The French law adopted in 2018 on the freedom to choose a professional future seeks to shift from an obligation of means to an obligation of results in terms of the gender pay gap. It is based on the notion that discriminations can be measured, providing quantification as a means to steer action. The article recalls that in fact, this goal is unachievable from a statistical point of view, coming back to the two most widely used approaches: “all other things being equal” statistical analysis and statistical testing. It explains why these tools cannot quantify the scope of discrimination, without denying their broad usefulness. The lack of measurement does not mean giving up on the need for action. Rather, it should open the path for ways of thinking and interdisciplinary approaches that do not rely on numbers for proof.