The Queue at the Terminal
Why badge-passing is usually a queue problem before it is a dishonesty problem, how to tell the difference, and what each control actually stops.
Six responses to badge-passing at one site, as the employer costed them
Of the two cheapest items, the one that stops nothing directly is the one that changed the numbers at this site. What each control records about people, and whether it is permitted where you operate, are separate questions and the second is for somebody qualified in the place concerned.
Badge-passing is usually a queue before it is a dishonesty. One terminal, sixty people, a shift that starts on the hour and a rule that treats anybody punching after the hour as late — and passing a badge forward becomes the obvious way for everybody to be on time. Nobody involved describes it as theft, because in most cases nobody is being paid for time they did not work.
The practical lesson in “The Queue at the Terminal” is that visibility is not certainty. For teams researching key person dependency, what teams should check in key person dependency can add time and project context to the operational record, provided the purpose is explained, access is restricted and any material inference is checked through conversation and proportionate human review.
That distinction matters, because the two problems have completely different fixes and the expensive controls address only one of them.
The Europol prevention guides offers another lens on the issue raised in “The Queue at the Terminal”. Compare its principles with the actual record, ownership model and review route rather than importing a generic checklist unchanged.
Telling the two apart
The question is whether the person whose badge was presented was there. If they were, the record is wrong and the pay is right. If they were not, both are wrong.
Those produce different evidence and warrant different responses, and the investigation that does not separate them tends to treat every instance as the second.
- Punches clustered in the two minutes either side of the hour.
- Several badges read at one terminal in rapid succession.
- The same small group, in the same order, most days.
- No corresponding gap in output, task completion or anything else.
- The practice stops immediately when a second terminal appears.
That last one is the test, and it is cheap to run.
What the controls actually stop
A PIN stops casual passing and fails the moment the PIN is shared, which in a group that was already passing badges takes about a week.
A photo at each punch stops most of it and depends entirely on somebody reviewing photos, which nobody does after the first month unless the review is scheduled and owned.
Biometrics end credential sharing as a category and do not touch anybody who punches in and then leaves, which is the other half of the problem and the half that actually costs money.
None of them addresses the queue.
The cost nobody puts in the business case
Biometric systems introduce obligations of their own — what is collected, how it is stored, what happens when somebody will not or cannot enrol, and what is permitted in the first place. Those are questions for somebody qualified in the place concerned, and they arrive before the procurement rather than after it.
They also change the relationship. A workforce that was passing badges to avoid being marked late, and is met with fingerprint readers, learns something about how the employer read the situation.
Running the cheap test first
- Measure the queue: stand at the terminal at a shift start and time it.
- Count punches in the five minutes either side of the hour.
- Check whether lateness has any consequence, formal or informal.
- Add a second terminal, or move the shift start by five minutes, for a month.
- Measure the same things again.
- Then decide what problem is left.
Step four costs a few hundred pounds and a conversation. In a sizeable proportion of sites it removes most of the behaviour, and what remains is a smaller and clearer problem that can be dealt with on its own terms.
Where the real losses are
Where somebody is genuinely being paid for hours they did not work, the pattern is different: it does not cluster at the hour, it is not a group activity, and there is usually a second record that disagrees — an output figure, a vehicle that did not move, a system nobody signed into.
That is a case worth investigating properly, and the method for it is the subject of its own section here. Confusing it with the queue wastes the investigation on the wrong people.
What the terminal's position does
A single physical fact drives much of this: where the terminal sits. On the way in, before the lockers, after the gate, at one end of a long building.
Walking the route people take at a shift start, with a watch, is the single most informative half hour available in this whole area — and it is the step that gets skipped in favour of a procurement decision.
What the terminal was sold to do
Time terminals are sold on their controls and bought on their price, and almost nobody specifies throughput — how many people per minute the device can process in the conditions of an actual shift change.
That number exists and the vendor will give it. Comparing it with the headcount at a shift start is a two-minute calculation that decides whether the queue described here is inevitable, and it is the single specification most worth asking for before buying anything.
Saying what was decided
Whatever is done, write down what the problem was understood to be and why the response was chosen. One paragraph.
An employer that installed biometrics to solve a queue has made an expensive and irreversible decision on a misreading, and the only thing that would have caught it is somebody writing the reasoning down while it could still be questioned.