We evaluated different in-solution and FASP-based sample preparation strategies for absolute protein quantification. Label-free quantification (LFQ) was employed to compare different sample preparation strategies in the bacterium Pseudomonas aeruginosa and human embryonic kidney cells (HEK), and organismal-specific differences in general performance and enrichment of specific protein classes were noted. The original FASP protocol globally enriched for most proteins in the bacterial sample, whereas the sodium deoxycholate in-solution strategy was more efficient with HEK cells. Although detergents were found to be highly suited for global proteome analysis, higher intensities were obtained for high-abundant nucleic acid-associated protein complexes, like the ribosome and histone proteins, using guanidine hydrochloride. Importantly, we show for the first time that the observable total proteome mass of a sample strongly depends on the sample preparation protocol, with some protocols resulting in a significant underestimation of protein mass due to incomplete protein extraction of biased protein groups. Furthermore, we demonstrate that some of the observed abundance biases can be overcome by incorporating a nuclease treatment step or, alternatively, a correction factor for complementary sample preparation approaches.