Towed-streamer marine broadband data have been key contributors to recent petroleum exploration history, in new frontiers and in mature basins around the world. They have improved the characterization of reservoirs by reducing the uncertainty in structural and stratigraphic interpretation and by providing more quantitative estimates of reservoir properties. Dedicated acquisition, processing, and quality control (QC) methods have been developed to capitalize on the broad bandwidth of the data and allow their rapid integration into reservoir models. Using a variable-depth steamer data set acquired in the Campos Basin, Brazil, we determine that particular care that should be taken when processing and inverting broadband data to realize their full potential for reservoir interpretation and uncertainty management in the reservoir model. In particular, we determine the QC implemented and interpretative processing approach used to monitor data improvements during processing and preconditioning for elastic inversion. In addition, we evaluate the importance of properly modeling the low frequencies during wavelet estimation. We find the benefits of carefully processed broadband data for structural interpretation and describe the application of acoustic and elastic inversions cascaded with Bayesian lithofacies classification, to provide clear interpretative products with which we were able to demonstrate a reduction in the uncertainty of the prediction and characterization of Santonian oil sandstones in the Campos Basin.