Spatial navigation is supported by visual cues (e.g., scenes, schemas like arrows, and words) that must be comprehended quickly to facilitate effective transit. People comprehend spatial directions faster from schemas and words than scenes. We hypothesize that this occurs because schemas and words automatically engage space-based attention, allowing for less costly computations. Here, participants completed a spatial cueing paradigm, and we calculated cue validity effects – how much faster participants responded to validly than invalidly cued locations – for each cue format. We pre-registered Experiment 1 and found significant cue validity effects with schemas and words, but not scenes, suggesting space-based attention was allocated more efficiently with schemas and words than scenes. In Experiment 2, we investigated whether these effects were due to costly computations to extract spatial direction and found that increasing cue duration had no influence. In Experiment 3, significant cue validity effects were observed for orthogonal but not non-orthogonal spatial directions, suggesting space-based attention was allocated more efficiently when the spatial direction precisely matched the target location. These findings confirm our hypothesis that faster spatial direction comprehension is guided by efficient allocation of space-based attention. Moreover, schemas and words may be more effective supports than scenes for navigation performance.