Metabolomics is a rapidly advancing field, and much of our understanding of the subject has come from research on cell lines. However, the results and interpretation of such studies depend on appropriate normalization of the data; ineffective or poorly chosen normalization methods can lead to frankly erroneous conclusions. That is a recurrent challenge because robust, reliable methods for normalization of data from cells have not been established. In this study, we have compared several methods for normalization of metabolomic data from cell extracts. Total protein concentration, cell count, and DNA concentration exhibited strong linear correlations with seeded cell number, but DNA concentration was found to be the most generally useful method for the following reasons: (1) DNA concentration showed the greatest consistency across a range of cell numbers; (2) DNA concentration was the closest to proportional with cell number; (3) DNA samples could be collected from the same dish as the metabolites; and (4) cell lines that grew in clumps were difficult to count accurately. We therefore conclude that DNA concentration is a widely applicable method for normalizing metabolomic data from adherent cell lines.