Cybercriminals often leverage Bitcoin for their illicit activities. In this work, we propose back-and-forth exploration, a novel automated Bitcoin transaction tracing technique to identify cybercrime financial relationships. Given seed addresses belonging to a cybercrime campaign, it outputs a transaction graph, and identifies paths corresponding to relationships between the campaign under study and external services and other cybercrime campaigns. Back-and-forth exploration provides two key contributions. First, it explores both forward and backwards, instead of only forward as done by prior work, enabling the discovery of more addresses and relationships. Second, it prevents graph explosion by combining a tagging database with a machine learning classifier for identifying addresses belonging to exchanges. We evaluate back-and-forth exploration on 30 malware families. We build classifiers for four families using Bitcoin for C&C and use them to demonstrate that back-and-forth exploration identifies 13 C&C signaling addresses missed by prior work. Our approach uncovers a wealth of services used by the malware including 43 exchanges, 11 gambling sites, 5 payment service providers, 4 underground markets, 4 mining pools, and 2 mixers. It also identifies relationships between the malware families and other cybercrime campaigns highlighting how some malware operators participate in a variety of cybercriminal activities.