The 8pm test: why POS reliability is a revenue number

Reliability gets filed under IT. Someone in the back office signs a contract with an uptime figure in it, feels reassured, and moves on. That is the wrong drawer. For a restaurant, reliability is not an IT footnote — it is a revenue number, and you find its real value at 8pm on a busy Friday.

That is the only test that matters. Not the demo. Not the spec sheet. What happens to your point of sale when the room is full, twelve tickets are open, and the internet wobbles.

The 8pm test

At 8pm on a Friday the floor is at capacity. Runners are moving, the pass is loaded, and the card machine is going every couple of minutes. Then the line drops — a router reboots, the ISP hiccups, a construction crew two streets over nicks a cable. It does not matter why.

Here is the fork in the road. A cloud point of sale with no real offline mode stops taking orders the moment it loses the connection. The screen spins. The host can't seat, the server can't fire a ticket, the kitchen goes quiet, and a full room turns into a queue of people deciding whether to wait or leave. A point of sale with a genuine offline mode does something different: it keeps writing tickets locally, the kitchen keeps cooking, cards keep getting captured, and every order syncs the moment the line comes back. The guest never sees the outage.

Same building, same wobble, two completely different Friday nights. One of them shows up in the daily total.

Do the math on one bad hour

The cost is not abstract. A cloud point of sale without robust offline capability is, plainly, an operational disaster waiting for a bad Friday — and the bill is easy to size.

Take a mid-sized venue running 60 orders an hour at peak. When the system goes dark, that is roughly $3,000 an hour walking out the door — not deferred, not delayed, gone. The people who left don't come back at 9pm to place the order they abandoned at 8pm.

Now stop treating the outage as a freak event. One widely used cloud point of sale suffered two separate outages lasting more than ten hours each in a single month, and operator-facing incident logs record more than 300 outage events on that platform since 2022. That is not a run of bad luck. That is a pattern you are underwriting every Friday you stay on it.

Reliability is a number on your P&L

Put those two figures side by side and the reframe writes itself. A ten-hour outage at $3,000 an hour is $30,000 off a single day. That is not an IT ticket. That is a line item bigger than most restaurants' monthly software spend, monthly rent, or a line cook's quarterly wages — spent, involuntarily, on downtime.

Which is why "reliability" belongs on the P&L conversation, not the IT one. The question is not "is the uptime figure good enough for the back office." The question is what the room does when the connection drops mid-service, because that answer has a dollar value and it lands in your total.

The honest answer to the first question sorts the market fast. A point of sale that only lives in the cloud has to say no — no connection, no orders. That is the whole risk in one word.

dojofood's restaurant POS is built around offline mode for exactly the 8pm scenario: it keeps taking orders during a connectivity loss, keeps the kitchen firing, and reconciles every ticket and payment the moment you're back online. Nothing waits for the router. If you want to see how that behaves on a bad Friday rather than a good demo, book a 20-minute walk-through and ask us to pull the cable.