Falcon 012009

Fair Isaac, Falcon Fraud Manager and Debit Card Fraud Ryan Ganji Falcon Product Management Global Fraud Solutions [email_address]

Fair Isaac Overview Falcon’s Contribution to Combating Fraud Evolving Fraud, Multiple Channels and Sophisticated Fraudsters Falcon Fraud Manager Overview Protecting Debit with Falcon Fraud Manager © 2008 Fair Isaac Corporation. Confidential. Agenda

We make decisions smarter Fair Isaac is the leader in Enterprise Decision Management powered by advanced analytics #1 in Worldwide Analytic Applications Software Revenue

Fair Isaac Fraud Solutions IP and analytic innovation: 25 fraud patents (including dynamic profiling) - additional 50 patents pending New breakthroughs such as adaptive analytics Industry collaboration Consortium that helps whole industry identify fraud schemes Fraud Alert Network to share best practices Domain expertise 16 years helping 4,000+ fraud clients all over the world Professional Fraud Consulting Services - more than 20 countries, 6 continents Over 100 neural net models distributed in 20 countries Proven strength: Reduced card fraud losses by 66% in US from 1992 to 2005 (measured in basis points)

Fair Isaac Card Fraud Solutions Falcon™ Fraud Manager Protects over 1.8 billion active accounts worldwide Processes approximately 65% of credit card transactions worldwide, 90% in US 20 of the top 25 financial Institutions worldwide – 9 of the top 10 card issuers in the US Provides fraud detection on credit cards, debit cards, deposits, on line activity, application fraud and merchant acquiring Card Alert : Card Alert analyzes traffic for over 11,000 financial institutions – 400,000 ATMs Card Alert monitors 60% of US ATM and PIN POS transactions. ATM coverage is over 70% and PIN POS is 50%

Fair Isaac’s Falcon Software Has Contributed To The Industry’s Success Combating Fraud U.S. fraud losses, as a percentage of sales, have declined from a peak level of over 18 basis points in 1992, to 6.25 basis point in 2005 Falcon Introduced Credit Card Fraud Source: Nilson data

Fraud’s Four Key Dimensions Deposit Account Securities Loans Criminal Employee Criminal Outsider ATM POS Teller Window Telebanking and Call Center PC Account File Changes Info Gathering Account Debit Account Credit Mail Account Opening Mobile Credit Card Criminals Products Access Methods Activity Customer

Growing Complexity, Changing Risks New sophisticated fraud across channels and lines Regulatory mandates for better risk management Criminals evolve and organize Sophisticated, worldwide criminal organizations Frequent pattern changes New criminal tools and new channels New compromise methods New payment types Social Engineering Internet Attack Bots Data breaches / mass compromise a recurring risk Lost, stolen, counterfeit cards remain concern Check fraud still high First-party (deposit) Third-party (forgery) Insider fraud top of mind The next unknown risk More Channels More Lines of Business More Regulations More Technology More Change

Fair Isaac Protects the Whole Lifecycle Account Opening Account Access Real-time Monitor and Response Suspect Case Management Investigations Recovery Fair Isaac Fraud Solutions Lost and Stolen First Party Identity Theft Flash Fraud Cross-Channel Cross-Line Phishing Employee Collusion Hacking Fraud in New Markets Counterfeit Mass Compromise Premeditated Credit Abuse

Fair Isaac Falcon Fraud Manager Solution

Protecting Against Payment Card Losses Sophisticated Analytics Case Management Patented Profiles Rules Management Review, action and document suspicious transactions Customizable, user-friendly interface Improves operational efficiency Establish work queues Reporting Regional and industry-specific Neural Network models Consortium-based models utilize data from billions of transactions Adaptive Analytics. Neural Network and Outlier Models Trained by specialized fraud scientists with hundreds of man years experience Provides customization of solution Establish case creation and authorization strategies Create real-time without IT involvement through user-friendly interface Combat ‘flash fraud’ Falcon ® Fraud Manager Compares historical behavior with recent patterns Developed at the cardholder level New Profiling: Intelligent and Calibration Provides lower false positive rates to minimize impact on good customers Multiple profiles provide significant lift: ATM, POS, merchant

Account Holder Profile Falcon Fraud Manager today creates Account Profiles to detect unusual cardholder behavior A cardholder profile includes: First stage application model output Time-based variables illustrating user behaviour over time (e.g. average number of transactions per week) Numeric variables indicating volumes. (e.g total spend last day) Purchase data (type of goods purchased) Ratios and Rates (combinations of inputs to optimise information value, e.g. fraction of cash transactions)

What Makes a Good Fraud Management Solution? Statistical representation of likelihood of fraud Rank order risk systematically Finely-tuned to your specific operating objectives Precise fraud detection, minimizing false-positives Flexibility to manage various organizational strategies, constraints, and objectives Ability to evaluate fraud risk in relevant time Effective fraud management solutions provide: Reduced losses Increased profitability Improved automation Enhanced customer service Appropriate resource allocation

Benefits of Analytics Focus effort on transactions with highest likelihood of fraud Good transactions Fraud transactions No protection – fraud/good transactions mixed and difficult to identify Rules-based solution - some fraud captured, but much is missed and still lot of good transactions captured as well Analytics-based solution – most of fraud captured, with minimal good transactions

Fraud Accounts Score Higher Score Range Illustrative Score Distribution and Detection Rates Fraud Accounts Non-Fraud Accounts

Falcon Fraud Manager 6.0 The Next Generation Enables Multiple Profiles Per Transaction Falcon 6 will expand Fair Isaac’s patented profiling technology to capture a broader view of debit transactions across lines of business and multiple channels, impacting a customer or their account. Intelligent Global Profiling provides substantial performance improvement by profiling not only cardholders, but a wide range of other entities. Multiple profiles help detect unusual behavior of other entities in the financial services environment such as ATMs, traditional swipe, POS devices, mobile phones and telecom switches

Multiple Profiles and Fraud Detection – Cont’d Intelligent profiles are “global” because while they are involved in cardholder transactions, they are NOT specific to any cardholder Intelligent Profiles enable FFM to detect fraudulent activity that may not look suspicious from the point of cardholder behavior patterns alone. These advancements in profiling improve Falcon performance by providing additional transactional views that provide a more complete picture of fraudulent activity.

Profiling Advancements Multiple Profile/UDV Capability Falcon 6 will enable the ability to create, read and update multiple profiles per transaction Cardholder Profiles Customer Profiles Global Profiles Calibration Profiles Tuning parameters for score calibration Intelligent Profiles Real time concise profiling will allow only high interest entities to be profiled to improve fraud detection Scoring on a “global” entity, such as device profiling for debit. Shared memory profile support Profiles for high interest items (i.e., device profiling) can be held in shared memory to maximize performance Provides fast response times to solutions using concise profiles

Patented Profiling System $ $ $ $ * $ $ # $ Falcon Profiling System Deposits Non-monetary Transactions ATM & POS Authorizations Customer Master Changes # Updated Profile To models

Analytic Advancements Adaptive Analytics Feedback score that will be powered by fraud and non-fraud decisions. Second score that will provide a view of the current transaction risk relative to recent fraud and non-fraud dispositions. Secondary score will enable fall back to original Falcon score Outlier Models Deployment capability Outlier models may be useful where sufficient data doesn’t exist for neural network model training

Enterprise Fraud Open Platform Bank Data Warehouse Online Banking & Telephony Access Branch ACH & Wire Transfers Application Originations Credit Card Transactions Channel Models, Rules, Profiles Scoring Module Source Systems Bill Pay Checks On us & Deposits Customer Model, Merchant Model, Rules, Profiles Research Module Modules Control APIs, rules, queues, hotlists Management Module Mobile Payments Fraud Risk Research Analysis & Investigations Debit Card Transactions Third-Party Case Intervention Case Database Case Module Intervention Case Database 3 rd Party Data Third-Party Detection Engine Consortium Outlier Bank Third-Party … Acquirer

Falcon Fraud Manager for Debit

Incidence of Debit Fraud Growing INCIDENCE RATES FOR EXISTING CARD FRAUD BY TYPE Source: Javelin Strategy & Research, 2008 Identity Fraud Report

Amount of Debit Fraud Increasing TOTAL ANNUAL FRAUD AMOUNTS FOR EXISTING CARD FRAUD BY TYPE $ Billion Source: Javelin Strategy & Research, 2008 Identity Fraud Report

Today’s Fraud – A Convergence of Threats In today’s challenging risk environment, debit card portfolios are under siege Debit cards being used for other fraud scenarios such as Deposit Fraud Typically not recognized as a debit card loss (often managed elsewhere in an organization) Data security issues Consumer behavior – easy to guess PIN #’s, enter PIN into anything mentality Signature- Based POS Fraud PIN-Based POS Fraud ATM Fraud Deposit Fraud VIA Debit Card

Debit PIN Fraud Trends Debit card usage continues to grow and outpace credit card usage Increased usage translates to larger opportunities for criminals, resulting in… An extremely challenging fraud environment for debit cards and deposit accounts Rapid evolution in criminal focus Shift in targeted segments: PIN-BASED FRAUD Gives criminals opportunity for cash Criminals have evolved their methodology Increased sophistication resulting in more data acquisition and aggregation VoIP phishing, pharming, man in the middle attacks Obtaining multiple pieces of customer data from various sources and creating their own data warehouses of stolen identities and information Using stolen data faster

Why Debit Fraud? Criminals are targeting DDA accounts via ATM No ID or embossing No fencing - A $500 ATM withdrawal is less risk than purchasing and fencing a big screen TV for the same reward More difficult to identify transactions as fraud Credit card fraud less lucrative than in the past Bank defenses are not as hardened on DDA More channels to target the accounts, and those channels are not as secure Few banks have the ability to tie fraud that crosses channels “ That’s where the money is.” – Willie Sutton

Debit PIN Fraud Types Compromise fraud Family/Friendly fraud Lost/Stolen Counterfeit Bad check deposits Electronic Funds ATM Transfer fraud PIN-less Bill Pay fraud

PIN Points of Compromise Source: Fair Isaac CardAlert Fraud Manager PIN Points of Compromise 2001-2007 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2001 2002 2003 2004 2005 2006 2007 ATM-Bank ATM -NonBank POS

POS Compromises PIN POS terminals and pay at the pump Skimmer wired inside the POS terminal, attack is 5 years old, some wireless transmitters and video cameras PIN compromised prior to encryption Transaction goes through the system Takes about a minute to swap out terminal Los Angeles criminal group sending tiger teams out to PA, DE, NC, NV, CA, RI, WA, FL, OR Recent EPP and PED POS standards limit this type attack, however these devices will remain with us for several more years

ATM Compromises EPP is forcing criminals to capture PINS on the other side of the PIN pad Euro criminals have always had to deal with FI ATMs Chip & PIN = cross border fraud + card not present ATMs compromised for unencrypted card track information only Combining card track breaches with alternative PIN capture Phishing Online banking IVR Hardware solutions

ATM Points of Compromise Source: Fair Isaac CardAlert Fraud Manager

2007 FI ATM Points of Compromise

ATM/POS Profiling Profiling cardholder and device provides complete profile picture for a transaction Looks for unusual device behavior Large amounts Rapid transactions Suspicious patterns of transaction types Many patterns involve multiple cards that are best detected by device profiling Strongly predictive variables can be created Device score can be combined with an account score to improve fraud detection

Predictive Lift from Device Profiling 80% relative performance lift in real-time value detection at a 10-to-1 false positive rate for Debit Cash Transactions Cardholder + ATM Profiles Cardholder Profiles

Predictive Lift from Merchant Profiles 40% relative performance lift in real-time value detection at a 10-to-1 false positive rate Cardholder + Merchant Profiles Cardholder Profiles

Falcon ® Fraud Manager Reduces Fraud Losses Fraud is a major problem worldwide Increased variety and use of cards is accompanied by more sophisticated fraud methodologies Financial institutions must deploy an effective fraud management system to stay competitive

Falcon ® Fraud Manager Reduces Fraud Losses A neural network-based fraud management system provides maximum protection against fraud losses. Falcon Fraud Manager from Fair Isaac has helped the leading financial institutions minimize fraud losses through sophisticated profiles , advanced modeling , and comprehensive case management capabilities

Working Together Fair Isaac is committed to working with our clients in new and innovative ways Our commitment to you: We will innovate and collaborate with you We will execute with integrity in all of our dealings with you We will deliver high performance, delivering on our promises Our requests of you: We ask that you join us in our efforts to innovate and collaborate with you: Participate with us in crafting solutions and contributing data to better meet your needs Participate as early adopters of our new and innovative solutions – in software and analytics

Falcon 012009

More Related Content

What's hot (20)

Viewers also liked (17)

Similar to Falcon 012009 (20)

Recently uploaded (20)

Falcon 012009

Editor's Notes