Currently the 25-hydroxy vitamin D (25(OH)D) concentration is thought to be the best estimate of the vitamin D status of an individual. Unfortunately, its measurement remains complex, despite recent technological advances. We evaluated the biological variation (BV) of 25(OH)D in order to set analytical performance specifications (APS) for measurement uncertainty (MU). Six European laboratories recruited 91 healthy participants. The 25(OH)D concentrations in K3-EDTA plasma were examined weekly for up to 10 weeks in duplicate on a Lumipulse G1200 (Fujirebio, Tokyo, Japan). The linear regression of the mean 25(OH)D concentrations at each blood collection showed that participants were not in a steady state. The dissection of the 10-sample collection into two subsets, namely collections 1–5 and 6–10, did not allow for correction of the lack of homogeneity: estimates of the within-subject BV ranged from 5.8% to 7.1% and the between-subject BV ranged from 25.0% to 39.2%. Methods that would differentiate a difference induced by 25(OH)D supplementation at p < 0.05 should have MU < 13.6%, while at p < 0.01, the MU should be <9.6%. The development of APS using BV assumes a steady state of patients. The findings in this study suggest that patients are not in steady state. Therefore, APS that are based on MU appear to be more appropriate.