The identification of "driver mutations" in cancers initiated rapid development of targeted drugs for the clinic. Despite promising outcomes initially in patients, the ultimate success of oncogene-targeted drugs has been hampered by the redundancy and remarkable complexity of cellular signaling pathways. Two studies in Science Signaling show that understanding these intricate networks and considering them during tumor classification and drug design can better predict drug response. These studies exemplify the potential of using systems analysis and computational modeling approaches to improve therapeutic strategies and outcomes in cancer patients.