Why a denied scan followed by an entry deserves its own review
A rejected scan that is immediately followed by someone walking through is not a random error. It is the clearest sign that a non-member got in behind a paying one. Treating this pair of events as its own review category turns a vague suspicion into something you can check and act on.
The pattern you are actually looking for
Think about what happens at 06:40 in an unstaffed club. A card is presented. The reader says no. Ten seconds later, the barrier turns and two bodies pass. Each piece on its own looks harmless. Together they describe a specific behaviour: someone tried, failed, and came in anyway with help.
Most gyms never see this because they look at entries and denials as separate lists. Grouped as one category, a handful of names start repeating. That repetition is the evidence, not the single incident.
Scans prove acceptance, not passage
A scan only tells you a credential was approved. It says nothing about how many people walked through afterwards. That is a separate layer: comparing what happened at the entrance with the access events, and flagging the mismatches for a person to review.
The output should be small and specific — a short list of flagged moments with supporting footage, processed locally at the site, so a manager spends minutes, not hours. The review never tells you who someone is. It tells you that more people passed than were approved, at a known time, in a known lane.
No system catches everything. Some flags will be innocent. But a club that reviews ten honest flags a week stops guessing about shared memberships and starts enforcing its own rules with evidence — which is what protects the subscription base you already sold.
Frequently asked questions
Why not just count denied scans on their own?
Denials alone are mostly noise — expired cards, wrong side of the reader, frozen memberships. The value appears only when a denial is tied to a passage that follows it within seconds.
Does this tell me who the extra person was?
No. It confirms that more people passed than were approved, at a specific time and lane. Identification is a human and contractual matter, not a machine output.
How much time does reviewing flagged events take?
If the flagging is narrow, a site typically produces a short weekly list rather than hundreds of items. Review is measured in minutes, and most of the value comes from spotting repeat offenders.
Will this damage the member experience?
Honest members never appear in the category, because they scan and walk through alone. The ones who notice are the ones sharing access, and that is the behaviour you want to correct.