What Is Anti-Passback? Stopping the Oldest Trick at the Door

Every counterfeit-detection technique ever taught assumes the document is fake. There is one very common method of getting underage people into a venue where the document is completely genuine, has never been altered, and will pass any inspection you can perform — including a barcode check.

Someone of age gets scanned and goes in. A minute later the card comes back out: through a railing, round the side of the building, out of the smoking area, or carried by a friend who walks back to the queue. It is handed to someone nineteen, who joins the line and presents a real licence to a door supervisor who has no way of knowing it was used four minutes ago.

This is passback, and it is probably the largest single hole in a well-run door. Not because it is clever — it is the least clever technique there is — but because every other layer of your process is looking at the card, and the card is fine.

What anti-passback actually does

Anti-passback is a simple rule: the same document cannot be used twice inside a set window.

When an ID is scanned, the system stores an identifier for that document — the licence number, or a hash of it — with a timestamp. Every subsequent scan is checked against what has already come through tonight. If the same document reappears within the window, the door gets an alert instead of a green light.

That is the whole mechanism. It is not sophisticated. What makes it valuable is that it catches a failure mode nothing else can see: the card is genuine, the barcode is correctly encoded, the front matches the back, the person in the photo is over 21. Every check you have passes. Only the repetition is wrong, and repetition is invisible to a human processing a hundred people an hour.

Nobody remembers licence numbers. Software does not have to.

Why the human cannot catch this

Four reasons, and they compound on exactly the nights you can least afford it.

Volume. A busy door processes well over a hundred people an hour. Recognising that a specific card has been seen before requires having retained it among hundreds, in the dark, while managing a queue. That is not a training problem. It is not something a person can do.

Lookalikes. Passbacks usually happen between people who resemble each other — a sibling, a friend of the same build and colouring. The photo check that catches a borrowed card handed to a stranger does not fire here, because the match is plausible.

Multiple entrances. The hand-off frequently goes around the corner. A card used at the main door reappears at the VIP line or the re-entry point, in front of a different member of staff who was not there for the first scan and has no way to know about it.

Conditions. Dark, loud, crowded, everyone in a hurry. These are the precise conditions the technique is designed around, and they are the conditions of every profitable night you will ever run.

The part that gets designed badly: re-entry

Here is where anti-passback goes wrong in practice, and why the setting matters more than the feature.

Your genuine customers leave and come back constantly. Smokers, phone calls, the cash machine, someone walking a friend to a taxi. On a normal night that is a large fraction of your patrons, and every one of them will present the same card twice.

A system that treats every second scan as a violation will fire dozens of false alerts a night. What happens next is entirely predictable: after the fifth false alarm, door staff stop reading the alerts and start clearing them reflexively. The feature is now worse than not having it, because it consumes attention and returns nothing — and the one real passback of the night gets waved through with the rest.

Any anti-passback implementation worth using has to distinguish an unauthorised passback from a legitimate re-entry. That is partly a software problem and largely an operational one:

• A configurable window, because venues differ. A neighbourhood pub with a busy smoking area might set fifteen minutes — under that, treat it as suspicious; over it, assume re-entry. A ticketed club night where nobody should be re-entering at all might set the window to the whole event.
• An alert, not a lock. The correct behaviour is to tell the door supervisor that this card came through at 23:14 and let a trained human decide. A hard block turns every smoker into an argument, and arguments at the door cost more than the risk they prevent.
• Somewhere to record the decision. If staff override an alert, that override should be logged with a reason. It keeps the record honest, and it tells you within a fortnight whether your window is set wrong — a venue overriding thirty alerts a night has a configuration problem, not a passback problem.

Set the window by watching your own door for a week. Guessing produces either noise or nothing.

What it does not solve

Worth being straight about, because this feature gets oversold.

It does not stop the first use. If the person the card genuinely belongs to never comes in, and it is handed straight to the minor at the front of the queue, there is no earlier scan to compare against. Anti-passback catches reuse; it cannot catch a borrowed card on its first presentation. Only a human looking properly at the face catches that.

It does not survive a determined perimeter problem. If your patio backs onto a public street with a low fence, cards will move over it faster than any policy can respond. Software will tell you it is happening — which is genuinely useful, and often the first time a venue discovers the fence is the problem — but the fix is the fence.

It is only as good as your scanning discipline. Anti-passback works by comparison against what has been scanned. Every unscanned entry is a gap in the record: the guest-list arrivals, the regulars waved through, the group rush at half eleven when the queue got long. Those exceptions are exactly where the technique is aimed.

Operational practices that matter more than the software

Close the perimeter. Walk your own site at 23:00 on a Saturday and look for where a card could be handed out: railings, fence gaps, the smoking area boundary, fire exits propped open for air, the gap beside the bins. Most venues find at least one. Fixing it costs less than any software.

Scan everyone. No exceptions for friends of the house, no skipping the back half of a group because the queue is building. An exception policy is a passback policy with extra steps.

Manage re-entry deliberately. If you have a smoking area, decide how re-entry works and make it the same every night — a wristband, a stamp, a dedicated re-entry point where staff know to expect repeat scans. Ad hoc re-entry is what makes the alert unreadable.

Train the response. When an alert fires, the door needs a calm, rehearsed script — a second look at the person, a question the real holder would answer easily, a decision. Staff who have not practised this either escalate a smoker into a confrontation or wave the alert away. Both are failures.

Read your own data. The pattern of alerts tells you things about your venue that nothing else will: which entrance, what time, how often. That is worth a manager’s ten minutes on a Monday.

Where we stand

Anti-passback is planned for the first release, for the reason described above: it catches something no amount of inspection can, and it needs no extra hardware — just the fact that the scans are already happening on connected devices.

The part we care about getting right is the re-entry distinction, because that is what determines whether door staff read the alerts or start ignoring them by midnight. If you run a door with a smoking area, or several entrances, the early access list below is where to tell us how re-entry actually works at your venue — that is precisely the detail that decides whether this feature is useful or noise.