Neurological conditions are extremely heterogeneous. For example, after stroke, survivors range from being unable to walk at all right through to being able to walk at a normal speed. Research over the last two decades is providing clearer evidence that intervention needs to be targeted. For example for the non-ambulatory person, mechanically assisted walking with partial weight support via an overhead harness provides the opportunity to complete large amounts of walking practice. More evidence has become available that this form of intervention is effective at establishing more walking without detriment to the quality of walking. For those who can already walk, there is evidence from systematic reviews that treadmill walking and cueing of cadence are effective in improving walking speed and distance. However, there is emerging evidence that interventions are not equally effective across the range of stroke survivors, ie, one size does not fit all. This has implications for both the clinic and for future research. Interventions need to be tailored to the individual, perhaps using walking speed to do the tailoring. Systematic reviews and clinical trials should take into account the inclusion criteria of the participants investigated.