Stripe Radar Rules
Custom fraud rules for Conservative Dating subscription checkout
How to apply these rules
- Open Stripe Dashboard Stripe Dashboard → Radar → Rules
- Click Add rule and paste each rule string below
- Add the rules in order — Block rules first, then Review rules
- Enable each rule after saving
- Radar rules apply to all new payment attempts immediately
Radar rules cannot be configured via the Stripe API — they must be entered in the Dashboard. The checkout session handler already sends enriched metadata (member_id, fingerprint_visitor_id, client_ip, user_agent) to Stripe for use in these rules.
Metadata sent with every checkout session
:metadata:platform:"conservative_dating":metadata:member_id:BetterAuth user ID:metadata:member_email:Authenticated email:metadata:fingerprint_visitor_id:FingerprintJS Pro visitor ID:metadata:client_ip:Real client IP (X-Forwarded-For):metadata:user_agent:Browser user-agent string:metadata:checkout_at:ISO 8601 timestampBlock Rules
8 rulesStripe Radar assigns a 0–100 risk score using its global fraud network. Scores ≥ 75 have a very high fraud probability. Dating platforms are a common target for card testing.
Prepaid cards are frequently used for fraud and chargebacks. Legitimate dating subscribers almost always use credit or debit cards tied to a real bank account.
Conservative Dating is a US-focused platform. Cards issued in these regions have disproportionately high fraud and romance-scam rates. Adjust this list based on your actual customer geography.
A card issued in one country used from an IP in a different country, combined with an elevated risk score, is a strong fraud signal. Legitimate US members traveling abroad are unlikely to trigger this (their card country is US).
More than 2 distinct cards tried on the same email address in a rolling window indicates card testing or stolen card cycling. Legitimate subscribers rarely try more than 2 cards.
Multiple cards from the same IP address is a card-testing signal. Three is a generous threshold that avoids blocking shared NAT/VPN IPs while catching automated attacks.
A failed CVC check means the cardholder entered the wrong security code — a strong indicator of a stolen card number. There is no legitimate reason to proceed after a CVC failure.
Failed AVS ZIP check means the billing address does not match the card issuer records. Combined with a dating subscription, this is a high-risk signal.
Review Rules
4 rulesScores in the 50–74 range are elevated but not conclusive. Flagging them for review lets you inspect before charging rather than blocking legitimate customers.
Every checkout session should have a member_id in metadata (set by the server). A missing ID means the checkout was initiated without a valid session — suspicious.
Disposable email addresses are a strong fraud signal on subscription platforms. Legitimate members use real email addresses they can receive notifications on.
A second card attempt on the same email warrants a closer look, even if not an outright block. Could be a legitimate retry after a declined card, or early-stage card cycling.
Ongoing tuning
- Review blocked payments weekly in Stripe Dashboard → Radar → Reviews to catch false positives
- Adjust the high-fraud country list based on your actual member geography — remove countries where you have legitimate subscribers
- Lower the block threshold from 75 to 65 if chargeback rates remain elevated after 30 days
- Enable Stripe Radar for Fraud Teams (paid add-on) for machine-learning-enhanced rules and team review queues
- Cross-reference FingerprintJS banned_fingerprints with Stripe Radar reviews — a banned device attempting checkout is a strong signal