In this talk we review reduced order modelling techniques for efficient computational and geometrical reduction with a special focus on computational fluid dynamics parametrized problems. Parameters may be physical or geometrical. The proposed computational reduction of parametrized PDEs is carried out by stable reduced basis method, certified by rigorous error bounds; geometrical reduction is performed thanks to efficient shape parametrization to model boundaries and/or interfaces by means of free-form deformation or radial basis functions. The proposed framework is suited for the solution of inverse problems, like shape optimization, optimal flow control, data assimilation, parameter estimation and uncertainty quantification and in general in the solution of iterative and repetitive problems as well as in the real time context.