Account Takeover Fraud: Protecting Customer Logins and Stored Cards
Account takeover fraud lets criminals spend customers' saved cards with orders that look legitimate. Learn the warning signs and login defenses that work.

Photo: Pixabay
On this page 7 sections
- How Account Takeover Fraud Differs From Stolen Card Fraud
- Credential Stuffing and Other Ways Accounts Get Hijacked
- Warning Signs of Stolen Customer Accounts
- Multi-Factor Authentication and Login Alerts That Customers Accept
- Protecting Stored Cards and Loyalty Balances
- Handling Disputes From Hijacked Accounts
- Where to Start This Week
Key takeaways
- Account takeover orders pass address and history checks because they use the real customer's stored data, so "trusted repeat customer" rules can work against you.
- The most reliable warning sign is sequence: account changes from an unfamiliar device, followed quickly by a purchase.
- Add friction only at risky moments (new device, new address, contact-detail changes, points redemptions) instead of on every login.
- Hijacked-account disputes are real fraud from the cardholder's side and hard to win, so prevention and fast refunds usually beat fighting.
Account takeover fraud happens when a criminal logs in to a real customer's account and spends what's saved there: the stored card, the address book, the loyalty points, the gift card balance. The orders look legitimate because the account has history and the saved card has been charged before without trouble. That is why account takeover (ATO) slips past filters built to catch stolen card numbers, and why your defense has to start at the login screen instead of at checkout.
The damage lands twice. You absorb the chargeback and the lost merchandise. Your real customer, who did nothing wrong, finds a charge they didn't make on a site they trusted.
How Account Takeover Fraud Differs From Stolen Card Fraud
With ordinary stolen card fraud, a criminal has a card number and tries it on your checkout, usually as a guest or through a brand-new account. Fraud screening catches a good share of this because the details don't line up. The billing address fails the address check, the email was created yesterday, the shipping address sits far from the billing address, and the buyer has never ordered from you. If many of your legitimate orders come from shoppers abroad, see what dLocal's Q3 results mean for selling abroad.
ATO flips almost every one of those signals. The account may carry years of clean order history. Because the saved card belongs to the customer, the address check passes. Most stores don't ask for the security code again on a saved card, so that check never runs. To a scoring model, the order looks like it came from your best kind of customer.
Many stores make it worse with rules that ease up on repeat buyers, such as skipping manual review once a customer has a few past orders or letting "trusted" accounts bypass velocity limits. Those rules make sense against stolen card fraud. They also hand anyone who breaks into a good account a lighter screening path. What's left to catch ATO is behavior: an unfamiliar device, or an account change made right before money moves.
ATO fraud in ecommerce hits some business models harder than others. Subscription and membership businesses see attackers redirect a recurring box to a new address or buy gift subscriptions on the saved card. Online marketplaces face it on both sides, since buyer accounts hold cards and seller accounts hold payouts, and a hijacked seller account with a new bank account on file can drain earnings before anyone notices.
Digital products are the most exposed of all. Game credits, gift cards, software keys and downloads deliver instantly and resell easily, so there's no shipping window in which to catch the problem. And any store with a loyalty program or stored credit carries a balance that can be spent or transferred without touching a card, which means no chargeback arrives to warn you.
Credential Stuffing and Other Ways Accounts Get Hijacked
The most common entry point is credential stuffing. Criminals buy lists of email and password pairs leaked from breaches at other companies, then use bots to try those pairs on your login page. It works because people reuse passwords. Each attempt has a low chance of success, but the bots run them by the thousands, so a stuffing run looks like a flood of failed logins with a few successes buried inside. It is a cousin of card testing attacks, which hit your checkout with stolen card numbers instead of your login form with stolen passwords, and the tooling and defenses overlap.
Other routes matter because each one gets around a different defense. Phishing emails dressed up as "a problem with your order" copy your branding and lead to a lookalike login page that captures the password. If the attacker controls the customer's inbox outright, your password reset link goes straight to them, and so do any email codes. SIM swaps target text-message codes: the attacker talks a mobile carrier into moving the customer's number to a new SIM and starts receiving them.
Session theft is the sneakiest. Malware on the customer's computer copies the browser cookies that keep them logged in, so the attacker never sees a login page and neither passwords nor one-time codes stand in the way. Then there's your own support team. "I lost access to my old email, can you update it?" is one of the oldest moves in the book, and AI-generated voices make it more convincing. We covered that trend in how deepfakes are fueling account takeovers.
Warning Signs of Stolen Customer Accounts
Look in two places: site-wide traffic for attacks in progress, and individual accounts for takeovers that already worked.
Site-wide, a stuffing attack usually shows up as a spike in failed logins well above your normal baseline. Many of those attempts target email addresses that have no account with you, a giveaway that someone is working through a purchased list. The traffic often comes from data-center IP ranges, or from many different IPs sharing the same device fingerprint. Timing helps too, since a burst in the middle of the night for your customers' time zone deserves a look.
On a single account, the strongest signal is sequence. One odd event means little; several in a row, minutes apart, is the pattern that matters. Watch for these combinations:
- A login from a new device or country, followed quickly by a change to the email or password.
- A new shipping address added and used on the very next order, especially a freight forwarder or reshipping service.
- Purchases that break from the customer's history, such as gift cards, digital codes, high-resale items or expedited shipping.
- A loyalty balance redeemed or transferred all at once, or a dormant saved card suddenly put to use.
- Several different accounts accessed from the same device in one session.
If your platform or gateway records device fingerprints and account changes, keep those records. You'll need them for detection now and for disputes later.
Multi-Factor Authentication and Login Alerts That Customers Accept
You can prevent account takeover without forcing every shopper through a code on every visit. The approach that keeps conversion intact is risk-based step-up: no extra friction for a customer on a recognized device, and a second factor at the moments that matter.
Second factors vary in strength. Passkeys are the strongest because they resist phishing and don't depend on a phone number. Authenticator apps avoid SIM-swap risk, but a lookalike login page can still capture their codes and replay them in real time. Text-message and email codes are weaker, since SIM swaps and compromised inboxes defeat them. Even so, any second factor raises the cost of an attack well above a password alone. Offer passkeys and let customers opt in, and fall back to text or email for customers who won't use anything else.
Here is one way to map events to responses. Treat it as a starting point and tune it to your margins and your customers' tolerance:
| Event | Suggested response |
|---|---|
| Login from a recognized device | No extra step |
| Login from a new device or country | One-time code or passkey prompt |
| Email or password change | Re-authenticate; alert the previous contact details |
| Saved-card order to a new shipping address | Ask for the card's security code or run 3-D Secure |
| Loyalty redemption soon after an account change | Hold for a set period or require a second factor |
| Seller payout account change (marketplaces) | Second factor plus a hold on the next payout |
Login and change alerts cost little and catch a lot. Send them to the contact details on file before the change, so the real customer hears about it even if the attacker swapped the email. Include a "this wasn't me" link that locks the account and signs out every session.
Quieter controls do much of the work behind the scenes. Rate-limit login attempts by IP address and by account, and add bot detection that challenges only suspicious traffic so ordinary shoppers never see a puzzle. At signup and reset, reject passwords that appear in known breach lists. Return the same error ("email or password is incorrect") whether or not the account exists, so bots can't confirm which emails are registered. When a password changes, end all other sessions, which shuts out anyone holding stolen cookies.
Protecting Stored Cards and Loyalty Balances
A common misunderstanding is that tokenization protects a saved card from takeover. It doesn't. Network tokens replace the card number with a token tied to your business, which limits the harm if your database leaks and keeps cards current when they're reissued. But the token works for whoever is logged in. Once an attacker is inside the account, it charges just as smoothly for them as for the customer.
So protect the moment of use. Re-ask for the security code when a saved card ships to a new address or is used from a new device. The real customer has the card in their wallet; the attacker usually has only the account.
For higher-risk orders, 3-D Secure authentication goes further. The challenge comes from the card issuer, often through the customer's banking app, which an account thief typically can't open. Card network rules can also shift fraud chargeback liability from you to the issuer on authenticated transactions, though conditions apply, so confirm the details for your account with your processor. Our explainer on 3-D Secure 2 and fewer fraud chargebacks covers when to trigger it selectively instead of on every order.
The account page itself should never become a card viewer. Show the card brand and last four digits, and don't let a logged-in user see or export a full card number. A cooling-off period on new shipping addresses helps as well: a saved card can't ship to a freshly added address until a set time passes, unless the customer clears a step-up check.
Loyalty points and store credit deserve the same care as cash, and arguably more. Many points systems have no chargeback path, so a drained balance becomes a direct loss you'll probably reimburse to keep the customer. Require a second factor to redeem or transfer points, and hold redemptions that follow a recent account change. Subscription businesses can apply the same logic by confirming delivery-address and plan changes to the original email before the next cycle ships. Sellers of digital goods have the least time to react, so a short fulfillment delay on orders that follow a new-device login buys the window that physical goods get for free.
Handling Disputes From Hijacked Accounts
When a customer says "I didn't place this order," act on the account first. Lock it and end every session. Force a password reset, cancel anything unshipped, and refund the charges you can confirm came from the attacker. Then decide whether you're looking at ATO or friendly fraud, where the cardholder made the purchase and disputes it anyway. Your logs settle it. An order from the customer's usual device and address points to friendly fraud, while a new device paired with a fresh email change points to takeover.
That distinction tells you whether to fight. Dispute evidence for card-not-present orders often leans on the customer's history, such as the same device or IP address seen on earlier undisputed orders. In a true takeover, that evidence works against you, because the fraudulent order didn't come from the customer's device. Unless the transaction was authenticated with 3-D Secure, these disputes are hard to win, and accepting them can cost less than the staff time spent fighting.
Fraud disputes also count toward network monitoring, so a wave of takeovers can push you toward the levels described in chargeback ratio thresholds and how close is too close. Refunding before a chargeback is filed usually avoids the dispute itself, but the issuer may still record the fraud report. Ask your processor how refunds interact with the fraud monitoring that applies to your account. If you sell through a platform, see what a merchant of record is and who carries chargeback liability.
Price One Hijacked Order Before You Choose Your Friction
Before deciding how much friction to add, put a real cost on a single takeover order using your own figures. Start with the wholesale cost of the goods that left your warehouse, since that's what you actually lose, not the retail price. Add the outbound shipping you paid. Add the chargeback fee in your processing agreement, which you'll find on your statement or in your fee schedule. Then add the processing fee on the original sale, because many processors don't return it when the sale is reversed; your agreement will say. If you process with Stripe, see what Stripe processing fees really cost beyond 2.9% + 30¢.
Set that total next to your profit on an ordinary order. If your net margin is thin, one hijacked order can wipe out the profit from many good ones, because you lose the cost of goods and the fees rather than just the margin. Then weigh what the fix costs you: a security-code prompt on a new shipping address takes the real customer a few seconds, and only on the orders that trip it. When the loss on one bad order is large relative to your margin, step-up checks on risky orders almost always earn their keep. When it's small, as with low-cost digital items you can revoke, lighter controls and quick refunds may be the better trade.
Where to Start This Week
Pull the past month of account activity and count one thing: logins that led to an email, password or shipping-address change followed shortly by an order. That number is your current exposure, and it shows where step-up checks will pay for themselves first. Then ask your gateway provider whether it can re-prompt for the security code on saved cards and trigger 3-D Secure only on risky orders.
If your current online payment gateway can't apply those checks selectively, request a free quote and ask a MannyPay specialist which options fit your platform. Quote requests get a response within one business day.
This article is for general educational purposes and is not legal, tax or financial advice. Card network rules and state laws change; confirm current requirements for your business.


