Cash-up and reconciliation: catching problems the same day

Cash-up is the daily version of a stock take: reconciling what the till expected against what actually came in, across cash, card, and EFT. It's easy to treat as a formality at close of shift. It's actually the fastest, cheapest point at which a store catches a problem - because the person who can explain it is still on shift, and the details are still fresh.

Three numbers, every shift

The comparison comes down to three figures. The opening float is the cash staff started their till with, entered when the shift begins. Expected cash, card, and EFT is what the system calculates should be there, based on the actual sales rung up during the shift, plus that float, minus anything paid out of the till for things like courier fees or supplies. Counted is what staff physically count at close. The gap between expected and counted is the variance - shown to staff and managers by exactly how much, in either direction, or as zero when the till matches perfectly.

Where the variance actually comes from

• Split payments - part cash, part card - logged against the wrong tender
• A refund processed against the till but not matched back to the original sale
• Change given incorrectly, which is a training issue more often than a dishonesty one
• A sale rung up correctly but the physical cash miscounted at close

None of these are dramatic on their own. A R40 variance on a single shift isn't worth an investigation. The same R40 variance, unexplained, on every shift for a month, is a process problem hiding behind a small number.

Why same-day beats month-end

A discrepancy found at close of day has a short list of possible causes: this shift, this till, these staff. A discrepancy found at month-end has thirty shifts' worth of possible causes and nobody who remembers any of them clearly. The gap between the two isn't just convenience - it's the difference between a fixable process and an unsolvable mystery.

What a cash-up should actually produce

• Expected cash, card, and EFT totals, generated from the till - not reconstructed from memory
• A single number for what was actually counted, checked against that expectation
• A record tied to the shift and the staff member on it, so a pattern across shifts is visible

Budstack generates the expected totals automatically from the day's sales, so cash-up is a comparison rather than a reconstruction - the same evening a shift closes, not weeks later when nobody can explain what happened.

Checking the number instead of trusting it

Staff can open a live breakdown of every sale that makes up the expected cash figure - not just a total to trust, but a list to actually check against what they remember selling. If a variance shows up, that breakdown is the first place to look, before it turns into a conversation nobody can resolve because nobody can see what the number is actually made of.

Offline sales used to break the count

A dropped connection mid-shift used to be enough to throw a cash-up off. The till kept working and sales still rang up offline, but a shift could be closed while one of those sales was still waiting to sync - which made the count look wrong even though nothing was actually missing. Offline sales now queue quietly in the background until the connection returns, and the app blocks closing a shift until everything's caught up, showing staff exactly what's still pending rather than letting them close on an incomplete count.

A visible online/offline indicator sits in the corner of the till at all times, so a connection issue is something staff notice as it happens, not something they discover only once the numbers don't add up at close.

Managers don't have to go looking for it

A cash difference bigger than a Rand amount set per store now notifies the manager automatically the moment a shift closes with a variance over that line - instead of the difference only surfacing later, if and when someone remembers to check a report. And on the rare occasion a sale still arrives late despite the offline protections, the report flags it directly next to the variance it explains, so a manager sees "this is a late sale" instead of an unexplained number to chase down.

The same logic applies to stock, just on a longer cycle - see what a stock take should catch before it becomes a shrinkage problem. And when a variance does turn out to need real investigation rather than an easy explanation, who had access to that till matters more than the size of the number.