AbstractMarine diazotrophs fix dinitrogen gas into bioavailable nitrogen that drives the ocean nitrogen cycle; yet, efforts to infer global diazotroph distributions have been limited by a sparsity of observations. In situ measurements of nifH gene abundance (essential for nitrogen fixation) are increasingly being used to inform the biogeography of diazotrophs. However, comparing such gene abundances spatially, temporally and between diazotroph species remains difficult. We synthesize existing data on gene‐to‐cell and cell‐to‐biomass conversions for four major diazotroph groups to convert nifH gene counts to abundance‐ and biomass‐based biogeographic “currencies.” Results suggest up to two orders of magnitude uncertainty converting from nifH gene abundance to cell abundance, and up to four orders of magnitude uncertainty from nifH gene abundance to biomass. Uncertainty arises due to large taxonomic variation in cell size and presumed polyploidy, that is, variability in the number of genomes per cell. Such uncertainties hinder comparing biogeographies of different species. Additionally, numerical models need biogeographies for validation, typically in the currency of carbon biomass. Here, we show that conversion uncertainty from nifH gene abundance to biomass overwhelms biomass variability simulated in such models. These results demonstrate a basic currency problem in converting gene abundance observations to biogeographically meaningful quantities for synthesizing studies and modeling approaches. Such issues may also have relevance to other genes and organisms beyond diazotrophs. To avoid biases in interpreting gene counts as a measure of abundance, we suggest converting gene counts to a binary presence/non‐detect metric to map broad biogeographical distributions more robustly.