Cash Drawer Management in a Cloud POS: Counts, Floats, and Reconciliation

Cash drawer management in a cloud POS comes down to three habits: set a consistent opening float, record every time cash moves in or out, and count the drawer at close so the system can reconcile the counted total against what it expected. Do those three things every shift and the drawer stops being a mystery box. Skip them and every shortage ends in a shrug.

Cash is a smaller share of sales than it used to be, but it has not gone anywhere. In the Federal Reserve's 2026 consumer payments diary, four out of five US consumers used cash in the last 30 days, and the average consumer made six cash payments a month¹. Treat those figures as a snapshot as of publication. The operational difference is simple: cards write their own audit trail, while cash only has the trail you build for it. Building that trail is the whole job. (If terms like float and reconciliation are new, the 30 POS terms every new merchant should know covers the vocabulary.)

What is a cash float, and how big should it be?

The float is the cash already sitting in the drawer when a shift opens: the bills and coins kept on hand to make change. It is not revenue, and at close it comes off the top before anyone talks about takings.

Two rules matter more than the amount. Keep the float identical every day, because a number that never changes makes every variance visible at a glance. And weight it toward coins and small bills, since the drawer's job is making change, not storing value.

On size, the honest answer is that no universal figure exists. You need enough change to survive the first busy hour without a run to the bank, and a cafe ringing up $4 coffees needs far more coins than a furniture showroom. Watch the drawer for a week: if staff keep breaking large bills for small sales, the float is too thin. If it opens heavy with twenties, it is storing cash rather than making change.

What should be recorded while the drawer is open?

Every movement that is not a sale, with a reason attached. That covers change top-ups (adding a roll of quarters), cash drops (moving excess bills to the safe mid-shift), and paid-outs (the courier you paid from the till). In a cloud POS each of these is logged against the session ledger the moment it happens, so the expected drawer balance is a live number rather than a reconstruction at midnight.

Cash drops deserve a habit of their own. Cash sitting in a drawer is risk: miscount risk, theft risk, and a bigger loss if the worst happens. Skimming the drawer back down toward float level whenever it gets heavy keeps the exposure small, and because each drop is recorded, it never muddies the count.

What does the closing count actually reconcile?

At close the math is fixed: opening float, plus cash payments, minus cash refunds, plus recorded additions, minus recorded removals, equals expected cash. The person closing counts the physical drawer, and the system compares the two. The difference is the over/under (counted cash minus expected cash), and it is the single most useful number cash drawer management produces.

Count blind. A blind count means counting the drawer before seeing the expected figure, so the count reports reality instead of confirming a target. It is the difference between an audit and a formality.

A cloud POS adds two things a locked cash register never gave you. The reconciliation is tied to a session, with a record of who opened it and who closed it. And the result is readable from anywhere: the closing report shows opening cash, every movement, the counted over/under, and the closing amount, whether you are behind the counter or at home.

Why do drawers come up short?

Rushed change-making, unrecorded paid-outs, cash refunds handed over without being logged, and occasionally theft. The response to all four is the same: chase patterns, not pennies. A one-off variance of a few dollars is noise, and hunting it costs more payroll than it recovers. The same drawer coming up short on the same shift under the same login is signal, and session-level reconciliation is what makes that signal visible.

This is the quiet argument for treating cash tracking as infrastructure rather than an afterthought. On Final, the loop above is built into the station itself: Station Home prompts for the float at open, keeps the cash ledger during the shift, and requires the count at close, while the End of Session report in the Merchant Hub carries the full drawer reconciliation.

So, what does good cash drawer management look like?

The same float every morning, a recorded reason every time the drawer opens outside a sale, a blind count at every close, and a report that ties each variance to a session and a person. None of it is complicated. All of it is discipline the software should carry for you, not a spreadsheet maintained out of guilt. The rule of thumb: you should be able to say what the drawer holds right now without opening it.

If cash handling is one step in a bigger move off paper, the gentle migration path from cash box to POS maps the stages. And when tonight's close lands on your desk, X report vs Z report explains what the end-of-day numbers actually mean.