When the Card Machine Goes Down

23:20 on a Saturday. Cards stop working.

Twenty minutes later it comes back. Everyone breathes out and the night continues.

The twenty minutes cost more than anyone counts, and almost none of it was unavoidable.

What actually happens in the twenty minutes

Not one failure. Several at once:

Tabs cannot be closed. People who wanted to leave cannot, and the queue at the till becomes the visible problem while the real one is upstream.

New orders slow or stop. Staff hesitate to serve when they are not sure the bill can be collected.

Somebody has to decide. Take details and bill later, hold the tab, accept cash only, or stop serving. ⚠️ That decision has to be made in the moment by whoever is present, and in most venues nobody has decided it in advance.

The room notices. Which is worse than the money.

The decisions to make now, not then

Four, and they take one conversation on a quiet afternoon:

One — what happens to open tabs. Held, converted to cash, or details taken. Pick one, in advance, and tell everyone.

Two — do we keep serving? Almost always yes, and saying so in advance stops five people making five different calls.

Three — who can authorise an exception, and up to what amount. If the answer is "the manager", the manager will be at the till, not on the floor.

Four — what we tell guests. One sentence, agreed, so the room hears the same thing from everyone. Inconsistent explanations do more damage than the outage.

The backup that is not a backup

Most venues have a second way to take payment. Fewer have one that works when the first fails, because they frequently share the same dependency — the same internet connection, the same provider, the same router.

⚠️ The check: does the backup use a different path? A second terminal on the same connection is not a backup. A mobile-network terminal, or a phone-based reader on cellular, is.

Test it during service, once. A backup nobody has used at 23:00 is a backup that will not be found, charged, or logged into when it is needed.

Cash

Worth saying plainly: the venues that handle an outage best are the ones that can still take cash.

That means a float, someone who knows the procedure, and the ability to give change for a large note at 23:30 — which is exactly what a mostly-cashless venue no longer has.

⚠️ This is a real trade-off, not a recommendation. Holding cash has its own cost and its own risk. But if you have gone cashless, the outage plan cannot quietly assume cash is available, and many do.

The part almost everyone skips

The recovery afterwards.

Tabs held during the outage, details taken, promises to settle later — these are the ones that go uncollected, because the night ends, the crisis passes, and nobody owns the list.

The fix is small: one place where every outage tab is written down, and one person who checks it the next day. Without that, the twenty minutes turn into a permanent loss that nobody ever attributes to the outage.

What it actually cost

Worth counting once, after it happens:

Tabs never collected. The direct loss.

Time to clear the queue afterwards, and what was not being served during it.

Guests who left without ordering again. Hard to measure, usually the largest, and visible only if you look at whether that night's guests came back.

Three to hold

Outages per quarter, with duration. If nobody records them, they are all the first one.

Uncollected tabs traced to outages. The recoverable loss.

Whether the backup path has been used in the last ninety days. ⚠️ Untested is the same as absent.

Where the record has to sit

If a held tab exists only as a note on paper behind the till, it does not survive the end of the night — and the follow-up the next day has nothing to work from.

tasteck keeps open tabs against the guest record, so a tab held at 23:20 is still there on Sunday afternoon with a name attached.

The machine will come back. The tabs from those twenty minutes are the part that does not come back on its own.