Raman spectroscopy provides chemical-rich information about the composition of analytes and is a powerful tool for biological studies. With the ability to investigate specific cellular components or image whole cells, compatible methods of sample preservation must be implemented for accurate spectra to be collected. Unfortunately, the effects of many commonly used sample preservation methods have not been explored with cultured cells. In this study, two human cell lineages of varying phenotypes were used to investigate the effects of sample preservation methods. Cells were cultured directly onto quartz substrates and either formalin-fixed, desiccated or air dried. The results indicate that the methodology applied to cell cultures for Raman analysis significantly influences the quality and reproducibility of the resulting spectral data. Formalin fixation was not found to be as universally efficient as anticipated for a commonly used fixative. This was due largely to the inconsistency in sample preservation between cell lines and loss of signal intensity. Sample air-drying was found to be largely inconsistent in terms of spectral reproducibility. Our study shows that sample desiccation displayed good spectral reproducibility and resulted in a good signal-to-noise ratio. Lipid and protein content in both activated and inactivated cells were maintained and provided a more controlled method compared with air-drying, revealing that the speed of drying is important for sample preservation.