Shotgun lipidomics is a powerful tool that enables simultaneous and fast quantification of diverse lipid classes through mass spectrometry based analyses of directly infused crude lipid extracts. We present here a shotgun lipidomics platform established to quantify 38 lipid classes belonging to four lipid categories present in mammalian samples and show the fine-tuning and comprehensive evaluation of its experimental parameters and performance. We first determined for all the targeted lipid classes the collision energy levels optimal for the recording of their lipid class- and species-specific fragment ions and fine-tuned the energy levels applied in the platform. We then performed a series of titrations to define the boundaries of linear signal response for the targeted lipid classes, and demonstrated that the dynamic quantification range spanned more than 3 orders of magnitude and reached sub picomole levels for 35 lipid classes. The platform identified 273, 261, and 287 lipid species in brain, plasma, and cultured fibroblast samples, respectively, at the respective optimal working sample amounts. The platform properly quantified the majority of these identified lipid species, while lipid species measured to be below the limit of quantification were efficiently removed from the data sets by the use of statistical analyses of data reproducibility or a cutoff threshold. Finally, we demonstrated that a series of parameters of cell culture conditions influence lipidomics outcomes, including confluency, medium supplements, and use of transfection reagents. The present study provides a guideline for setting up and using a simple and efficient platform for quantitatively exploring the mammalian lipidome.