Seascape variability may confound assessments on the effectiveness of no-take marine reserves (NTMRs) in conserving biodiversity. In most cases baseline data are lacking, resulting in evaluations of NTMR effectiveness being Control Impact (CI) assessments. Even with independent replicate areas among management zones, this approach can make it difficult to detect zone effects if seascape attributes, such as habitat structural complexity varies among experimental areas. To determine the importance of structural complexity in evaluations of NTMR effectiveness we performed assessments on the abundance of a targeted fish, yellowtail kingfish (Seriola lalandi), in the Lord Howe Island Marine Park (LHIMP). We compared assessments which did and did not account for structural complexity, quantified using high resolution multibeam bathymetry. Despite almost 3 times more S. lalandi in NTMRs, the traditional CI assessment explained only 3% of the variation in the abundance of S. lalandi and revealed no clear effect of protection. Incorporating structural complexity into the assessment increased the deviance explained to 65% and uncovered an important interaction between zone and structural complexity. Greater abundances of S. lalandi were detected in NTMRs compared to fished zones but only on highly complex reefs. By accounting for structural complexity, we demonstrate that the precision and accuracy of NTMR assessments can be improved, leading to a better understanding of ecological change in response to this conservation strategy. Consequently, where marine park zones vary greatly in structural complexity, we strongly advocate for quantifying and accounting for such variability in assessments of NTMR performance.