Performance management of manufacturing systems requires applicable methodologies to be implemented in practice. Both the literature and the experiments we led in this paper show the need of a precise methodological tool to support the setting-up of dashboards. In a previous work, we proposed a KPI meta-model framework as a means to implement KPI computations in practice. In this article we go one step further and focus on the methodological aspects of this framework to build performance dashboards using the DMAIC (Define, Measure/Analyse, Improve/Control) method from 6 sigma. Starting from strategic objectives, we drive KPI trees until measure means to collect performance feedback. The methodology has been evaluated by an experimentation with students on the Lampex learning factory from Strasbourg University.