Stripe Fraud Prevention with Radar Rules

Fraudulent transactions cost more than the lost funds. Each chargeback incurs a $20 Stripe fee regardless of outcome. High dispute rates trigger account reviews or termination. Proactive fraud rules protect both revenue and account standing.

The Fraud Problem

Stolen credit cards get tested on digital products. The cardholder eventually notices, disputes the charge, and you refund the full amount plus a $20 fee. Scale this across dozens of fraudulent transactions and the financial damage becomes severe.

Worse, Stripe monitors dispute rates. Accounts with elevated fraud activity face restrictions or closure. The account itself becomes at risk, not just the individual transactions.

Enabling Radar Rules

Upgrade to Stripe Radar for Fraud Teams to access custom rule creation. The standard Radar catches obvious fraud but allows configurable rules for nuanced situations.

Custom rules combine conditions like card country, risk score, and transaction history. When conditions match, Stripe blocks the transaction before it processes.

Effective Rule Configurations

Block high-risk countries above risk thresholds. Cards from Philippines with risk scores above 30 warrant automatic blocking. Cards from Singapore above 40 similarly. These thresholds catch fraudulent patterns while allowing legitimate customers.

Block cards with dispute history. Any card that has previously disputed a charge on your account should be blocked from future purchases. The rule :dispute_count_on_card_number_all_time: > 0 implements this.

Block customers using many cards. Legitimate customers rarely cycle through multiple payment methods. A customer associating more than two cards suggests card testing. Block these with :card_count_for_customer_all_time: > 2.

3D Secure for Borderline Cases

Rather than blocking all medium-risk transactions, require 3D Secure verification. Cards that can’t complete verification fail automatically. Legitimate cardholders complete the extra step; fraudsters typically can’t.

Configure rules to request 3DS when risk scores fall in ambiguous ranges. This preserves legitimate sales while filtering fraud that passes basic checks.

Global High-Risk Threshold

Set a blanket rule blocking any transaction with risk score above 50. This catches fraud that doesn’t match specific patterns but still exhibits suspicious characteristics. The threshold can be adjusted based on your observed fraud rates.