Proteomics will celebrate its 20th year in 2014. In this relatively short period of time, it has invaded most areas of biology and its use will probably continue to spread in the future. These two decades have seen a considerable increase in the speed and sensitivity of protein identification and characterization, even from complex samples. Indeed, what was a challenge twenty years ago is now little more than a daily routine. Although not completely over, the technological challenge now makes room to another challenge, which is the best possible appraisal and exploitation of proteomic data to draw the best possible conclusions from a biological point of view. The point developed in this paper is that proteomic data are almost always fragmentary. This means in turn that although better than an mRNA level, a protein level is often insufficient to draw a valid conclusion from a biological point of view, especially in a world where PTMs play such an important role. This means in turn that transformation of proteomic data into biological data requires an important intermediate layer of functional validation, i.e. not merely the confirmation of protein abundance changes by other methods, but a functional appraisal of the biological consequences of the protein level changes highlighted by the proteomic screens.