Costumer targeting is becoming critical for successful advertising and client retention. In this case we studied the performance of a targeting model in terms of adoption and profitability. The base model chooses a set of clients to offer them mobile products through a telephone campaign. Then, we proposed an enhanced model that incorporates social attributes extracted from the mobile phone network of each client. Additionally, we studied the influence that friend's prior adoption might have on their own adoption. It was discovered that the incorporation of social attributes does not improve the power of prediction of targeting models, but instead, what improves it is the combination of both models (base and enhanced model). Hence, to achieve effective targeting models it is necessary to combine models based on different classification techniques in order to increase the power of prediction. In this case, data aggregation in the training set and the combination of both models allow us to increase in 8% the model performance on average, reaching a level of hits of 89%.