Home › What a 3-Strike Login Lockout Does to Support Tickets

What a 3-Strike Login Lockout Does to Support Tickets

What a 3-Strike Login Lockout Does to Support Tickets

A three-attempt login lockout cuts credential-related support contacts by roughly 60% in the first month of deployment, but the residual ticket mix shifts toward a harder, more expensive category: account recovery. The policy solves the problem it is designed to solve and creates a second one in its place. What follows is an examination of that trade, drawing on the operational logic of lockout design rather than on any single operator's data.

The Mechanism and the First-Order Effect

A three-strike lockout is a rate limiter applied to authentication. After three consecutive failed credential submissions, the account is frozen for a defined cool-down — commonly 15 to 30 minutes, sometimes until manual release. The design intent is narrow: defeat online guessing and credential-stuffing scripts by making the cost of each guess rise faster than the attacker's willingness to pay it.

The first-order effect on support volume is well established in the fraud literature and consistent across operators who have published internally. Password-reset and "I can't get in" tickets are the single largest category in most casino support queues, typically 25–40% of inbound contacts. A lockout intercepts a large share of these before they are ever filed, because the user who mistypes twice and succeeds on the third attempt never contacts anyone. The 60% figure above is a reasonable planning assumption for the credential-adjacent slice of the queue, not for total volume.

The mechanism is also cheap. A lockout is a few lines of state logic and a timer. It requires no vendor, no model, and no per-contact cost. This is why it propagates: it is the rare control that reduces both fraud loss and headcount pressure simultaneously.

Why three and not five

The threshold is a policy choice dressed as a technical one. Five attempts reduces false positives — genuine users with a stale password manager entry, a keyboard layout mismatch, or a shared device — at the cost of doubling the attacker's free guesses. Three is the common compromise because the marginal attacker gain from attempts four and five is small against a well-chosen password, while the marginal support cost of a lockout is high. Operators who move to three typically do so after a credential-stuffing incident, not after a support analysis.

The Second-Order Effect: Recovery Becomes the Queue

The tickets that survive a lockout are not a random sample of the tickets that existed before it. They are selected for difficulty.

A pre-lockout "can't log in" ticket is often resolved in one interaction: the agent confirms identity, issues a reset link, closes. A post-lockout ticket is frequently a user who has already failed three times, does not know why, and may be locked out of the email address tied to the account as well. The agent cannot simply reset — the account is in a frozen state that may require a supervisor override, a KYC re-check, or a wait for the cool-down to expire while the user remains on the line.

The measurable consequence is handle time. Recovery tickets routinely run two to four times the duration of a standard reset, and a meaningful fraction escalate to identity verification, which pulls in a second team. Total ticket count falls; total agent minutes fall less. An operator that staffs to ticket volume rather than to ticket minutes will under-provision exactly when the lockout goes live.

There is a further selection effect worth naming. Sophisticated attackers do not trigger lockouts on the accounts they intend to take; they trigger them on decoy accounts to probe response, or they spread attempts across many accounts at one or two tries each. The lockout therefore filters out the casual and the clumsy — the same population that generated the cheap tickets — while leaving the determined attacker and the genuinely locked-out legitimate user in the queue together. The agent cannot easily tell them apart, and the policy has removed the signal (repeated failures) that previously helped.

The Lockout-to-Fraud Pipeline

Not every lockout is a support event, but every lockout is a data point. A user locked out three times in an hour is a different risk profile from a user locked out once a quarter, and the distinction is only visible if the lockout counter is joined to the fraud system rather than siloed in authentication.

Operators who treat the lockout as a signal, not just a gate, report a secondary benefit: the locked-out population becomes a high-precision feed for manual review. The cost is that this feed generates its own tickets — the "why is my account under review" contact — which lands in the same queue and is even harder to resolve because the agent is often not cleared to discuss the review.

The honest accounting

A lockout's support impact is best expressed as a net figure across three lines: credential tickets avoided (large, fast to realize), recovery tickets made more expensive (moderate, realized within weeks), and review-related contacts created (small but persistent). The net is usually positive. It is rarely as positive as the first-month number suggests, because the avoided tickets were the cheapest ones in the queue and the new ones are among the most expensive.

Design Choices That Move the Number

The lockout threshold gets the attention; the surrounding parameters do most of the work on support load.

Cool-down length is the dominant lever. A 15-minute lockout generates a support contact only if the user cannot or will not wait. A 24-hour lockout converts a wait into a ticket almost every time. Short cool-downs preserve most of the anti-guessing benefit — the attacker's cost is in the delay, not the finality — while keeping the "I'll just try again later" user out of the queue entirely.

Counter reset semantics matter more than they appear to. If the counter resets on any successful login, a user with one bad password and two good ones never locks out. If the counter resets only on a time window, the same user accumulates strikes across sessions and locks out on a day they did nothing wrong. The second design is stricter and generates more recovery tickets for no measurable fraud benefit.

Notification is the cheapest intervention available. An email or in-app message at the moment of lockout — "your account is locked for 15 minutes; if this wasn't you, here is what to do" — converts a would-be ticket into a self-service resolution for the majority of legitimate users. Operators who add it typically see the recovery-ticket share flatten within a quarter.

Channel asymmetry is underused. Locking the web login while leaving a verified mobile session active means the user with a remembered device never notices. The lockout still defeats the remote attacker, who has no such session. This is a policy decision, not a technical one, and it is frequently overlooked because authentication teams and support teams do not share a dashboard.

Where the Evidence Is Thin

The 60% figure is a planning heuristic, not a measurement, and the operators best positioned to measure it rarely publish. What is missing is a controlled comparison: the same operator, the same population, lockout on versus off, with ticket categories, handle times, and fraud loss tracked together. Without that, the support impact of a three-strike policy is inferred from adjacent evidence — the composition of queues before and after, the handle-time distribution of recovery contacts — rather than demonstrated.

The open question is whether the lockout is the right instrument at all. Its support cost falls on the legitimate user who forgot a password, while its fraud benefit accrues against an attacker who, at three attempts per account, was never going to succeed against a decent password anyway. Passkeys and device-bound sessions remove the password from the equation entirely and would collapse both the credential queue and the recovery queue. Until those are ubiquitous, the three-strike lockout will keep doing what it does: clearing the cheap tickets and handing support a smaller, slower, more expensive pile in their place. Whether that trade is worth it depends on a number most operators have not yet bothered to calculate — the fully loaded cost of a recovery contact against the fraud loss it prevents.