Objective: To compare speech outcome following different sequencing of hard and soft palate closure between arms and centers within trial 3 and compare results to peers without cleft palate. Design: A prospective randomized clinical trial. Setting: Two Norwegian and 2 British centers. Participants: One hundred thirty-six 5-year-olds with unilateral cleft lip and palate were randomized to either lip and soft palate closure at 3 to 4 months and hard palate closure at 12 months (arm A) or lip and hard palate closure at 3 to 4 months and soft palate closure at 12 months (arm D). Main Outcome Measures: A composite measure of velopharyngeal competence (VPC), overall assessment of VPC from connected speech (VPC-Rate). Percentage of consonants correct (PCC), active cleft speech characteristics (CSCs), subdivided by oral retracted and nonoral errors, and developmental speech characteristics (DSCs). Results: Across the trial, 47% had VPC, with no statistically significant difference between arms within or across centers. Thirty-eight percent achieved a PCC score of >90%, with no difference between arms or centers. In one center, significantly more children in arm A produced >3 active CSCs (P < .05). Across centers, there was a statistically significant difference in active CSCs (arm D), oral retracted CSCs (arm D), and DSCs (arms A and D). Conclusions: Less than half of the 5-year-olds achieved VPC and around one-third achieved age-appropriate PCC scores. Cleft speech characteristics were more common in arm A, but outcomes varied within and across centers. Thus, outcome of the same surgical method can vary substantially across centers. Keywords RCT, 2-stage procedures, UCLP, speech