This retrospective review of voice outcomes following a diagnosis of unilateral vocal fold paralysis divided patients into three groups according to the time of initiation of voice therapy following the onset of paralysis. The ‘early’ group started voice therapy within four weeks, the ‘intermediate’ group started therapy between four and eight weeks and the ‘delayed’ group started treatment over eight weeks after symptom onset. The early, intermediate and delayed groups comprised 78, 49 and 44 patients respectively with median ages of 57, 55 and 55.5 years. In all groups, the commonest aetiology was iatrogenic following partial or total thyroidectomy. Of the 171 patients, 106 (62%) recovered vocal fold motility. More patients in the early group recovered (65%) than in the intermediate (61%) and delayed (56%) groups. Among patients whose vocal cord paralysis persisted in spite of voice therapy, 93% of those in the early group and 95% of those in the intermediate group showed glottal compensation after treatment. By contrast only 26% of patients in the delayed group showed glottal compensation. Significant improvements in the mean values of Jitter and Shimmer were also noted in the early and intermediate groups following voice therapy but not in the delayed group. On the basis of these results, the authors recommend that voice therapy should ideally commence within four weeks after the onset of vocal cord paralysis to optimise functional outcomes, and certainly no later than two months after injury. Whilst a prospective, randomised trial might provide more robust evidence to support this assertion, it would certainly seem reasonable to prioritise patients with unilateral vocal cord paralysis for early initiation of voice therapy.