Two hundred one serum samples from individual dairy cows with a range of results on initial testing with a commercial Johne's disease enzyme-linked immunosorbent assay (ELISA) kit were repeat tested 5 times in each of 2 laboratories with kits produced by the same manufacturer. The results for the samples with all 10 replicates showed that the values for individual samples often had a coefficient of variation greater than 20%. As expected, the standard deviation for the results increased as the average value increased and the coefficient of variation was greater in samples with low mean values. The different lots of the commercial ELISA kit used in this study had a significant effect on both the optical density and the calculated sample to positive (S/P) ratio for test replicates. Based on the variability detected in S/P ratios of replicate samples, application of a single cutoff point to interpret individual test results as positive or negative for antibodies to Mycobacterium avium subspecies paratuberculosis could result in inconsistent classification of animals as positive or negative for Johne's disease. Such inconsistency in test interpretation leads to frustration in large animal veterinarians or producers trying to make management decisions based on individual test results. Instead of dichotomizing the test results as positive or negative based on a single cutoff value, reporting numerical values and supplying a classification scheme that includes a suspect category reflecting the uncertainty inherent in the test is recommended to provide more reliable result interpretation.