Gas-station reconciliation, start to finish

A gas station runs on two sets of books, and both of them lie a little every day. The nightly close says what happened at the counter. The bank statement says what happened to the money. Keep them honest and you know where you stand; leave them alone and the gaps compound quietly until month-end, when they’re just numbers you can’t explain.

This guide walks through all of it — the close, cash, lottery, fuel, and the bank statement — with the thresholds that matter and links to deeper pieces on each. If you read one thing about the back office of your stores, read this.

Two sets of books

Reconciliation just means checking that two records of the same thing agree. At a station you’re doing it in two places.

The nightly close is the day’s story told in paper: the register report, the payout sheet, lottery slips, tank readings, the deposit. The question it answers is did tonight add up?

The monthly bank statement is the money’s story: every debit and deposit that actually cleared. The question it answers is did the money go where the books say it went?

Most owners watch the first and skim the second. Both hide different kinds of mistakes, and you need both to know your real numbers.

The nightly close: over, short, and cash

Over-and-short is the oldest number in retail: what the drawer should have held versus what it did. The arithmetic is simple — register rung, minus legitimate payouts, compared to the deposit. The hard part is timing, not math.

A consistent shortage means an error that repeats, or a hand in the till. A consistent overage is also a warning — it often means under-ringing, where cash comes in for a sale that was never rung. Either way, a single day tells you nothing; the pattern tells you where to look. And a pattern is only visible if you’re reconciling daily, while you can still ask the clerk who closed.

More on this: why your register is short every day and doing over-and-short by morning instead of at month-end.

Lottery: two cash flows, one drawer

Lottery is the one thing at the counter that moves cash in both directions — customers buy tickets, and the same drawer pays out winners. If lottery cash is blended into the register, every payout reads as a shortage and every unrecorded sale reads as an overage. The register looks wrong for reasons that have nothing to do with the register.

The fix is to reconcile lottery against its own record — the slips, the settlement, the scans — not the till, and to keep lottery cash tracked separately. A lottery discrepancy that repeats is a known way stores lose a few thousand dollars a year without seeing where it went. See lottery cash isn’t register cash.

Fuel: the tank equation and the BOL

Fuel is the highest-dollar thing that moves through a station, and the easiest to stop scrutinizing because it mostly runs itself. Reconciling it is one line of arithmetic:

Opening tank level + gallons delivered − gallons dispensed = expected closing level.

Compare that expected level to what the tank actually reads and the difference is your variance. Positive means gallons are missing; negative means the tank gained fuel it can’t account for — usually an unrecorded delivery or a receipt that never reached the books.

Two cautions that trip everyone up. First, variance is only meaningful over tanks you can read on both days — one unreadable stick and the math invents a swing that was never there. Second, the Bill of Lading is the only proof of how much fuel you actually bought; when it goes missing, the loss doesn’t disappear, it just stops being checkable. We’ve seen a single month at one store with six drops and no receipts — tens of thousands of gallons nobody could verify.

More on this: fuel variance explained, the BOL you didn’t check, and $165,000 of fuel nobody could check. You can also estimate what unverified fuel is worth at your throughput.

The bank statement: the second set of books

Not every leak is at the register. The bank statement hides a quieter kind: a vendor bill filed to the wrong expense, a check matched to the wrong payee, a charge that crept up unnoticed. We’ve seen a Pepsi bill booked under payroll taxes — for two months — because nobody went line by line.

Skimming the statement for anything obviously huge will never catch this. Only classifying every debit to the right place does, and the useful version remembers the correction so the same mistake can’t run again. See a Pepsi bill filed under payroll taxes.

The rule that ties it together: flag, don’t guess

Every check above depends on one discipline: a number you can’t verify counts as a failure, not as fine. Not averaged, not filled in with a plausible guess, not skipped so the day looks clean. A bad scan or an unreadable stick doesn’t just lose one figure — it breaks the check that figure was part of, and a guessed value manufactures a variance that’s either fake or hiding a real one.

This is the difference between a tool that helps and one that quietly hurts. Green everywhere looks reassuring, which is exactly why a system that guesses to stay tidy is dangerous — it paints over the days you most needed to see. Across one operator’s 21 reviewed days, 18 came back clean and 3 needed a look — and every day that was actually wrong was already flagged. Nothing wrong sat green.

More on this: read, then check, what happens when it can’t read something, and 21 days, 3 flagged, 0 missed.

Doing it across every store

All of this is manageable for one store. The trouble starts at three, or eight, or twenty — because the checking has to happen every night, per store, and the numbers live on different pieces of paper nobody has time to line up. So it slips to month-end, or never, and the gaps that were catchable become gaps that are just gone.

Past a certain number of stores you stop being able to see them, and the quiet middle store — neither great nor obviously broken — is exactly where a slow leak lives. What you need isn’t more dashboards; it’s a single morning screen that already did the triage: which stores are clean (skip them), which days didn’t reconcile (named, with the figure and the page), and which packets never arrived at all.

More on this: your audit department is you, you can’t drive to every store, and the paperwork that never arrived.

How to actually get it done

The big back-office platforms solve this with a POS integration per store: months of IT work, enterprise pricing, and a rip-out you don’t want. The reason they need all that plumbing is that the last step of reconciliation is getting your numbers into your books — and your workbook isn’t theirs.

There’s another way. Your stores already produce everything reconciliation needs, every night, and already photograph and email it. Read that paper and the plumbing disappears — nothing to install, nothing that changes at the counter. That’s the whole idea behind reading the paper you already have.