Extraction of the lipid fraction is a key part of acquiring lipidomics data. High-throughput lipidomics, the extraction of samples in 96w plates that are then run on 96 or 384w plates, has particular requirements that mean special development work is needed to fully optimise an extraction method. Several methods have been published as suitable for it. Here, we test those methods using four liquid matrices: milk, human serum, homogenised mouse liver and homogenised mouse heart. In order to determine the difference in performance of the methods as objectively as possible, we used the number of lipid variables identified, the total signal strength and the coefficient of variance to quantify the performance of the methods. This showed that extraction methods with an aqueous component were generally better than those without for these matrices. However, methods without an aqueous fraction in the extraction were efficient for milk samples. Furthermore, a mixture containing a chlorinated solvent (dichloromethane) appears to be better than an ethereal solvent (tert-butyl methyl ether) for extracting lipids. This study suggests that a 3:1:0.005 mixture of dichloromethane, methanol and triethylammonium chloride, with an aqueous wash, is the most efficient of the currently reported methods for high-throughput lipid extraction and analysis. Further work is required to develop non-aqueous extraction methods that are both convenient and applicable to a broad range of sample types.