How to stop cash theft at a parking lot
Cash theft at parking operations is not a rumor problem, it is a court-record problem: attendants at one Smithsonian museum lot took $1.4 million over three years, a transit contractor settled for $5.6 million after employees skimmed millions more, and fraud examiners say a typical skimming scheme runs a year and a half before anyone notices. The schemes are well documented, and so are the controls that stop them. This is the playbook, from the cheapest process fix to the record-keeping that makes skimming visible at all.
How the money actually leaves
Almost every parking cash loss is a variation on one move: money is taken and no record is created, so there is nothing to reconcile against. The classic forms all show up in prosecutions and audit reports. Skimming: the attendant collects the fee and never rings it. Ghost vehicles: the customer pays, gets no ticket, and the entry is never logged. Manual gate opens: a car is waved out on a claimed system failure, with no matching payment. Void and comp abuse: a real cash sale is zeroed out on paper and pocketed. Report forgery: the shift summary is edited downward to match a short deposit. Free favors: friends and regulars wave through, booked as nothing at all. And off-the-books monthlies: a regular pays cash to a person instead of a parking lot.
The court cases show how far this goes when nothing checks it. At the Smithsonian's Udvar-Hazy Center lot, a site manager and booth attendants took about $1.38 million in cash entrance fees from more than 92,000 visitors over three years, by withholding ticket stubs and unplugging the electronic vehicle counters so the car count could never be compared with the cash. At a New York garage, a shift manager was charged with taking over $110,000 in a year and a half by rewriting cashiers' daily reports. In Massachusetts, a parking contractor paid a $5.6 million settlement after employees skimmed cash for roughly three years across commuter lots.
Why it runs for so long
Fraud examiners' data puts a number on the quiet: the median skimming scheme runs about eighteen months before detection, precisely because skimmed money never enters the books, so ordinary bookkeeping cannot see the hole. Small operations get hit hardest. Organizations under a hundred employees suffer higher median fraud losses than large ones, for the plain reason that one person often takes the cash, counts the cash, and reports the cash.
A truck yard is a near-perfect environment for this: cash-heavy, lightly staffed, busiest at night, and often run on a paper log that the same attendant both writes and totals. If the only record of a truck's stay is a line the attendant chose to write, then the attendant, not the owner, decides what the yard earned.
The one control that actually catches skimming
Every flagship case ends the same way, in both directions. The Massachusetts scheme was caught by a routine comparison of vehicle counts against parking receipts at one station: more vehicles than money, scheme over. The Smithsonian scheme survived three years only because the conspirators had unplugged the counters, killing exactly that comparison. The lesson is symmetrical and worth stating plainly: cash theft is caught by an independent count of vehicles reconciled against recorded revenue, and it survives wherever that count does not exist or can be quietly disabled.
Everything else, cameras, gates, honesty policies, is secondary to this one question: does a record of every vehicle exist, independent of the person handling the money? On a truck yard the answer does not require hardware. A photo of every truck at check-in, timestamped, with the plate read automatically, is an independent count, and unlike a loop counter in a booth, a missing photo is obvious the moment anyone looks at the log.
The control stack, cheapest first
Most of what auditors recommend costs nothing but discipline. One till per person, never shared, so every dollar has exactly one owner. A counted, dual-signed starting bank at every shift change. Reconciliation at shift close against recorded transactions, with a small tolerance and a written note for anything beyond it. Variance trending per attendant over a month or three, because a pattern of small shortages says more than one big one. Surprise counts on no schedule. A weekly look at every comp, void and discount against who authorized it. Rotation and enforced vacations, since the longest-running schemes belong to people who never leave their post. And daily deposits, double-counted, so cash never pools.
The hardware tier, gates, ticket dispensers, loop counters, license plate cameras, exists mostly to automate the independent count and the exception log. It works, but it is capital-heavy, and the Smithsonian case is the standing proof that equipment alone is not a control: the lot had gates, tickets and counters, and lost $1.4 million anyway, because the people running the reporting layer were the people taking the cash. The control is the reconciliation, not the machine.
What this looks like on a truck yard
TIRpark implements the control stack as a side effect of how the yard already works. The attendant photographs each truck at the gate, so the independent vehicle record builds itself, one entry per truck, timestamped, with the plate read on the phone. Every payment is taken under the attendant's own PIN, so attribution is automatic and per-person totals are live: shift reconciliation stops being arithmetic on trust and becomes comparing the box to a known number. The owner sees occupancy, revenue and debt remotely, per shift and per person, which leaves report forgery nothing to forge. And patrol mode is the surprise count on demand: walk the lot, re-photograph what is standing, and anything that never went through the log, the friend in the corner, the night that was collected but never recorded, surfaces immediately.
Process still matters. Software does not deposit the cash, and it does not replace the dual-signed bank or the no-notice count. What it removes is the gap every scheme in the court records lived in: the distance between what stood on the lot and what the paper said. When the vehicle record cannot be unwritten and the money cannot be anonymous, the ordinary schemes stop being worth the risk, and most of them simply stop.
Close the gap this month
Put TIRpark on one Android phone and run a month on it, free, no card. Then compare the receipts to the same month last year. The difference is what the paper log was costing you.