How Protectmaxx works

Online payment fraud #

What is online payment fraud? #

Online payment fraud is when a fraudster uses false credentials to make purchases online. Fraudsters can obtain these credentials in several ways, which include account takeover (ATO), purchasing credit card details on the dark web, or gathering information via phishing.

Through online payment fraud, fraudsters can gain access to products at no cost for their personal use, or they can generate income by reselling these goods.

How Alphacomm prevents online payment fraud #

To prevent online payment fraud, we utilise the data sent to us via our API. Our anti-fraud solution evaluates whether the order appears fraudulent and provides a recommendation on whether you should accept it. Alphacomm solely offers guidance on accepting or declining, thereby clarifying how the order should be managed.

Alphacomm uses a unique combination of machine learning, rules and human supervision to provide advice:

🛡️ The Protectmaxx Three-Pillar Defense System

At Alphacomm, we combine sophisticated technology with dedicated human expertise to create a layered defense against online payment fraud. Our strategy is built on Rules, Machine Learning, and our Expert Analyst Team.

1. Rules: Dynamic Business Logic

Rules serve as an effective, agile mechanism to enforce critical business logic and rapidly respond to new fraud patterns and emerging threats. They provide immediate, granular control that complements our machine learning models.

Rules enable you to:

• Act Fast to Stop Attacks: Fraud analysts can deploy or modify rules instantly to halt a fraud attack while it is actively ongoing.
• Proactively Block Trends: Safeguard against emerging fraud trends before the machine learning model fully adjusts to the new fraudulent behavior.
• Allow Good Customers (Frictionless Flow): Rules can be used not just to prevent, but to create “Allow” lists, minimizing friction for high-value, trusted customers.

Alphacomm diligently performs around-the-clock surveillance (24/7, 365 days a year) on rule performance, proactively offering improvements whenever necessary.

Machine Learning: Adaptable & Scalable Defense

Alphacomm employs advanced machine learning models, custom-tailored to fit your specific transactional data, to deter online payment fraud.

Key Advantages:

• Robust & Hard to Bypass: Our models offer a highly robust defense due to their adaptability and scalability.
• Continuous Optimization: We are proactive in monitoring and improving model performance. This includes:
• Consistent retraining of the models.
• Threshold adjustments based on real-world outcomes.
• Identification and integration of significant new data features.

3. Analyst Team: 24/7 Human Expertise

Our dedicated team of expert analysts works tirelessly, ensuring maximum security and conversion rate optimization.

Their Role:

• 24/7 Transaction Monitoring: Provides continuous surveillance for rapid fraud detection.
• Swift Response: Quickly responds to sudden changes in attack vectors and potential high-impact incidents.
• Conversion Rate Focus: This human oversight ensures we consistently deliver the highest possible conversion rate all year round.

 

📊 Comprehensive Data Used for Fraud Models

Our online payment fraud models analyze hundreds of customer features by pulling data from multiple, integrated sources:

 

Data Type	Example Features Collected
Merchant Data	Transaction details (what was purchased), purchase timing, payment method chosen, and purchase location.
Payment Information	Data provided by Paymaxx (or other PSPs) on payment status, method, and specific customer-related data regarding the transaction.
Internal Data	Protectmaxx’s historical data, including previous transactions, chargebacks, and profile construction.
Third-Party Data	Data used to enrich customer profiles, such as email validation status and Device ID.
Customer-Specific Data	Examination of factors like the customer’s email domain used and the age of the customer’s account.
Location Information	Analysis of the difference between the customer’s billing address and their delivery address.
Device Information	Evaluation of whether the device has been used for a previous order, and if it has been associated with other customers.