How to balance a store till: blind counts and variances

The afternoon shift ends. The cashier prints the report, sees the system expects Bs 1,580 in cash, counts the drawer and writes down 1,580. Everything balances, apparently. Three weeks later the accountant finds that deposits do not match sales, and nobody can say which shift the shortage started in. The problem was not the count: it was counting while knowing what the answer should be.

Balancing a till is not adding up banknotes. It is a procedure with fixed steps, owners and clear limits. This article builds it end to end, and it works the same with a notebook, a spreadsheet or a system.

1. Opening with a fixed float

Every shift starts with a defined change float (for example, Bs 500 in notes and coins). The procedure:

1. The supervisor hands over the float, counted in front of the cashier.
2. Both sign or record the opening amount.
3. If the float is short, it is noted before selling, not at the end.

The float is always the same amount. If change runs out, it is requested from the supervisor and recorded; it is never “borrowed” from another till.

2. Blind count vs count with the expected amount

	Count with expected amount	Blind count
What the cashier sees	The amount the system expects	Nothing: they count and declare
Risk	Adjusting the count to the expected figure	None of that kind
What it reveals	Only variances the cashier is willing to show	The real variance
Who compares	The cashier	The supervisor or the system

In a blind count the cashier counts cash by denomination, declares card and QR totals from their receipts, and hands over. Only then does someone compare against the expected amount. It is the only way the variance that shows up is the variance that exists.

3. Balancing by payment method

Balancing only the total hides errors. Each method has its own formula and its own evidence:

• Expected cash = opening float + cash sales + other till income − cash refunds − cash drops − payments made from the till.
• Card: system total against the card terminal’s batch close.
• QR: system total against payments received in the bank account or app, matched by reference number.
• Other methods (vouchers, gift cards, credit): against their receipts or balances.

Every cash drop and every payment from the till needs a signed receipt. Without it, the system cannot deduct it and it will show up as a shortage.

Worked example

Suppose a shift with these hypothetical figures:

• Opening float: Bs 500.
• Cash sales: Bs 3,200.
• Cash drop to the safe: Bs 2,000.
• Payment to a small supplier from the till, with receipt: Bs 120.

Expected cash = 500 + 3,200 − 2,000 − 120 = Bs 1,580.

The blind count gives these results:

Method	Expected	Declared	Variance
Cash	1,580	1,552	−28
Card	2,100	2,050	−50
QR	1,450	1,500	+50
Total	5,130	5,102	−28

Looking only at the total, you would see a Bs 28 shortage. But balancing by method shows more: card and QR have variances of equal size and opposite sign. Most likely a Bs 50 sale paid by QR was recorded as card. It is not lost money, but it is a recording error that, left uncorrected, breaks the day’s bank reconciliation.

The real shortage is in cash: Bs 28, which is 28 ÷ 3,200 ≈ 0.9% of cash sales.

4. Tolerances

Tolerance is the variance your company accepts without a formal investigation. It is an internal policy, not a legal rule, and it is worth defining:

• As an amount, not only a percentage, so a large till cannot hide big shortages.
• Per shift and cumulative per cashier: three small variances in a row in the same direction are already a pattern.
• Separately for shortages and overages. An overage is also a variance: it can point to wrong change or an unrecorded sale.

Suppose your policy tolerates up to Bs 10 per shift. In the example, the Bs 28 cash shortage exceeds the limit and is investigated; the card/QR mix-up is corrected even though the net is zero.

5. How to investigate a variance

1. Recount the cash, preferably by another person, by denomination.
2. Check the receipts for cash drops and payments from the till: are they all recorded?
3. Match by reference the card and QR payments against the terminal batch and the bank.
4. Review voids, refunds and manual discounts for the shift, with their time and who authorized them.
5. Look for sales recorded under a different method than the one used, as in the example.
6. Check the opening float: was it counted correctly at the start?
7. Record the outcome: cause found, amount, action taken and owner.

If the variance involves an invoice issued by mistake, remember that voiding it follows the SIN electronic invoicing procedure; it is not a tool for balancing the till. Validate the approach with your accountant.

6. Shift and supervisor controls

• One cashier per till per shift; nobody else uses that drawer.
• Cash drops when cash exceeds a set amount, with a receipt signed by both.
• Voids and refunds authorized by the supervisor.
• Surprise counts during the shift, also blind.
• The supervisor reviews each close the same day, not at the end of the week.

Indicators

Indicator	How to calculate it	How to read it
Shifts out of tolerance	Shifts with variance > tolerance ÷ total shifts × 100	If it rises, review training or controls
Net variance	Overages − shortages for the period	Negative and persistent: money is leaking
Absolute variance over sales	Sum of variances without sign ÷ cash sales × 100	Measures cash handling quality even when the net is zero
Variances per cashier	Each cashier’s variances ÷ their shifts	Detects individual patterns
Voids per shift	Number and amount of voids	Unexplained peaks deserve a review

Common mistakes

• Letting the cashier see the expected amount before counting.
• Balancing only the total and not each payment method.
• Accepting drops or payments without a receipt.
• Offsetting today’s shortage with yesterday’s overage.
• Reviewing closes once a week, when nobody remembers the shift.
• Not recording the cause of each variance investigated.

Taking it into a system

With a notebook and a spreadsheet this method works, as long as nobody sees the expected amount before counting. In an ERP such as MKA the till close can be blind from the till itself: the cashier counts without seeing the expected amount and the variance by payment method goes to the supervisor’s incident queue (see point of sale).