<p>Complex assays such as immunoassays can be affected by robustness and ruggedness factors. Associated risks can be reduced by a systematic performance assessment across key factors followed by control strategies. However, the large number of factors and their interactions can represent an experimental challenge. Statistical Design of Experiments (DOE) allows efficient evaluation of more factors with fewer total assay runs in addition to assessing potential factor interactions. We applied DOEs to the robustness evaluation of a vaccine potency ELISA. Test factors were selected based on a review and ranking of development data, scientific experience, and commonly expected sources of variability. Comparing different design options with 16–20 runs which was a laboratory limit, a 16-run Resolution III design was selected based on the total number of runs, the degree of factor confounding, and the potential projection properties. DOE data were first visually analyzed by plotting the concentration-responses of reference curves against DOE Runs followed by detailed statistical models of the maximum fluorescent curve signal and the WRMSE fit values. Initial confounding between factors and their interactions was reduced by eliminating factors with no impact from the models and by removing factors or interactions based on their likelihood of an impact after applying statistical and scientific expertise. Despite initial confounding, the designs allowed discerning an impact of plate manufacturer with interaction of coating concentration and time out of 15 factors with only 16 runs.</p> Graphical Abstract <p></p>