Reconciling the till without a long evening

Every cafe has a daily figure that should match and often does not. Most treat a small gap as the cost of being busy.

A tolerated gap teaches everybody that the number is approximate, and an approximate number grows.

Count the same way, every close

A float each closer counts differently produces variances that mean nothing. The method has to be identical whoever closes.

Same float, counted twice, recorded before the takings leave the drawer. That sequence removes most of the noise people blame on busyness.

Separate the payment types before looking for a cause

Cash, card and platform each go wrong for different reasons. A single variance figure mixes them and hides all three.

Card rarely disappears; it arrives late or arrives net of fees. Cash is the one that genuinely moves, and it deserves the attention.

Set a threshold and mean it

Investigating every dollar wastes an evening a week. Ignoring twenty dollars a day loses thousands a year.

A named threshold, with a short investigation above it, keeps the effort proportionate. Nobody then decides it while tired.

Most variance is process, not theft

Wrong change, an item rung at the wrong price, a refund handled badly, a staff meal not recorded. Those cause far more variance than dishonesty.

Treating every gap as a suspicion damages a team quickly. Treating it as a process question usually finds the real answer within a week.

Discounts and staff meals need a button

A discount given verbally and not entered creates a variance that looks exactly like missing cash.

A button for each reason turns an invisible gap into a reported figure. It also tells the owner what discounting really costs across a month.

Reconcile the card settlement, not just the terminal

The terminal total and the money in the bank differ by fees and timing. A cafe checking only the terminal never sees a missed settlement.

Matching the bank deposit against the daily figure catches that, and it is the only way a whole missing day is ever noticed.

Look at variance by shift, not by day

A daily figure hides which shift produced it. The same pattern on the same shift is a training conversation rather than a mystery.

That comparison needs nothing more than the closer's name beside the number, recorded every day.

Two people count anything large

A single person counting a big takings figure carries a suspicion they should never have to carry.

A second signature protects the staff far more than it protects the money, and that is reason enough.

Banking needs a rhythm

Cash held on site is a risk to the business and to whoever is holding it. A predictable banking day with an unpredictable time is the usual balance.

Recording what went to the bank, against what the tills reported, closes the last gap in the chain.

A recurring variance is a training gap

The same shift, the same size, the same direction is rarely mysterious. Somebody is doing one step differently.

Watching that shift for an hour usually finds it, and the conversation afterwards is short.

A variance that moves between shifts is different again, and usually points at the till configuration rather than at any person. A button priced wrongly produces a steady error nobody can find by watching people.

What to change first

• One counting method, used by whoever closes
• Cash, card and platform variances separate before anybody looks for a cause
• A named threshold decides what gets investigated
• Discounts and staff meals each get their own button
• The bank deposit matches the daily figure, not just the terminal
• The closer's name sits beside the variance, so shifts compare