You cannot fix a leak you cannot see, and there is no honest industry number to hand you. Anyone who quotes you a tidy "the average garage loses X percent" is guessing, borrowing someone else's math, or selling you something. Your leak is a function of your gates, your validation rules, your permit population, and your local habits around tailgating and ticket-swapping. The only number that matters is yours.
The good news: you can get a defensible first estimate in one week with data you already have. This is the operator's version of the full method in our parking revenue leakage guide — less theory, more clipboard. Block out a normal seven-day window (no holidays, no events that skew traffic) and run the two counts below.
Part one: the count leak
The count leak is the blunt instrument. It asks a simple question: over one week, how many cars came in, and how many exits can you actually account for?
"Account for" is the load-bearing phrase. Accounted exits are exits you can explain: a paid transaction, a valid permit or monthly pass, a validated ticket. Everything else — cars that left without a matching, explainable exit event — is your suspect pile. That pile is mostly tailgating (two cars, one arm-lift) and drive-offs.
Here is the method, honestly stated:
- Pull entry count for the week. Every arm-lift, loop trigger, or plate read at entry.
- Pull every accounted exit. Paid exits, permit/monthly exits, and validated exits. Sum them.
- Adjust for cars parked at the boundaries. Count what was on the deck when your window opened and again when it closed. If more cars are sitting inside at the end than at the start, those entries have not exited yet — they are not leak, they are still parked. Subtract the difference. Skip this and a busy Sunday night will make your leak look enormous.
- Multiply the leftover by your average ticket value. That is your first-order count-leak estimate.
The critical honesty check, and the thing most back-of-napkin audits get wrong: entries minus paid exits is not your leak. That subtraction silently counts every permit holder and every validated visitor as a thief. On a site with a big monthly-parker base, it will hand you a scary, useless number. You must credit permit and validated exits as accounted before you call anything a leak.
In an illustrative, modeled scenario — 7,000 weekly entries, 6,600 accounted exits after crediting permits and validations, a net of 50 more cars parked at week's end, and a $14 average ticket — the count leak pencils to roughly (7,000 − 6,600 − 50) × $14 = $4,900 for the week. Plug in your own four numbers; the shape is what matters, not the total.
Part two: the underpayment leak
The count leak misses the quieter half, because underpayment looks exactly like a normal paid exit. The arm goes up, a transaction fires, the report shows a clean sale. Underneath it: a ticket-swap that charged the short stay, a validation applied that should not have been, an overstay that talked its way to the early rate.
You cannot count your way to this one. You have to sample.
- Pull a random sample of paid exits from the week — fifty is plenty to smell a pattern.
- For each, find the real entry event (the actual plate or ticket first seen) and compare it to the duration that was charged.
- Flag every exit where the charged time is materially shorter than the real dwell time.
A cluster of "charged 40 minutes, actually parked 6 hours" is not random error. That is the signature of swapping, validation abuse, or overstay slippage. Extrapolate the flagged rate across the week's paid volume, price the gap, and you have your underpayment estimate. Add it to the count leak for a full picture.
The reason this stays invisible on ticket-based sites is that entry and exit are not tied to the same identity. A ticket is a piece of paper; it does not care which car holds it.
How plate access closes each leak
Both leaks trace back to the same root cause: the entry and the exit are not the same verified identity. When the car itself is the credential, that gap closes.
Plate-based access reads the plate on the way in and on the way out — a plate read plus an open/deny decision on-site in under a second, no ticket and no app. Every crossing writes a photo-verified record: a plate photo, a full-vehicle image, and a short clip. That does specific work against each leak:
- Tailgating and drive-offs (count leak): entry and exit are matched by plate, so a car that leaves has a true entry it maps to. Two cars under one arm-lift become two records, not one. The suspect pile shrinks because you can see who actually left.
- Ticket-swaps and overstays (underpayment leak): there is no ticket to swap. The session is anchored to the same plate's real entry-to-exit time, so the "charged 40 minutes, parked 6 hours" trick has nothing to stand on: the record shows the true dwell, and your billing rules run against that record.
- Overstays specifically: overstay detection is built in. Once your payment processor is connected, overstays are billed automatically on exit per your rules. Treat auto-billing as conditional on that processor link — it is not magic that appears the day of install.
Permits and validations move to the same plate-based logic: lists with access windows, bulk import, instant changes, and time-scoped guest passes for a specific plate at a specific site, every use tracked. That is what lets you credit permit traffic honestly in Part one instead of guessing.
None of this requires ripping out your gate. It installs on the barrier and opener you already run — most facilities qualify without replacing equipment — and the reads keep working fully local for up to 30 days if the internet drops.
What to do with your two numbers
Add the count leak and the underpayment leak. That sum is your weekly bleed, measured on your site, with no borrowed averages. Annualize it and compare it against the cost of closing it. If part of your current control is a staffed booth, the guard cost calculator helps you put a real figure on the labor side, and the gate delay calculator lets you model the throughput math around faster, ticketless crossings.
Frequently Asked Questions
Why not just quote me the industry average leakage percentage?
Because there isn't an honest one to quote. Leakage depends on your gate hardware, validation rules, permit mix, and local behavior. A borrowed percentage tells you nothing actionable about your deck. The one-week audit gives you a defensible number that is actually yours.
Isn't entries minus paid exits basically my leak?
No, and this is the most common mistake. That subtraction counts every permit holder and validated visitor as a loss, because they exit without a cash transaction. You have to credit permit and validated exits as accounted before calling the remainder leak. Otherwise sites with large monthly-parker bases get a wildly inflated, useless number.
Why do I need to sample exits when I already counted the leak?
Because the two leaks hide in different places. The count leak catches cars that left without an accounted exit — tailgating and drive-offs. Underpayment hides inside exits that look perfectly normal: the transaction fires, the report looks clean, but the charged duration is shorter than the real dwell time. Sampling real entry-to-exit time against what was charged is the only way to see it.
Do I have to replace my gate to fix this?
Usually not. Plate-based access installs on the barrier and opener you already have, and most facilities qualify without replacing equipment. Send a few photos of your gate and you get a compatibility read and a quote, usually within 48 hours.
Does overstay auto-billing work the day it's installed?
Overstay detection is built in from the start, but automatic billing is conditional. Once your payment processor is connected, overstays are billed automatically on exit per your rules. Until that processor link is in place, you get the detection and the record, not the automatic charge.
Run your own numbers
Spend one normal week measuring your count leak and your underpayment leak — then decide. Start with the full method in the parking revenue leakage guide and model the cost side with the gate delay calculator. When you want a compatibility read, send a few photos of your gate and we will come back with a confirmation and a quote, usually within 48 hours.
Wondering if it fits your gate?
Send two photos and get a clear yes/no in about 48 hours, no sales call required.