Enterprise fraud costs hundreds of millions of dollars every year. Enterprise fraud management (EFM) solutions analyze transaction data to see if it matches known fraudulent patterns. If it does, an alert is raised, and a fraud investigator is assigned to the alert. EFM solutions typically are sold as software products with a fixed-term license. We highlight best practices for these product-based solutions and consider an alternative architecture that leverages cloud computing. Cloud computing is an emerging system architecture where the user accesses software services using the Internet. Users only need Internet connectivity to utilize computation and storage services. Users needn’t know about the inner workings of the server software or the location of the servers. Thus, cloud computing relieves the user from the burdens of designing and building an in-house server architecture and purchasing licenses to run software. In this paper, we develop a framework that assesses the costs and benefits of enterprise fraud management solutions. We extend our framework to weigh the costs and benefits of EFM solutions deployed as a licensed product and also in the cloud. Finally, we present a costbenefits analysis of comparable EFM solutions deployed using these two system architectures.