For wayfinding at museums, galleries, and heritage sites, the majority of people who are blind or partially sighted favour verbal directions over tactile maps. Those who provide such directions are usually fully sighted. Can we assume that details considered important by those with vision are equally helpful to those without? This article presents a study comparing descriptions of familiar routes recalled by fully sighted, partially sighted, and blind participants, highlighting the differences in navigation strategies. All participants cited landmarks, but those with impaired vision also detailed significantly more objects, or ‘milestones’, en route. In the absence of such milestones, they relied on metrics that were often environmental or body-based. Distances were precisely specified, rather than ‘vague’. Landmarks and milestones were accessed multimodally by blind and partially sighted participants, or a combination of landmarks and milestones were sought for corroboration. In contrast, fully sighted participants reported predominantly unimodal information, that is, visual. The findings may have practical implications for improving the type of verbal orientation information provided to assist blind or partially sighted people.