short tactile mode than with the 2 other modes. Compared to spatial language, tactile guidance, regardless of the vibration rhythm, also allowed faster responses and did not impair accuracy in the absence of vision. These findings quantitatively demonstrate that tactile guidance is particularly effective when it is reciprocally related to movement. We discuss implications of the benefits of perception-action coupling for the design of tactile navigation devices.