Standardized sampling of fishes is either probabilistic (i.e., random) or non-probabilistic (i.e., fixed), with each having advantages and disadvantages. Our study objective was to evaluate whether a fixed site-survey design for early detection monitoring of non-native fishes could be improved by modifying fixed site selection and adding random sampling. Results were largely similar but varied by evaluation metric; accumulation curves of species suggested that performance of the fixed survey design was 10% higher than the modified survey and also accumulated species significantly faster from 1 to 42 units of effort. Results indicated the fixed site survey previously took advantage of the patchiness in fish distribution, leaving little potential efficiency to gain using the modified survey design. We would like to emphasize that the modified survey design was effective, just not as effective as the fixed site survey, and could likely optimize early detection monitoring (EDM) for programs with less available fish data to draw observations from or if only random sampling was used for monitoring.