Credit Card Fraud Detection Client Presentation
Download as PPTX, PDF5 likes3,181 views
Tango Technologies, founded in 2008 with a global presence, specializes in fraud detection in the BFSI sector, helping businesses enhance growth, profitability, and customer satisfaction. The document details their innovative fraud analytics approach, including advanced techniques such as transaction-based scoring and global profiling, aimed at improving customer experience while minimizing fraud risks. It also highlights positive feedback from clients regarding their expertise in data science and successful project outcomes.
Credit Card Fraud Detection Client Presentation
- 2. AGENDA About Tango Technologies Fraud Analytics – An Overview Your Business, Our Understanding Tango's Approach We Sow, You Reap Why Tango?
- 3. About Tango Technologies Founded in 2008 Helping businesses to make better decisions that drive higher levels of growth, profitability and customer satisfaction. With 25 locations worldwide Tango has a global presence with its head quarters in Bangalore, India. Area of Expertise: Fraud detection in BFSI
- 4. About Tango Technologies Our Customers
- 5. Fraud Analytics What is Fraud? Criminal activity. Abuse of position, or false representation, or prejudicing someone's rights for personal gain. Traditional Methods. "Understand not only who your customer is, but who they’re becoming“. FICO Falcon Fraud Assessment System is the first Fraud Detection system using analytics.
- 6. Fraudulent Activities in BFSI Electronic fraud Identity theft Cheque fraud Credit/Debit card fraud oSkimming oBIN Attack oTele Phishing
- 7. Analytics in Credit Card Fraud Detection Risk Scoring based on Neural Network - Risk scoring is a system of predictive fraud detection that payment processors use to identify the highest-risk transactions to capture patterns of fraudulent activity, and to differentiate these patterns from legitimate purchasing activity. Geolocation tracking – Uses the users system information, location information, devices used by the user, etc., to track the customer’s activities and detects fraud.
- 8. Your Business, Our Understanding 9% 11% 16% 27%7% 10% 15% 26% 3% 5% 9% 19% 4% 7% 11% 15% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2011 2012 2013 2014 Market position of different products Deposits Debit card usage Mobile banking Credit Card • Well positioned in market - huge deposits and transactions – YOY increase rate is more than 20%. • Corporate account & Debit card section growth – Increased almost 70% in 2013-2014 financial year. • Credit card section lagging behind the other products of the bank. • Growth rate reversed for credit card section.
- 9. Your Business, Our Understanding
- 10. Your Business, Our Understanding 14 11 13 25 47 Customers complaint against Credit Card Intrest charges others Annual fees Customer relationship Security Issues • Major complaints - Customer Relationship, Interest Charges and Security Issues. • 47% fall under Security Issues. Almost 13K complaints booked during the year 2013- 2014. • Online frauds - a predominant issue with the customers. • No separate customer care section for credit card. • ICICI survey says 63% of people will switch to other credit card if they feel security issue.
- 11. Tango's Approach Transaction Based Scoring Real Time Decision making Adaptive Modelling Global Profiling
- 12. Tango's Approach Could fraud analytics do a better job of protecting not just Anil’s account, but his buying experience? Situation : Anil who is living in Kolkata, is on a vacation to J & K. He is buying a Snow Board from a high-end sporting store. • Bank A’s Credit Card declined the transaction (Traditional Detection System) Anil’s Feedback on this : “Being declined was annoying, but not enough to leave the bank. Still, if they keep bombarding me with irrelevant offers while declining the purchases I really do want to make, I’m going to think twice about jumping ship.” • Bank B’s Card approved the transaction (Innovative Streaming Analytics)
- 13. Tango's Approach – Techniques Incorporated •Behaviour Sorted Lists – Using Transaction Based Scoring, Behavioural Analytics capture the granularity of customer habitual patterns. •Collaborative Profiles - Using Global Profiling, we determine probability based on the behaviour patterns of other cardholders. with similar characteristics
- 14. Tango's Approach – How it works? • Fraud risk is lower if some aspects of the transaction matches the cardholder favourites. • Even with no favourites match, there is a way to anticipate new cardholder’s behaviour. • It maps actual individuals to multiple behavioural archetypes in real time and updates the mapping with each new transaction. • These archetypes are not conceptual, hand-built categories. Rather, they are machine- learned by an algorithm that analyzes historical transactional data which can come from multiple sources in any form.
- 15. Tango's Approach – How it works? >>How wide is the difference << Actual customer behaviour is mapped to archetypes in real time. The wider the change in archetype distribution, the riskier the transaction.
- 16. Tango's Approach – How it works? • As the quantity and variety of data being captured continues to expand, one of the most important things analytics can do for us is to help us understand change. • Knowing ‘what is changing’ and ‘whether it’s important’ is the key in creating vitality and value in customer relationships, while operating in a lean and efficient manner.
- 17. Support The Product includes one-year free Tango Care subscription. Tango Care package includes: One-Day Hands-on Product Training at Client’s Location. Dedicated Support Assistant with 24 x 7 Support. Access to Tango Ticket System for general queries. Product Quality Inspection every 3 months.
- 18. What’s Next? Product Consultant Assignment Data Collection POC Submission Final Discussion Contract Signoff Test Implementation Production Implementation Today 3-Weeks T T + 2 Weeks T + 3 Weeks
- 19. We sow, you reap !!! Fraud Detection to contribute 145 Cr. to the total revenue by 2022. inCrores
- 20. We sow, you reap !!! Merchant Acquisition is expected to be doubled (100% increase) upon Fraud Detection Implementation
- 21. Why Tango? Data Science Offer Services in Advanced Data Science & Quantitative Modelling Usage of Mainstream Tools- SAS, R, Python, SQL. Model requirement Tools: GAUSS, MATLAB. Data Visualization Tools: Tableau & Spotfire Experts in Econometric, Machine Learning and Operation Research Experienced Research Scientists & Patent Holders
- 22. Why Tango? Why do you need us? Data Science Offer Services in Advanced Data Science & Quantitative Modelling Experts in Econometric, Machine Learning and Operation Research
- 23. Business Instincts Delivering combination of data & art of business Timely delivery of information using custom reports via Dashboards, drill- downs, roll-ups and score cards Presentation of data using visualization and info-graphics Understanding the client and ensuring their needs and expectations are met Why Tango?
- 24. “Tango Technologies was very much like our internal team, working on several data science and analytics problems. They worked closely with me, building cutting-edge social media campaign management algorithms, and insights frameworks. This enhanced our offerings, helping us win some of our biggest clients in competitive trials against prominent names in the industry.” VP of Data Science and Products, Bank of America “We had an extremely productive association with Tango Technologies. They brought deep expertise in business analytics and data sciences, along with top notch professionalism. They blended seamlessly with our operations, engineering, and sales teams. They conceptualized and implemented advanced statistical frameworks on top of transaction data, helping us identify credit cards open to fraud. I would strongly recommend them for any data-oriented business problems.” VP Engineering, RBS Success Stories
- 25. “Tango Technologies were excellent as always. Partnership is a buzz word, but Tango Technology truly embody the concept of partnership, making sure that they understand their client stakeholders to tailor their service accordingly.” Global Analytics Manager (WealthCare Financial Group) “Tango Technologies team’s commitment, dedication, team work and follow through are big factors to ensure the success of project.” Senior Director - Planning and Insights (ICICI) “Tango Technologies generated many insightful slides on their own based on their analysis of our data, which were not part of the scope/deliverable. Such slides helped us to understand our product cross-sell and customer behaviour better and were thought provoking.” Business Intelligence Manager, Deutsche Bank Success Stories
Editor's Notes
- #15: this heterogeneous data is combined into a single mixed-bag “document,” which the algorithm examines as a whole to find similarities in customer behavior that support the discovery of archetypes. This document is unstructured—we can add data to it without having to know what predictive characteristics the various data sources contain or specify relationships between them. In other words, no data model is required. An advantage of this approach is flexibility to work with whatever transactional data becomes available. We also gain speed, since once the data has been assembled, analytic deployment is straightforward.