First of all: what going out actually means
Going out means what is in the drawer does not match what should be there. And what should be there is a simple sum: opening float, plus cash takings for the period, minus cash-outs.
If that subtraction does not equal what you count, there is a difference. A discrepancy of a few cents is normal rounding noise. One of tens of euros repeated across several days is not.
Opening float β¬150.00 + Cash takings for the shift + β¬842.30 β Cash banked β β¬400.00 βββββββββββββββββββββββββββββββββββββββββββββ = Expected in the drawer β¬592.30 Physically counted β¬589.80 βββββββββββββββββββββββββββββββββββββββββββββ Discrepancy (short) ββ¬2.50
The seven causes, most common first
They are ordered by real frequency, not by seriousness. Worth ruling them out in this order before thinking badly of anybody.
- Wrong change. Cause number one, and nearly always at peak time. Tends to give small discrepancies in both directions.
- Unrecorded cash-outs. Somebody takes money out to pay the bread delivery or the ice and nothing gets written down. Gives round discrepancies and always short.
- Comps and staff drinks. A coffee given away that leaves the stock but not the drawer. It skews the stock more than the till, but it muddies the analysis.
- Bad voids. A payment voided after the money went in, or the other way round.
- Tips mixed into the drawer. The quietest one: the money is where it should not be and vanishes from the reconciliation without anyone taking anything wrongly.
- The midnight cut-off. If the system closes the day at 00:00 and you close at 02:00, two hours of takings fall into the next day and nothing balances.
- Theft. It exists, but it is last on the list. You tell it apart because it is systematic, always in the same direction and usually lines up with the same shifts.
The one nobody looks at: the midnight cut-off
It is a design fault, not anybody on the team, and that is why it takes so long to find. A venue that shuts in the early hours has a shift crossing two calendar days, and if the system splits the period at 00:00 the shift is cut in half.
The consequence is that the close never balances and nobody knows why: the takings after midnight are missing from one day and surplus on the other. The fix is not counting better, it is having the period run close to close.
How to find yours in one night
The technique is always the same: stop writing down a final number and start recording each item separately. A discrepancy can only be located if you can isolate where it came from.
- Count the float when you open too, not only when you close. If it is already out at nine in the morning, the problem is not the shift.
- Record every cash movement in and out with its reason, at the moment it happens.
- Note cash tips separately and never put them into the drawer's sum.
- Count by denomination. A discrepancy of exactly β¬50 is usually a note; one of β¬2.50 is change.
- Keep the history. One discrepancy is noise; the same sign three days running is a pattern.
When to stop looking
A reasonable tolerance in a busy bar is a few euros per shift. Chasing two euros costs more in manager time than the difference is worth.
What is worth doing is setting a threshold in writing β five euros, say β beyond which the close requires an explanation or a second signature. It is not distrust: a large unexplained discrepancy repeats, and an explained one does not.
Frequently asked questions
How much of a discrepancy is normal?
A few euros per shift in a busy venue is rounding and change noise. What matters is not the isolated number but the pattern: same sign, same size and same shift, repeated, is a signal.
Do cash tips go into the reconciliation?
They should not. They go from the customer to the waiter without passing through the drawer, so putting them into the count balances one number and falsifies another. Record them separately, as information.
Why does my close never balance if I shut in the early hours?
Almost always because your system splits the day at 00:00 and your shift crosses it. The period has to run from one close to the next, not midnight to midnight.
Should I have two people sign the close?
In venues with several managers, or with recurring discrepancies, it helps. Not out of distrust, but because it forces the difference to be explained at the time instead of argued about three days later.
Should I count the drawer in front of somebody?
In venues with several managers it helps, and not out of distrust: two people counting reduce miscounts, which are as real a cause of discrepancy as any other.
What do I do with a discrepancy I cannot find?
Record it with its amount and a note of what you checked, and move on. What you should not do is balance it by hand by adding or removing cash: that erases the only clue you would have if it happens again.