Skip to content

Research and Journey

Nilson Report

Nilson Report is the most trusted source of news and analysis for the card and mobile payment industry.
The latest Nilson Report from 2019 reveals that global card fraud losses reached $28.65 billion. Card fraud losses are projected to reach $35.31 billion by 2025, highlighting the urgent need for advanced fraud detection systems.

nilson_report
Nilson Report

Online payment fraud is a growing concern for businesses and consumers alike. As e-commerce continues to expand, fraudsters are finding new ways to exploit vulnerabilities in online transactions.

European Central Bank

European Central Bank (ECB) data shows that card fraud losses in the Single Euro Payments Area (SEPA) reached €1.55 billion in 2019.
The ECB has been working to enhance payment security through the implementation of the Revised Payment Services Directive (PSD2).

european_bank
European Central Bank Report

The PSD2 aims to improve the security of online payments and protect consumers from fraud. By requiring strong customer authentication for electronic transactions, the directive seeks to reduce the risk of fraud and enhance the overall security of the payment ecosystem.

Well Known PIN Codes | Advanced Techniques (Biometric identification)

Purpose

  1. Automated systems should optimize the workload of fraud investigators.

  2. Automated systems and human investigators work at different time scales:

    • Automated systems:

      • Usually provide risk scores for transactions in less than a second.

    • Fraud investigators:

      • Usually require contacting a client to confirm a fraud, which can take days or weeks.

Scenarios

flowchart TD
    A[Scenarios] --> B[Card Present]
    B --> C[Lost or Stolen Card]
    B --> D[Card Not Received]
    B --> E[Counterfeited Card / Skimming]
    A --> G[Card Not Present]
    G --> H[Physical Card Not Present]
    G --> I[Payment performed on Internet]
    I --> J[Mail]
    I --> K[Mobile]
    I --> L[3rd Party Apps]

CP Scenarios have existed for more than 2 decades and are robust to fraud attacks, notably due to EMV (Europay Mastercard & Visa) technology.

Layer Diagram

flowchart TD
    A[Layer Diagram] --> B[Terminal Blocking Rules]
    B --> C[Within Milli-seconds]
    A --> D[Transaction Blocking Rules]
    D --> E[Before Authorization]
    A --> F[Scoring Rules]
    F --> G[Near Real time]
    A --> H[Data Driven Model]
    H --> I[To Block the Card]
    A --> J[Investigators / Human Intervention]
    J --> K[In Offline Mode]
layer_control_diagram
Layer Control Diagram

Investigators design transaction-blocking & scoring rules layers.

  1. Check few alerts/day as the process is long & tedious.
  2. Investigate the wrong/false alarms raised by DDM (Data Driven Model) and reported back to real-time FDS (Fraud Detection System).
  3. High scores of DDM for fraud will pass the Investigators layers, and contact/send alerts using SMS.

Variation in Transaction Data

graph LR
    A[Transaction Data] --> B[Account Features]
    B --> C[Account Number]
    B --> D[Date of Account Opening]
    B --> E[Account Limit]
    B --> F[Card Limit]
    B --> G[Card Expiry Date]

    A --> H[Transaction Features]
    H --> I[T/X Refrence No.]
    H --> J[Account No.]
    H --> K[T/X Amount]
    H --> L[POS No.]
    H --> M[T/X Time]
    H --> N[Location]

    A --> O[Customer Features]
    O --> P[Customer ID No.]
    O --> Q[Type of Customer]
    Q --> R[Low Profile]
    Q --> S[High Profile]
    O --> T[Mobile No.]
    O --> U[Address]

    A --> V[Generate Dataset]
    V --> W[Highly Imbalanced]
    V --> X[Contain both numerical & Categorical values]
    V --> Y[Feature time dependent fraud scenarios]

    A --> Z[Transaction Process]
    Z --> AA[Difference in spending habits]
    Z --> AB[Their geographical location]
    Z --> AC[Spending frequency]
    Z --> AD[Spending Amounts]
    Z --> AE[Spending Date & Time]