If you've ever looked at a chargeback case and thought,
How did this even get approved?
You're not alone. Across merchant forums and private chats, one question keeps coming up: Why do fraud tools miss things that feel so obvious?
We're talking about stuff that real people spot in seconds. Like:
- Customers who ask about refunds before they even buy.
- Orders with mismatched names, emails, and addresses.
- Buyers who place one order cancel it and reorder five minutes later with a different card.
This isn't subtle fraud. It's the kind that leaves a paper trail a mile long. So why aren't automated systems catching it?
Let's break down the common patterns merchants keep flagging — and why your fraud tool probably isn't built to notice them.
1. Refund Questions Before Purchase
A lot of merchants say this is their number one red flag. Someone shows up in live chat or email asking:
- "What's your refund policy?"
- "Can I cancel after ordering?"
- "How quickly can I get my money back?"
Sometimes the message comes through before any order is placed. Sometimes it's seconds after checkout. And in too many cases, a chargeback hits within a week.
Fraud tools don't track pre-sale behavior like this. Most are wired into transaction data, not chat logs or help desk tools. So, unless your system integrates everything, these buyers slip through with clean scores.
What to watch for: If you're seeing refund-related questions before purchases, especially from new customers, flag those accounts manually. Fraud tools won't connect the dots for you.
2. Card Switching and Back-to-Back Declines
Another overlooked pattern is card switching. Here's what it looks like:
- A customer tries to pay.
- The card is declined.
- They try again — maybe with two or three different cards.
- One finally works, and the order goes through.
This happens fast. Sometimes in under a minute. Most tools don't flag this as risky unless the IP address changes or the cards come from different countries. But merchants say this pattern often leads to chargebacks, especially when the names on the cards don't match the account holder.
What's going wrong?
Fraud tools often judge each transaction separately. They don't always treat multiple declines from the same device or user session as suspicious behavior.
Fix tip: If you see a customer go through multiple failed payments in a row, consider adding a rule that pauses the transaction until it can be reviewed. Some merchants use velocity filters or custom scripts to catch this.
3. Erratic or Partial Address Info
Here's another classic: the customer gives an address with missing or weird formatting.
- Apartment number swapped with the street number
- ZIP code doesn't match the city
- Entire address in lowercase or ALL CAPS
- Billing and shipping info barely related
Some of these might look like typos. Others are signs of masking identity. Fraudsters often mess with address info just enough to avoid AVS (Address Verification System) rejection, especially when testing stolen cards.
The catch: AVS checks whether parts of the address match, not the full context. So even if a ZIP code and street number line up, the tool might approve it, even if the rest is garbage.
Pro move: Train your customer support team to eyeball addresses that feel off. A human can often spot what AVS doesn't.
4. Inconsistent Account Details
Merchants report another big one: mismatched account details that just don't make sense together.
- Email: [email protected]
- Name on order: Michael Jordan
- Phone number: random 7-digit local number
- IP geolocation: Vietnam
- Shipping address: New Jersey
This stuff should raise red flags, but fraud tools might not flag it unless your scoring system weights mismatches heavily. Most tools are designed to catch high-risk behavior across categories, but they may not flag inconsistencies unless you customize the risk logic.
What helps: Use multi-point verification. Match email domain, billing name, and phone number patterns. Consider adding manual reviews for orders with three or more inconsistent data points.
5. Friendly Fraud Patterns
This one's tricky. Some customers aren't professional fraudsters. But they behave like one when things don't go their way.
Examples include:
- Claiming they didn't receive the item, even though tracking shows delivery
- Filing disputes after the return window closes
- Saying their child used the card without permission
Fraud tools struggle with this because friendly fraud usually comes from accounts that look clean. The purchase looks normal. Shipping is confirmed. But something shifts after the sale.
Best protection: Document everything. Emails, delivery confirmations, refund attempts. The more context you have, the better your chances in a dispute especially with repeat offenders.
Real Talk from Merchants
One seller shared this:
"We had a customer ask three times about our refund policy, placed a $400 order, then disputed the charge two days later. The name and email didn't match, and the address was off by one number. But our tool said it was low risk. That's when we realized the tool isn't reading intent. It's just reading data."
Another said:
"We added a custom rule that flags if someone uses three cards in one session. Our fraud dropped by 30%. It was a simple fix, but our system never offered it by default."
Final Takeaway: Tools Aren't Enough Without Eyes On
Fraud detection systems are useful, but they're not built to think like a human. They miss context. They don't read tone. And they're not watching customer behavior before the checkout page.
If you're seeing sketchy patterns get through, it might be time to tighten your manual review process, rework your fraud logic, or combine your tools with human checks.
The bots can help, but they can't do it all.
Stay One Step Ahead with Chargeblast
If your fraud tools are missing the obvious, Chargeblast can help you build smarter alerts around the stuff machines don't catch. We plug into your existing system and surface the real signs of trouble, from card switching to refund-baiting patterns, so you're not stuck reviewing every order yourself. Get the upper hand before disputes hit.
