In recent years, the adoption of the FAIR principles has achieved notable success. This progress has led to the development of numerous assessment tools originating from diverse fields of application, thus addressing diverse object types, interpretations and implementations. Given the plethora of proposals available, it is crucial for users to precisely understand these measures, compare them effectively, make informed choices, and accurately interpret the obtained measurements. To meet these needs, we propose UReFM, a generic framework to formally represent and analyze measures with a tree like structure, FAIRness measures being a representative example. Besides the benefit of homogenization and of fine grained analysis, it allows for the formal definition of three characteristic quantities: coverage, granularity and impact. This latter reflects how much a given principle contributes to the final score. Our experiments show how these quantities (i) contribute to explain different scores obtained by digital artifacts using two different state-of-the-art assessment engines, (ii) enable a comparative study of different FAIRness measures, independently of any digital artifact.