Microfluidic CE with conductivity detection platforms could have an impact on the future development of smaller, faster and portable devices. However, for the purpose of reliable identification and quantification, there is a need to understand the degree of irreproducibility associated with the analytical technique. In this study, a protocol was developed to remove baseline drift problems sometimes observed in such devices. The protocol, which consisted of pre-conditioning steps prior to analysis, was used to further assess measurement variability from 24 individual microchips fabricated from six separate batches of glass substrate. Results show acceptable RSD percentage for retention time measurements but large variability in their corresponding peak areas (with some microchips having variability of approximate to 50%). Sources of variability were not related to substrate batch but possibly to a number of factors such as applied voltage fluctuations or variations in microchannel quality, for example surface roughness that will subsequently affect microchannel dimensions.