BackgroundWithin the plant kingdom, there is an exceptional amount of chemical diversity that has yet to be annotated. It is for this reason that non-targeted analysis is of interest for those working in novel natural products. To increase the number and diversity of compounds observable in root exudate extracts, several workflows which differ at three key stages were compared: 1) sample extraction, 2) chromatography, and 3) data preprocessing. ResultsPlants were grown in Hoagland's solution for two weeks, and exudates were initially extracted with water, followed by a 24-h regeneration period with subsequent extraction using methanol. Utilizing the second extraction showed improved results with less ion suppression and reduced retention time shifting compared to the first extraction. A single column method, utilizing a pentafluorophenyl column, paired with high-resolution mass spectrometry ionized and correctly identified 34 mock root exudate compounds, while the dual column method, incorporating a pentafluorophenyl column and a porous graphitic carbon column, retained and identified 43 compounds. In a pooled quality control sample of exudate extracts, the single column method detected 1,444 compounds. While the dual method detected fewer compounds overall (1,050), it revealed a larger number of small polar compounds. Three preprocessing methods (targeted, proprietary, and open source) successfully identified 43, 31, and 38 mock root exudate compounds to confidence level 1, respectively. SignificanceEnhancing signal strength and analytical method stability involves removing the high ionic strength nutrient solution before sampling root exudate extracts. Despite signal intensity loss, a dual column method enhances compound coverage, particularly for small polar metabolites. Open-source software proves a viable alternative for non-targeted analysis, even surpassing proprietary software in peak picking.