In veterinary multi-site studies, the detrimental impact of unequal site enrollment on statistical power is examined under various conditions and design features. Specifically, when the response variable is binary, we explore the effect of site imbalance in the presence of different magnitudes of treatment effect size and site-by-treatment interaction effects. The number of sites in relation to total sample size will also be considered. Finally, the use of unbalanced randomization is commonly believed to exacerbate the impact of unequal site sizes. Based on these findings, we discuss general approaches to mitigating the impact of unequal site enrollment.