Milk and milk derivatives are common ingredients in food products. Undeclared milk is one of the leading causes of recalls in many countries, including the USA, and cases of allergic reactions have been reported due to unexpected exposures. There are commercial enzyme-linked immunosorbent assay (ELISA) kits available to the food industry to comply with the law by ensuring label accuracy and to identify potential sources of cross-contact. These kits are also used by regulatory agencies as part of their compliance programs. However, none of the commercial ELISAs for milk have been validated. Performance of ELISA kits for food allergens is affected by matrix, food processing, and stability and solubility of target proteins, among others factors. The performance of different commercial kits for milk allergens was evaluated by comparing a standard [National Institute of Standards and Technology (NIST) SRM #1549] spiked in wheat flour. We also compared the effect of food processing on detectability of milk proteins from incurred peanut butter cookies baked at various times. Kits differed in their ability to detect heat-treated milk proteins in baked cookies. Immunoblots clearly showed differences in antibody specificities and in their ability to detect proteins in processed foods. Factors such as undefined antibody specificity and differences in sample extraction solutions, materials used for calibrators, and reporting units contribute to variability of results among test kits and, hence, to increased uncertainty regarding the most appropriate use of the kits. Moreover, the use of incurred vs. spiked samples may affect protein recovery and, therefore, jeopardize the quantitative nature of the kit.