Virtual Decision Making Unit (DMU) is a new technique and it has been applied on Data Envelopment Analysis (DEA) literature to enhance the discriminatory power in order to produce more comprehensive rankings. In the large set of Virtual DMUs, ideal DMU (IDMU) idea is widely used. An IDMU is a fictional unit which consumes the lowest inputs to produces the highest outputs levels. However, due its unreal performance, it is fair to inquire if IDMU is plausible. Therefore, we propose the quasi ideal DMU (quasIDMU) concept. At a glance, a quasIDMU is built by averaging all inputs and outputs and testing on a DEA model whether this DMU has a 100% efficiency score or not. In order to verify quasIDMU concept, a secondary goal cross efficiency model is used both with IDMU and quasIDMU. Benevolent and aggressive secondary goals are also provided in order to enhance analysis. Data are sourced on grades of Brazilian Institutes of Higher Education (IES). Our results shall be shown in congress.