Uncover and Prevent Fraud
Uncover and Prevent Fraud
Uncover and Prevent Fraud
www.feedzai.com
Next Generation Fraud Prevention Solution Stops Fraud Faster and More Accurately Than Ever Before
blocking capabilities, while achieving the industrys lowest percentage of false positives. Called FeedZai Fraud Prevention, the solution achieves these results by executing ultra-efficient analytics over big data using innovative in-memory architecture. FeedZai is so powerful that it learns and predicts financially optimized scoring thresholds for block transaction and approve but review transaction. Built on a highly flexible platform that includes pre-configured analytic models, dashboards, workflows and data connectors, the solution deploys and delivers value within two weeks. Thats just one of the reasons why Gartner named FeedZai a Cool Vendor in real time operational intelligence1.
1 Cool Vendors in Analytics and Business Intelligence, Bill Gassman et al., Research ID Number: G00211300, April 21, 2011
1
www.feedzai.com
Introduction
Creative criminals use a variety of tactics to perpetuate fraud against financial institutions, such as payment processors, acquiring and merchant banks, and credit card companies. Compounding the complexity of detecting and remediating fraudulent activities are evolving regulations and legislation as well as the unrelenting pressure to launchand protect from fraudnew financial products. While fraud prevention solutions have certainly become more sophisticated in their approach to uncovering fraud, they rarely enhance the customers experience, negatively impacting financial institutions ability to retain their best customers. These siloed and disjointed fraud management solutions cannot provide the customer-centric, enterprise-wide fraud prevention capabilities needed by financial institutions to halt fraud losses and grow revenues. Ironically, the emergence of big data presents both a problem and an opportunity for banks and clearinghouses. With its vast data sets, big data was supposed to provide financial institutions the insight needed to target fraud with scalpel-like precision to minimize losses and improve overall customer experience. But financial enterprises still struggle with how best to analyze, visualize, and correctly process tens of millions of transactions each hour at the speed of business. Because it often takes days, if not weeks, to detect and resolve suspected fraudulent activity, fraud management solutions are too slow and lack the precision needed to enhance, rather than degrade customer experience. That is, until now.
With FeedZai Fraud Prevention, financial institutions can leverage their investment in big data by applying real-time analytics to unprecedented amounts of structured, semi-structured and unstructured data
With FeedZai Fraud Prevention, financial institutions can leverage their investment in big data by applying real-time analytics to unprecedented amounts of structured, semi-structured and unstructured data, uncovering transaction anomalies in milliseconds. FeedZai creates profiles based on data gathered over the preceding two years for each credit card, merchant, wire transfer, automated teller machine, and historic fraud cases, analyzing far greater data sets than were previously possible. Real-time transactions are then scored against these rich historical baselines, and obvious fraudulent activity instantaneously identified and blocked while suspected fraud is automatically routed and tracked for manual review. FeedZai Fraud Prevention uses a complex-event processing engine for its core computation, enabling it to apply analytical techniques that are generally used only with data at rest to streams of
2
www.feedzai.com
business events as they arrive in motion. Its unique automatic base lining capability leverages historical data that describes normal for each KPI. Credit card companies and payment processors understand that its one thing to discover and halt a simple transaction anomaly, and quite another to associate the anomaly with the specific information that scored the transaction as suspicious and initiating deeper investigation. FeedZais pre-built and customizable case management workflow routes and tracks all suspicious cases until they are resolved, producing the industrys highest fraud detection percent and generating the lowest percent of false positives. As a result, this next generation EFM solution delivers increased productivity and lower operational costs, without negatively impacting the experiences of valued customers. Most EFM solutions require extensive customization, time-intensive analytic model training, and development of APIs to existing systems. FeedZais fraud prevention application is simply configured and quickly connects with existing systems with a robust set of pre-existing APIs. For instance, one European payment processor quickly configured and deployed FeedZai Fraud Prevention in a few weeks to protect its network of more than 15,000 automated teller machines and 200,000 point-of-sale devices, running more than 2 billion transactions per year and refreshing 300 KPIs per second. FeedZai finally delivers on the promise of big data. PIN and signature POS, and card-not-present transactions We asked three major vendors for a proof-of-concept of a realtime software solution capable of computing 260K real-time KPI/second for 50M transactions/month from 200,000 geographically dispersed terminals, FeedZai was the only vendor who could handle our data load requirements out of the box. We had it up and running in weeks instead of the 6-12 months required for other solutions. FeedZai not only met our requirements, but handled more volume and KPIs in real-time, compared to the real-time data with historical and reference values, and generated alarms on the fly.
Deposit fraud: Best practice rules and analytics for detection of suspicious incoming funds such as kiting and bust out fraud via electronic transactions to address schemes
Tracks anti-fraud department KPIs with pre-configured and easily customizable reports and dashboards. Today, most fraud departments use separate BI software to generate reports on their performance and achievements in fraud prevention, adding unnecessary administrative overhead to re-enter data and format reports.
and rules in less than 2 milliseconds for best-inclass in-transaction protection. Most fraud solutions are offline and require manual intervention and analysis.
users, cards, merchants and fraud for a comprehensive view of all available data when scoring a transaction. Other solutions often utilize averages and proxies for analytic models due to performance bottlenecks, yielding lower fraud detection rates and significantly higher false positive percents.
holders, terminals, merchants and fraud cases, enabling each transaction to be analyzed in
www.feedzai.com
has been compromised or behavior determined unusual. For example, if a clients profile were updated in the morning to record an ATM withdrawal in Paris, and two hours later a credit card transaction for the same client processed in San Francisco, this data is correlated and the likely fraudulent transaction appropriately scored. Card profiling: The card profiling module monitors card transactions, inferring the usual transaction amounts, merchants, countries and other relevant transaction attributes to properly generate a transaction score. This modules predictive capability is continuously and automatically improved using machine learning techniques. Merchant profiling: The merchant-profiling module determines normality as it relates to the merchant. With this data its possible to monitor merchants and their transactions to infer risk factors and other attributes that may impact scoring of an individual transaction. This module is also informed by machine learning algorithms. Fraud profiling: The fraud-profiling module is the centerpiece of the scoring process, and automatically recognizes patterns based on current transactions. Combined with the score provided by other profiling methods, it assigns ultra-accurate fraud scores to any given transaction. Fraud Prevention Optimizer: FeedZais Fraud Prevention Optimizer is the industrys first optimizer to provide companies complete control over the cost/benefit factors of fraud prevention. The Optimizer calculates expected results at different scoring thresholds for block, approve but review, and approve, and then graphically displays the volume and cost of manual reviews and fraud savings over any time period. The software uses predictive analytics that execute across each companys historical fraud metrics to predict expected
fraud detection rates and false positive rates. The Optimizer also automatically calculates threshold settings and delivers the highest return on investment for each company, while also allowing manual fine-tuning to adjust for vacation periods, staffing changes, etc. Second level rules: This module uses the scoring information from the profiling modulesas well as the optimal threshold information from the Optimizer moduleto determine the disposition of the transaction and whether to approve, approve with manual review, or block. Case management: The case management module is the front office of the entire FeedZai Fraud Prevention system and used daily by fraud analysts and system administrators. Case management serves to route and track suspected fraudulent transactions through manual review to disposition and reporting. Its features include: case workflow, a metrics dashboard, a reporting tool, and alert analysis workflow.
5
www.feedzai.com
Conclusion
With FeedZai Fraud Preventions highly advanced analytics and detection capabilities, financial institutions are dramatically reducing fraud by as much as 80 percent with lower false positive rates. There is no other off-the-shelf software solution on the market today that can analyze as much historical data against hundreds of thousands of KPIs in less than two milliseconds and produce the same results. As one FeedZai customer said, FeedZai was the only vendor that could handle our data load requirements out-of-the-box. We had it running in weeks instead of the six to 12 months required for other solutions. FeedZai not only met our requirements, but handled more volume and KPIs in real time, while comparing the real-time data with historical and reference values, and generating alarms on the fly. Download the complete Fraud Prevention case study to learn how Feedzai prevents credit card fraud in real-time before it occurs. For more information on Fraud Prevention or other FeedZai solutions, go to www.feedzai.com. Or contact us at either of our offices below:
crease in customer experience and satisfaction since fewer valid transactions were blocked or
Europe
6 University Way, London E16 2RD, United Kingdom Tel: +44.203.286.7590 info@feedzai.com
6
www.feedzai.com
Rev 6 11-13-2012