Dimensional analysis is a tool to quickly and reliably elaborate a dimensionless frame in which the experiments are favourably presented, because it is condensed as compared with the dimensional one. Further advantage lies in the 'scale-invariance' of the dimensionless frame, thus enabling the only reliable scale-up. There are only two real problems in dealing with dimensional analysis. The first is the construction of a complete list of relevant parameters which describe the process ('relevance list'). The second is the determination of the process characteristics and the establishment of the real operational number(s) particularly in case of large scale factors. This paper discusses possible pitfalls and their avoidance in the field of mixing.