In this work, we present a comprehensive evaluation of the Agilent Bioanalyzer, a microfluidics-based electrophoretic device that was used for impurity testing of a monoclonal antibody (mAb). We compared the system to SDS-PAGE, both operated under non-reducing conditions and found a significant improvement of accuracy for the Bioanalyzer. In addition, the latter exhibited a larger assay range and lower limit of quantitation (LOQ) based on a predefined total error limit of +/-30%. However, during method qualification applying a three-factor nested design with two operators performing duplicate measurements per day, each on 4 different days, we observed unpredictable recurring quantitative outliers using the chip-based system. In-depth analysis on multiple runs with various chip lots confirmed the above finding and indicated that most likely on-chip dye labeling and/or post-column background fluorescence elimination are not compatible with the large size of the intact antibody as similar findings were observed for myosin used as upper marker for time correction. Interestingly, after reducing the intact antibody into light and heavy chain, we resolved the outlier issue. Eventually, requalification of the micro-fabricated analytical device under reducing conditions revealed only 1 out of 32 quality control samples (QCs) exceeding the +/-30% total error limits.