When the internet goes down: a front desk checklist for cloud-based hotels
This guide provides a step-by-step hotel internet outage plan for cloud-based properties. It covers preparation, manual check-in procedures, and post-outage reconciliation.
Contents
Key takeaways
- In a browser-based hotel system the screen is a window onto a server elsewhere: with no connection the desk can neither read nor write, so the fallback is paper and a procedure, not a hidden offline mode.
- Preparation is the only part you control: a second connection, a nightly printed in-house list, pre-numbered slips and a named person for each job.
- During an outage one numbered paper log and one decision-maker prevent the two failures that cost money, a room sold twice and a charge posted twice.
- Identity notifications and fiscal documents do not disappear with the connection; record the data at check-in and complete the steps when the system returns, with timing confirmed by your provider and accountant.
- Catch-up is a reconciliation, not a typing exercise: post each slip once against its number, then compare the paper log with what the system believes happened.
A hotel internet outage plan is a written procedure for running the front desk on paper while the browser-based system is unreachable, then reconciling every manual record afterwards. It has three parts: preparation (a second line, nightly printed lists, pre-numbered forms), the outage itself (one log, one person in charge), and a clean catch-up.
The screen freezes halfway through a check-in. The guest is at the desk, the room is clean, the key is blank, and the system that holds the room, the rate and the bill cannot be reached. This article walks through what stops, what to prepare in advance, how to run the desk during the outage, and how to catch up afterwards without double postings or missing records.
What stops when the internet stops?
In a browser-based system the screen is a window onto a server somewhere else, so with no connection the desk can neither read nor write. What stops is not one feature but the shared record of rooms, guests and money.
The room inventory is the first loss. The only place that says which room is free, which is occupied and which is out of order becomes unreadable, and the desk starts working from memory. Key card encoders are usually driven by the same system, so no connection means no new key unless the encoder has a standalone mode that someone has actually tested.
Posting stops at the moment of the charge. Room, restaurant, minibar and spa all write into the same guest bill, and none of them can be written. Rates and availability stop flowing to the channels, so the risk moves off the desk and onto the booking sites, where a room can still be sold while the hotel cannot see it. Night audit cannot close the business day, which delays the date-based reports and postings that follow it.
Compliance steps that run through the system pause, but the duty behind them does not. Identity notifications and fiscal documents still have to be produced for the stays that happened. What still works is the building, the rooms, the guests, the staff, a printer that does not need the network, and paper.
| Function | With no connection | What the desk does instead |
|---|---|---|
| Check-in and room assignment | No record of free rooms | Printed in-house list and a paper floor plan |
| Key card encoding | No new key | Standalone or mechanical procedure; log every key |
| Posting a charge | No guest bill | Pre-numbered slip signed by the guest |
| Balance lookup | No folio | Printed balance column and the day's slips |
| Night audit | Cannot close the day | Close it in order once the system returns |
| Channel updates | Rates and availability frozen | Decide stop-sell rules in advance |
| Guest app and online offers | Unreachable | The front desk becomes the channel |
| Identity notification and e-invoice | Cannot be sent | Prepared on paper, completed later |
What should be ready before the connection drops?
Preparation is cheap and boring, and it is the only part you control. Four things matter: a second connection, a nightly printed list, pre-numbered paper, and a named person for each job.
A second line from a different provider, ideally entering the building by a different physical path, is the first item. Add a charged mobile router as a third option, with a known data limit. Test the failover on a schedule: unplug the primary line at a quiet hour, time how long the backup takes, and write down which devices do not come back by themselves, usually the printer, the key encoder and the card terminal.
Print the in-house list at a fixed point every night, after the business day closes: arrivals, room numbers, departure dates, balances, groups, special requests and VIP notes. Print a fresh copy at every shift handover, the same way you would for any front desk shift handover checklist. Keep pre-numbered charge slips and registration forms at the desk, with a log of which numbers were issued to which staff member.
Put a contact sheet on paper: internet provider, IT provider, PMS vendor, channel manager, key card vendor, e-invoice integrator and accountant, with names and mobile numbers rather than a shared inbox. Name the roles in advance: who calls the provider, who owns the paper log, who decides to stop selling, and who speaks to guests.
Checklist
- Second internet line from a different provider (owner: IT, checked monthly)
- Mobile router charged, data limit known (owner: duty manager)
- Monthly unplug test, results written down (owner: IT)
- Nightly in-house list printed after night audit (owner: night auditor)
- Fresh list at each shift handover (owner: incoming shift leader)
- Pre-numbered charge slips and registration forms in stock (owner: front office manager)
- Paper contact sheet at both terminals (owner: front office manager)
- Named decision-maker per shift (owner: general manager)
- Uninterruptible power supply on router and printer (owner: IT)
- Card terminal with its own connection where available (owner: IT)
During the outage: how do you check in a guest and take a payment?
Run the desk from one numbered paper log and one person's decisions. The goal is not to look normal; it is to leave a record you can post once, correctly, when the system returns.
One log, one writer: every arrival, key, charge and message goes into a single numbered sequence with a time and initials. Two parallel notebooks are how double postings begin. Check-in means completing the paper registration form, assigning the room from the printed list, and marking it as taken on a paper floor plan so a second terminal does not sell it again.
If the encoder needs the system, switch to the mechanical or master key procedure and record every key handed out against a room and a name. Charges go on a pre-numbered slip with room, item, amount, time and initials, signed by the guest, with a copy given to the guest at the moment of the charge. For payments, use a card terminal on its own connection where you have one, and record cash on a separate cash sheet that references the slip number, not a running total.
Stop-sell deserves its own decision. With no visibility you can sell a room twice or sell it at a stale rate, so decide in advance whether the answer is to reduce availability through the channel manager, ask the provider to close channels, or accept the risk with a manual block, the same trade-off behind any stop sale decision. Tell in-house guests what is down: the guest app and the online offers will not answer, so the front desk is the channel, and a short written note saves repeated explanations.
Key pointOne log, one writer, one decision-maker per shift
Every manual action gets a number, a time and initials, and nothing is posted later without that number.
Do identity notifications and invoices still have to go out?
The connection is a tool, not the duty. Identity notifications and fiscal documents still have to be produced for the stays that happened, so record everything on paper at the moment of check-in and complete it when the system returns.
Capture at check-in the same identity data you would type into the system, following your normal KBS check-in routine, and note the actual time the guest arrived, because that is the fact you will need later. Timing is set by the authority and interpreted by your provider and your adviser; ask both in advance and write the answer into the procedure. Fiscal documents are issued through a licensed integrator, so if the integrator is unreachable, prepare the document and send it when it is back, and keep a person reviewing e-invoices before they go out.
Never skip, always record. The failure that hurts is not a late document but a missing one, or a stay with no identity record at all. Tell the accountant on the same day, with the paper slips in hand, so the documents are issued in the right period and nothing is reconstructed from memory weeks later. Paper forms carry personal data, so lock them, keep them only as long as your retention rule allows and destroy them on schedule.
Watch outAn outage does not suspend the duty
Confirm timing, scope and any penalties with the authority's guidance, your PMS provider and your accountant before an outage happens, and record the check-in time on paper regardless.
How do you catch up afterwards without double postings?
Catch-up is a reconciliation, not a typing exercise. Every paper slip should be posted once, against a reference, and then checked against what the system thinks happened.
Post from the slips in number order and write the slip number into the system's reference field, so a second person can see at a glance that it is already done. Then compare the paper log against the system's arrivals, departures and postings, looking for the two failure modes: a slip posted twice and a slip never posted.
Channels need the same treatment. List every reservation that arrived by phone or email during the outage, compare it with what the channels sold, and check for overbookings and for rates that were pushed stale. Room status matters too: housekeeping worked on paper, so confirm clean, occupied and out-of-order rooms before the next check-in wave rather than trusting the screen. If night audit did not close, close the business day in order instead of jumping dates, or the reports that follow will not line up.
Finish with a one-page incident note: when it started, when it ended, what was done on paper, what broke, what to change. It is the only way the checklist gets better.
What should you ask your PMS vendor about offline behavior?
Ask before you sign, and ask again every year, because offline behavior is a product decision the vendor has already made, and it is worth asking directly.
Ask what the desk can see and do with no connection, and accept an honest "nothing, plan for paper" as a good answer. Ask whether the key card encoder works without the system, and ask for the standalone procedure in writing. Ask how manual entries get into the system and whether there is a reference field and a reconciliation report. Ask what happens to rates and availability on the channels when the system cannot be reached, and whether you can export the in-house list yourself at any time rather than waiting for a support ticket. Ask for the escalation path and response time written into the contract, and how planned maintenance is announced. Then ask the same questions of the channel manager and the e-invoice integrator, because the outage is shared across all three.
| Question | Why it matters | A good answer sounds like |
|---|---|---|
| What can the desk see offline? | Decides paper versus partial cache | A clear list, or an honest "nothing" |
| Does the key encoder have a standalone mode? | Keys are the first thing a guest notices | A named mode and a tested procedure |
| How do offline entries get posted? | Prevents double postings | A reference field plus a reconciliation report |
| What happens to channels when unreachable? | The channel risk outlives the outage | Held values or closed channels, stated plainly |
| Can we export the in-house list ourselves? | Removes dependence on a nightly print | An export, not a ticket |
| What is the escalation path? | A phone number is not a plan | Named contact, hours, definition of an incident |
| How is planned maintenance announced? | Most outages are scheduled | Advance notice with a window and a contact |
What to do next
Print tonight's in-house list after the business day closes, then run an unplug test at a quiet hour and write down what stops and who does what. That single exercise, one page and fifteen minutes, tells you more about your hotel's outage readiness than any vendor document. From there, build the three phases into a written procedure: prepare the connection, the paper and the roles; run the outage from one numbered log with one decision-maker; then reconcile every slip once and check the channels and room status before the next check-in wave.
Frequently asked questions
Can a cloud PMS work offline at all?
A browser-based system generally needs a connection for anything that writes; whether any local read-only copy exists is a product decision only your vendor can answer. Treat any offline mode as a bonus and plan paper as the fallback.
How long can a hotel keep running on paper?
There is no fixed limit, and the constraint is not paper but reconciliation: the longer the outage, the more slips, keys and channel bookings have to be matched afterwards, and the higher the chance of selling the same room twice. Decide in advance at what point you stop taking new reservations.
Is a second internet line enough for a hotel internet outage plan?
Only if it comes from a different provider and a different physical path, and only if someone tests it. Add a charged mobile router as a third option, put the router and printer on a small power supply, and check which devices fail to come back on their own.
What is the most common mistake after an outage?
Posting paper slips twice, or not at all, and trusting the screen for room status that housekeeping updated on paper. Post in slip-number order with the number in the reference field, then compare the paper log with the system before the next check-in wave.
Should we tell guests that the system is down?
No legal duty is stated here, but guests notice when the guest app, the offers and the television stop answering. A short written note at the desk and a working phone line reduce complaints and stop the front desk from explaining the same thing all evening.