Quantitative nuclear magnetic resonance (qNMR) spectroscopy is employed by an increasing number of analytical and industrial laboratories for the assignment of content and quantitative determination of impurities. Within the last few years, it was demonstrated that (1)H qNMR can be performed with high accuracy leading to measurement uncertainties below 1 % relative. It was even demonstrated that the combination of (1)H qNMR with metrological weighing can lead to measurement uncertainties below 0.1 % when highly pure substances are used. Although qNMR reference standards are already available as certified reference materials (CRM) providing traceability on the basis of (1)H qNMR experiments, there is an increasing demand for purity assays on phosphorylated organic compounds and metabolites requiring CRM for quantification by (31)P qNMR. Unfortunately, the number of available primary phosphorus standards is limited to a few inorganic CRM which only can be used for the analysis of water-soluble analytes but fail when organic solvents must be employed. This paper presents the concept of value assignment by (31)P qNMR measurements for the development of CRM and describes different approaches to establish traceability to primary Standard Reference Material from the National Institute of Standards and Technology (NIST SRM). Phosphonoacetic acid is analyzed as a water-soluble CRM candidate, whereas triphenyl phosphate is a good candidate for the use as qNMR reference material in organic solvents. These substances contain both nuclei, (1)H and (31)P, and the concept is to show that it is possible to indirectly quantify a potential phosphorus standard via its protons using (1)H qNMR. The same standard with its assigned purity can then be used for the quantification of an analyte via its phosphorus using (31)P qNMR. For the validation of the concept, triphenyl phosphate and phosphonoacetic acid have been used as (31)P qNMR standards to determine the purity of the analyte tris(2-chloroethyl) phosphate, and the resulting purity values perfectly overlap within their expanded measurement uncertainties.