Credential stuffing is how a lot of loyalty account theft starts. An attacker takes email and password pairs leaked from one site and tries them on another. Where a member reused a password, the login works. In a loyalty program, what sits behind that login is a points balance that spends like cash.

What is credential stuffing?

OWASP defines credential stuffing as stolen username and password pairs fed into login forms by machine, to take over accounts (OWASP). It works because people reuse the same email and password on many sites. When one site is breached, each site where that pair still works is at risk too.

OWASP’s list of automated threats gives it the code OAT-008. The list sums it up as mass login attempts that check whether stolen pairs are valid. The pairs come from another breach or a criminal market, or from a public dump. Nothing is guessed, and that is what sets it apart from cracking a password (OWASP OAT-008).

New York’s Attorney General calls them repeated attempts by bots to log in with passwords stolen from other services. In 2022 the office found working logins for more than 1.1 million accounts at 17 well-known companies. Most of those attacks had gone unseen until the office flagged them (New York, 2022).

How does a credential stuffing attack work?

OWASP lays out the attack in four steps (OWASP):

  1. The attacker collects usernames and passwords from a breach or a phishing attack. Password dump sites supply more.
  2. Bots try the stolen pairs on many websites at once.
  3. Each login that works confirms a valid set of credentials.
  4. The attacker cashes in. OWASP lists draining stored value and making purchases first, then data theft, spam and the resale of working logins.

The tools are cheap and take little skill. The New York guide says one attacker can send hundreds of thousands, or even millions, of login attempts to one service. Most attempts fail. The sheer volume still yields thousands of open accounts (New York Attorney General guide).

Attackers also spread their traffic across many IP addresses through proxy networks. OWASP warns that this keeps the volume from each address low. Simple IP blocklists and rate limits then miss the attack (OWASP cheat sheet).

How is credential stuffing different from brute force and password spraying?

All three hit the login page, but each one tests something different. OWASP covers them together because the same defenses work against all three (OWASP cheat sheet).

AttackWhat the attacker testsThe weakness it usesThe first defense
Credential stuffingUsername and password pairs from another site’s breachThe same password reused across sitesMulti-factor authentication, plus checks of new passwords against breached lists
Brute forceMany passwords against one accountA weak or guessable passwordA cap on failed attempts per account. NIST sets the ceiling at 100 in a row
Password sprayingOne weak password against many accountsCommon passwords spread across a user baseA blocklist of common passwords when members set one
Account takeoverNothing new. It is the outcome of the attacks aboveValue that can leave the account fastAlerts and step-up checks before points or stored value move

The NIST row comes from SP 800-63B-4. It says a verifier shall limit consecutive failed attempts on one account to no more than 100 (NIST). OWASP calls credential stuffing one of the most common ways attackers take over user accounts (OWASP).

Why do attackers target loyalty and rewards accounts?

Loyalty accounts hold value that converts fast. Points turn into gift cards and flights. Stored value and saved cards often sit behind the same login. The New York guide lists how attackers cash out. They buy with a saved card or steal a saved gift card to sell. They also mine the account for phishing data or sell the login itself (New York Attorney General guide).

Dunkin’ is the case on record. From early 2015, attackers ran credential stuffing against Dunkin’ customer accounts, and tens of thousands fell within months. Many held DD stored value cards, which attackers spent or sold online. In 2020 Dunkin’ settled with the New York Attorney General and paid $650,000 in penalties and costs. It also had to reset passwords and refund stolen funds (New York Attorney General, 2020).

The pattern repeats. Chick-fil-A said attackers used logins from a third-party source to get into Chick-fil-A One accounts from June 17 to June 19, 2026. The breach hit 13,322 people. The chain logged out those accounts and removed saved payment methods. It restored each lost balance and added rewards to say sorry (BleepingComputer). It had reported an earlier wave that hit more than 71,000 customers in March 2023.

For illustration, an attacker tests 400,000 leaked pairs against a program’s login, and 1 in 200 works. That opens 2,000 accounts. At an average balance of 25,000 points worth one cent each, $500,000 of member value is exposed. If the program restores drained balances, as Dunkin’ and Chick-fil-A did, it pays for the same points twice.

How do you detect and stop credential stuffing?

OWASP calls multi-factor authentication, or MFA, by far the best defense. It cites Microsoft’s finding that MFA would have stopped 99.9% of account break-ins (OWASP). The FTC says a 2015 attack on the tax site TaxSlayer ended once the firm turned MFA on (FTC).

A second factor on every login adds friction that members feel. OWASP suggests asking for it only when a login looks risky, such as one from a new device or an odd location. It also suggests a step-up check before high-risk actions like large currency transactions (OWASP cheat sheet). In a loyalty program, a large redemption or a points transfer is that action.

The New York review found these safeguards worked best when set up right:

  • Bot detection. One restaurant chain told the office that its vendor blocked more than 271 million login attempts over 17 months (New York Attorney General guide).
  • Multi-factor authentication that is tested. The office found more than 140,000 accounts compromised at a business whose MFA had been set up wrong.
  • Passwordless login. With no password to reuse, a stolen pair is useless there (New York Attorney General, 2022).
  • Breached-password checks. NIST says verifiers shall compare each new password against a blocklist, which can include passwords from earlier breaches (NIST).

Detection matters because some attacks always get through. The New York guide points to spikes in failed logins and to patterns in complaints, such as stolen gift card balances. In a loyalty program, the clearest sign shows up at redemption. A login from a new device, then an email change, then a full-balance redemption within minutes reads as a takeover. Our guide to loyalty fraud prevention covers the full set of controls, and the entry on account takeover covers what happens next.

How GRAVTY handles credential stuffing

Credential stuffing starts at the login page. The loss lands later, when stolen points leave the account. GRAVTY watches that second step. AI-Trust, its fraud layer, uses machine-learning models to flag anomalous earn and redemption behavior as it happens. Its scoring models are tuned to each program’s own risk profile (GRAVTY for financial services).

Teams set rules, caps and velocity limits in GRAVTY Visual Rules, a patented no-code rules engine. A change needs no engineering release. The platform is ISO 27001 and SOC 2 Type II certified. Data is encrypted, and access is set by role at every layer.

Frequently asked questions

Who pays when loyalty accounts are drained by credential stuffing?

In the cases on record, the brand did. The New York Attorney General treats credential stuffing as a well-known attack that a business must guard against. Its Dunkin’ settlement required refunds for stolen funds plus $650,000 in penalties and costs (New York Attorney General, 2020). A program that restores drained points also carries the cost of the rewards the thief already took.

Does a CAPTCHA stop credential stuffing?

A CAPTCHA slows credential stuffing, but it does not stop it. OWASP notes that tools exist that break many CAPTCHAs at a high success rate (OWASP cheat sheet). The New York guide adds that CAPTCHA farms pay people to solve the rest (New York Attorney General guide). Use it on risky logins only, behind bot detection and MFA.

Can a member tell if their account was hit by credential stuffing?

Often only after the points are gone, unless the program tells them. OWASP advises showing the date, time and place of the last login. It also advises alerts when a login passes the password but fails the second factor (OWASP cheat sheet). Alerts on email changes and large redemptions turn members into an early warning system.

Is credential stuffing a data breach?

It can be, even when the program’s own database is never touched. New York sued Dunkin’ under its breach notice law for not telling customers about credential stuffing on their accounts (New York, 2020). The FTC has also faulted a firm for failing to control that risk (FTC on Ring).

What should a loyalty program do right after a credential stuffing attack?

Reset the passwords on hit accounts and find every account the attacker reached. That is the response New York set in the Dunkin’ deal, along with notice to customers. Dunkin’ also had to move balances to new stored value card accounts (New York, 2020). Then restore stolen points and review why the redemption went through.