Digital & Data

POS System Down: A 4-Phase Protocol for Surviving a Crash Mid-Service

The screen freezes at 8:12pm, the room is full, and the kitchen keeps asking for tickets. This is not the moment to improvise — it's the moment to follow the plan you're reading right now.

In this article
  1. Before it happens: the outage kit
  2. The 4-phase protocol
  3. The protocol is the investment, not the technology

Every POS system fails eventually. The question isn't whether — it's whether your venue has a plan for that moment, or a crisis.

There is no article on this site — and barely one anywhere online — about what to actually do in the minutes right after your POS system dies. There are plenty of pieces on which POS to buy, on fiscal till rules, on cybersecurity and backups. Nobody writes about the 45 minutes that follow the moment the screen freezes, while the room is full and the kitchen is still waiting on tickets.

That's odd, because it happens more often than most owners admit. An update that hangs mid-service. An internet connection that drops while the card reader still needs to be online. A terminal that simply dies — dead battery, black screen — right as the table of eight asks for the bill. The failure itself is rarely long: most outages resolve within the hour. The real damage happens at the tables, in the kitchen and at the till while nobody is sure what to do.

This article covers that gap. Not a technical manual on rebooting a router — your provider knows that better than we do. An operational protocol: what to do in the first two minutes, how to switch to paper for the next eight without slowing service down, how to manage the room rather than just the till for the 35 minutes after that, and how to reconcile everything once service ends. Plus a calculator that shows what an outage actually costs you — with and without preparation.

Before it happens: the outage kit

The difference between a ten-minute hiccup and an hour-long mess is rarely how fast the underlying problem gets fixed. It's how fast the venue switches to a manual way of working without service grinding to a halt. That switch takes seconds if you've rehearsed it once, and it takes minutes — hunting for pads, looking up prices, arguing over who can authorize what — if you're doing it for the first time while the room watches.

Six things belong in that kit, together in one folder or box near the till or in the kitchen, so nobody has to search for them at the moment it matters:

1 Paper dockets or an order padenough for a full service, plus pens that actually write — not a half-dead one from a kitchen drawer.
2 A laminated price listthe whole menu with current prices, so nobody has to guess a dish from memory.
3 An offline card reader or backup terminalmany providers offer a cheap unit that works independently of your main POS — essential if the fault is the till itself.
4 A charged power bankfor tablets, phones or a spare terminal — a dead battery should never be the reason you can't switch over.
5 A mobile hotspot or spare data SIMif the fault is your internet connection, a second, independent network is the fastest workaround.
6 A card stating the comp-authority thresholdwho can give away how much without calling the owner — decided before the pressure builds, not during it.

Nothing on this list is expensive or hard to find. Most venues that already have one assembled it after an outage that ran too long — this article exists so you don't have to learn it the same way.

Ultimate guide Everything about digitizing your restaurant From POS systems to kitchen displays: the complete guide. Read the guide

The 4-phase protocol

Four phases, each with its own job. The first two are about the fault itself; the last two are about the room and the books — and those two get forgotten most often, even though they're the ones that linger longest.

How an outage escalates if nobody does anything

Each phase below stretches out further the longer nobody is sure what the next step is. With a plan, the first two phases overlap instead of running one after the other.

0–2 min
Terminal, network, card reader, or all of it?
2–10 min
Paper to the kitchen, price list at the till, card reader on its own
10–45 min
Honest communication, table triage, applying the comp threshold
After service
Rebuilding the books, checking stock, preventing a repeat

These windows are guidelines, not law — a fault that resolves itself after two minutes never needs to reach phase 3. The whole point of a protocol is stopping every outage from automatically escalating that far.

1. Work out what's actually broken (0–2 min)

The first instinct is usually to panic about the whole POS, when the fault often sits in just one link. Four questions, asked in this order, settle it inside two minutes: is the screen itself doing anything — does the cursor move, does a tap register? Is the internet still working on another device, like a phone on the same network? Is it only the card reader refusing, while orders can still be entered? And is it one device acting up, or have every till and tablet gone dark at once?

One person runs this check — not three people fumbling with the same terminal at once. Decide who that is at the pre-shift briefing, before it's needed: usually the shift lead, or whoever has run the system longest that night. While that person checks, everyone else keeps taking orders — service doesn't stop, only how it's recorded changes for now.

If the answer is clear inside those two minutes — say, the internet is down but the till itself still works offline — you already know which of the six kit items you need. If the answer is unclear, move to phase 2 anyway: the room doesn't wait for a diagnosis.

2. Switch to manual, on purpose (2–10 min)

Going manual isn't a stopgap you abandon the moment the system flickers back — for the length of the outage, it's simply how the venue now runs. Making that mental shift is what separates service that keeps flowing from service that checks the terminal every five minutes hoping it's fixed.

From here, the kitchen gets a paper ticket per order — table number, dishes, any modifications — exactly what a printer would produce, just handwritten. The price list from the outage kit stays open wherever orders are taken, so nobody has to guess a price. Payments run through the offline card reader or, for smaller amounts, cash — and every payment is logged on the same docket, with table number and amount, so nothing goes missing once service ends.

Tell the team what's changing in one sentence, not a panicked announcement to the whole room: "we're running on paper dockets for now, you'll get the price list in a second." Guests rarely need to know the detail — they just notice service carrying on normally, which is exactly the point.

3. Manage the room, not just the till (10–45 min)

By around the ten-minute mark, the outage becomes visible to guests even if service itself hasn't slowed — they see someone with a pad and pen instead of a tablet, or notice the card payment taking longer. That's the moment to name it briefly and honestly to anyone who explicitly asks: "our system's having a moment, we're running on paper tonight — your order and bill will be sorted properly." Nothing more than that, and certainly no apologizing to tables that haven't noticed anything.

Seating new tables during an outage doesn't have to stop, but the pace can drop — three tables served well beats eight served badly. If the waitlist grows, that's a fair reason to let people in a little slower than usual, rather than pressuring service to compensate for the system.

This is where the comp-authority threshold from the outage kit comes into play: who can give away a drink, dessert or discount, and up to what amount, to a table that's clearly waiting longer than usual? Without that agreement, it's a judgement call made under pressure. With it, it's a decision that was already made before the outage began.

4. Close the till and find the cause (After service)

The next morning — or, if the outage struck early in service, that same night — every paper docket gets manually entered into the POS once it's working again. This is where a sloppy round of dockets comes back to bite: if a table number, amount or payment method is missing, the reconciliation takes hours instead of minutes. That's exactly why every docket written during the outage needs to carry all three.

Alongside the books, run a quick stock check: was anything served during the outage that never got logged, because the system wasn't tracking it? That usually only shows up days later, when the stock count no longer matches what was actually sold.

Finally: which of the four questions from phase 1 gave the answer? Was it the internet, the terminal itself, the card reader, or something on the provider's end? That answer decides whether the outage kit needs restocking — an extra power bank, say, if a dead battery was the cause — or whether it's a conversation with the provider about why the system has failed a third time this quarter.

What a 30-minute outage costs — with and without a kit

Same outage, same 40-covers-an-hour venue: the only difference is whether the team already knew what to do.

No preparation€506
€470
Outage kit ready€156
€138
Revenue not processed Given away as goodwill

Red is revenue the room couldn't process because of the slowdown itself; amber is what got given away as goodwill to tables that waited too long. Both figures are above — plug in your own numbers in the calculator to see it for your venue.

The pattern behind these numbers is always the same: the faster the first switchover happens, the less the rest of the outage costs. That's exactly why phases 1 and 2 above are timed so tightly — not because two minutes is a magic number, but because every extra minute spent diagnosing multiplies through every phase that follows.

What would your next outage cost?

Plug in your own numbers. The calculator runs the same model as the chart above.

Yes, ready
Revenue not processed
gap between normal throughput and what actually got sold during the outage
That's roughly equal to
guests' worth of revenue, rounded
Estimated goodwill cost
drinks, desserts or discounts to tables that waited too long

This is a simplified model, not an accounting truth: it assumes diagnosis and switchover go faster with a prepared kit (2 vs. 6 minutes), that service recovers speed faster afterward (82% vs. 25% of normal throughput), and that guests waiting past 15 minutes are more often offered something. Adjust the numbers to whatever's realistic for your venue.

Also worth working out: what the kit itself costs to put together — a few order pads, a laminated price list, a secondhand card reader, a power bank. For most venues that's an amount one thirty-minute outage more than pays back, and the kit then lasts for years.

The protocol is the investment, not the technology

No POS system is failure-proof, whatever the sales rep promises. What is within your control is how long an outage takes before the room notices it — and that isn't decided by which system you bought, but by whether the team knows what to do in the first two minutes.

Print this article, or just the four phases and the outage-kit list, and pin it up where the till stands. Walk the team through it once at a pre-shift briefing — five minutes that pay for themselves the night the screen freezes.

And if you already know this happens to your venue twice a month rather than twice a year, this stops being a question of having a plan — it becomes a conversation with your provider about why, or about a second, independent system as a backup.

Frequently asked questions

Do we always have to tell guests the POS is down?

Only the ones who explicitly ask or are clearly waiting longer than usual. A short, honest line covers it — "our system's having a moment, we're running on paper tonight" — without over-apologizing to tables that haven't noticed a thing.

Can we just wait for the system to come back instead of switching to paper?

For a fault of two or three minutes, sure, especially if you already know it's a brief network blip. Beyond that, waiting costs more than switching — every minute without a clear way of working is a minute where orders get lost or entered twice.

Who decides to give something away free to a table that's waited too long?

Decide that in advance, not during the outage itself. A common setup: the shift lead can authorize up to a fixed amount without checking in, and anything above that goes through a call to the owner or manager — written on the card in the outage kit.

How do we stop paper dockets going missing after the outage?

One fixed spot where every docket from the service lands once it's been paid — a tray by the till, not loose in aprons. Number the dockets in advance so you can spot afterward if one's missing.

Do we really need an offline backup card reader if we also take cash?

Most guests carry little cash, and the night of an outage is exactly when you don't want part of the room unable to pay. An offline card reader is relatively cheap and covers exactly that gap — it doesn't need to be linked to your main system to work.

How often should the team practice the 4-phase protocol?

Walk through it thoroughly once at a pre-shift briefing, then a quick refresher whenever new staff join. The protocol itself doesn't need much rehearsal — the outage kit just needs to always live in the same spot, so nobody has to think about where.