Numerical modeling may be an effective tool to predict the behavior of instable slopes provided that it relies on appropriate slope physical model. This is particular true for large-scale landslides, that generally feature intrinsic high spatial variability. For these large mass movements, it is crucial to assess the required spatial scale and input data accuracy for a correct numerical analysis of landslide. This paper investigates the role of the spatial scale and data accuracy in landslide modeling by considering as a testbed a Deep Seated Gravitational Slope Deformation (DSGSD) located in the North of Italy. Thanks to a large amount of available data, collected by means of geological and geophysical surveys, several numerical simulations with different resolutions (in term of geological and geomorphological settings) were performed. A comparison between the outcomes of the numerical models was accomplished through the use of a Multi Criteria Analysis (MCA) in order to evaluate landslide modeling with respect to numerical performance, investigations cost and time constraints. As a matter of fact, the high costs of detailed investigations often prevent an adequate slope assessment. Therefore, it is of primary importance to understand the role of input data on the effectiveness of numerical simulations. Hence, a decision support system capable to consider all these aspects is presented in order to define the best model as a trade-off between effectiveness, in terms of outcomes, and cost investigations.