Although LC-MS untargeted metabolomics continues to expand into exiting research domains, methodological issues have not been solved yet by the definition of unbiased, standardized and globally accepted analytical protocols. In the present study, the response of the plasma metabolome coverage to specific methodological choices of the sample preparation (two SPE technologies, three sample-to-solvent dilution ratios) and the LC-ESI-MS data acquisition steps of the metabolomics workflow (four RP columns, four elution solvent combinations, two solvent quality grades, postcolumn modification of the mobile phase) was investigated in a pragmatic and decision tree-like performance evaluation strategy. Quality control samples, reference plasma and human plasma from a real nutrimetabolomic study were used for intermethod comparisons. Uni- and multivariate data analysis approaches were independently applied. The highest method performance was obtained by combining the plasma hybrid extraction with the highest solvent proportion during sample preparation, the use of a RP column compatible with 100% aqueous polar phase (Atlantis T3), and the ESI enhancement by using UHPLC-MS purity grade methanol as both organic phase and postcolumn modifier. Results led to the following considerations: submit plasma samples to hybrid extraction for removal of interfering components to minimize the major sample-dependent matrix effects; avoid solvent evaporation following sample extraction if loss in detection and peak shape distortion of early eluting metabolites are not noticed; opt for a RP column for superior retention of highly polar species when analysis fractionation is not feasible; use ultrahigh quality grade solvents and "vintage" analytical tricks such as postcolumn organic enrichment of the mobile phase to enhance ESI efficiency. The final proposed protocol offers an example of how novel and old-fashioned analytical solutions may fruitfully cohabit in untargeted metabolomics protocols.