Abstract Common garden experiments are valuable to study adaptive phenomenon and adaptive potential, in that they allow to study local adaptation without the confounding effect of phenotypic plasticity. The QST?FST comparison framework, comparing genetic differentiation at the phenotypic and molecular level, is the usual way to test and measure whether local adaptation influences phenotypic divergence between populations. Here, we highlight that the assumptions behind the expected equality QST = FST under neutrality correspond to a very simple model of population genetics. While the equality might, on average, be robust to violation of such assumptions, more complex population structure can generate strong evolutionary noise. Synthesis: We highlight recent methodological developments aimed at overcoming this issue and at providing a more general framework to detect local adaptation, using less restrictive assumptions. We invite empiricists to look into these methods and theorists to continue developing even more general methods.