Hotel technology

What an hour of system downtime costs a small hotel: a worksheet from your own numbers

Build a hotel system downtime cost worksheet from your own numbers. Count affected arrivals, manual work hours, delayed postings and goodwill to turn a bad afternoon into a defensible budget line.

  • 11 min read
  • AI Otel Editorial

AI draft · awaiting editorial review

Contents

Key takeaways

  1. A downtime worksheet has four lines: arrivals and departures inside the window, manual work hours including the recovery tail, postings delayed or lost, and goodwill given.
  2. Count only the arrivals due inside the outage window, not the day's bookings; an inflated worksheet loses the argument it was built for.
  3. Re-entering paper records after the system returns usually costs more hours than the outage itself, so time it with a stopwatch.
  4. Compare the annual loss from your own incident log with the annual cost of a second line or a fallback, and ask how many hours a year it must prevent to pay for itself.
  5. One page per incident, written within a week, turns a feeling into a budget line.

Downtime is any period when the systems the desk depends on stop working: the internet line, the property management system (PMS), the card terminal or the guest identity notification. Its cost is four lines you can count yourself: affected arrivals and departures, manual work hours, delayed or lost postings, and goodwill gestures. A hotel system downtime cost worksheet built from those four lines turns a bad afternoon into a number you can defend.

The front desk screen freezes at 16:40 on a Friday. Two coaches are due, the restaurant is taking dinner orders and the night audit will not wait. By 17:30 the system is back and everyone is relieved. Nobody writes down what the hour cost. After reading this you will be able to build the worksheet, hour by hour and line by line, from your own hotel's numbers, and compare the total with the price of doing something about it.

What stops when the system goes down, and what keeps running?

Usually one layer stops, not the whole hotel. The worksheet begins with a timeline of which layer failed and when, because each layer has a different cost and a different fix.

Separate the layers before you count anything: the internet line, the local network and Wi-Fi, the PMS itself, the card payment link, the guest identity notification that goes with check-in, electronic invoice issuance, the door lock encoder, and the channel sync that feeds online travel agencies. A failure in each of these produces different manual work, and the fix for one does nothing for another.

Distinguish a hard stop from a degraded hour. A system that is up but slow, or up but unable to print, still produces manual work and still belongs in the worksheet. Record the clock, not only the duration: the same hour costs more at 15:00 on a full Saturday than at 03:00 on a quiet Tuesday, and the worksheet should show which one you had.

LayerWhat stopsWhat keeps workingWho notices first
Internet lineEverything onlineCash, rooms, guests in houseFront desk
PMS applicationCheck-in, billing, reportsCard terminal, phonesFront desk
Card payment linkCard paymentsCash, all room recordsFront desk
Guest identity notificationKBS submission at check-inThe rest of check-inFront desk
Electronic invoice issuanceInvoices and e-documentsPayments, room recordsAccounting
Door lock encoderCutting new keysKeys already cutFront desk
Channel and OTA syncRate and availability updatesDirect bookingsRevenue manager

List what continues regardless: the building, the rooms, the guests already inside, cash, the paper registration form, and every key card already cut. Note who notices first, because that is where the first manual work starts: the front desk, then housekeeping, then accounting when postings fail to appear, then the guest.

Write the timeline to the minute. A forty-minute outage and a four-hour outage are different events and should never be averaged into one number. For what to do while the systems are down, keep a front desk kit for check-in, billing and keys on paper printed and within reach.

Which arrivals and departures does an outage actually touch?

Only the ones due inside the window. Counting the whole day's bookings is the most common way a downtime worksheet becomes a number nobody believes.

Take the arrivals due between the start and the end of the outage, not the arrivals for the day. A guest arriving at 22:00 was never affected by a 16:40 failure. Count the departures that could not be closed: a bill that cannot be closed holds the room, the room cannot be cleaned and resold, and the delay moves into housekeeping and the next arrival.

Count the in-house charges that happened during the window: restaurant covers, minibar, spa, laundry. These are the postings most likely to be lost, because the guest is already inside and nobody re-asks. Count walk-ins separately; a walk-in turned away is the only lost sale you can defend, because the guest was standing in front of you with a card.

Note group blocks and allotments whose pick-up was recorded on paper, since those numbers have to be reconciled later and the reconciliation is work. Record no-shows and cancellations that arrived during the window, and say plainly whether you could have converted them. If you cannot show it, leave it out of the total.

Keep a column for guests affected rather than rooms affected. A family in one room and a coach group are not the same unit of trouble.

How do you count the manual work hours?

Time the tasks, do not estimate them, and include the hours after the system returns. The recovery tail is usually longer than the outage and almost always left out.

Time one arrival written on paper, one in-house charge written on paper, one bill closed by hand and one key card made without the encoder. Four numbers give you the whole worksheet. Add the recovery tail: re-entering the paper records once the system is back, checking for duplicates, and the reconciliation the next morning. This is where the hours hide.

Count by role, because the loaded hourly cost differs: front desk, housekeeping, night audit, accounting, and a manager who stops doing their own work to stand at the desk. Use a loaded hourly cost from payroll, wage plus employer contributions and overheads, and confirm the figure with your accountant rather than guessing it.

For example, if a paper arrival takes twelve minutes instead of four, and eighteen arrivals fall inside the window, that is two hours twenty-four minutes before the recovery tail is added. Label it as an example and replace it with your own timing.

Include the manager hours spent on the phone with the internet provider or the software support line. That time is real and it is usually the most expensive hour in the incident. Do not count the same hour twice: if the front desk clerk is already paid for the shift, count only the extra work created, not the whole shift.

ChecklistTime these with a stopwatch during the next incident

  • One arrival written on paper
  • One walk-in registered on paper
  • One in-house charge written on paper
  • One bill closed by hand
  • One key card made without the encoder
  • The re-entry of one day of paper records after the system returns
  • The reconciliation the following morning

To understand what a vendor means when it promises a recovery time, read recovery time and recovery point explained for the front desk.

The two lines hotels forget: delayed postings and goodwill gestures

Postings that never reach the bill and gestures given to calm a guest are the two costs that disappear from memory by the next morning, so they must be written down at the moment they happen.

Split postings into recovered and written off. A minibar charge entered two days later is recovered; a restaurant charge nobody re-asks for is written off, and only the second one is a loss. Count the correction work as well: a charge entered twice, a bill reopened after night audit, a payment matched to the wrong room. Corrections are hours, and hours already have a line.

Give goodwill its own line and a value. A waived charge, a complimentary drink, a late checkout or an upgrade all have a menu price; state whether you valued them at menu price or at cost, and stay with one method. Record goodwill on the same paper as the incident, at the moment it is given, with the room number and the reason. Reconstructed goodwill is always understated.

Note what you cannot count: the guest who says nothing, pays and does not return. Put it in a note field, never in the total, because you cannot defend it. Keep the worksheet small enough to survive a challenge. A defensible figure of a few thousand lira is worth more in a budget meeting than an inflated one that falls apart on the second question.

Watch outWhat not to put in the total

Leave out lost revenue from rooms you cannot prove you would have sold, goodwill valued at menu price in one incident and at cost in the next, and any figure a manager remembers rather than a record shows. Each of these will be questioned, and each weakens the honest lines around it.

How does the total compare with the price of resilience?

Annualize the loss from your own log, then compare it with the annual cost of the option, and ask how many hours a year that option must prevent to pay for itself.

Build the annual figure from incidents per year multiplied by average hours multiplied by cost per hour. Use your own log, not a general assumption. Compare like with like: a recurring annual loss against a recurring annual cost, not one incident against one purchase price.

OptionWhat it protectsWhat it does not protectHow to price it
Second internet line, different mediumInternet outagesPMS and card terminal failuresAnnual line rental
Battery backup for router and desk devicesPower cuts at the deskLine and software failuresPurchase plus battery life
Local fallback for check-in and billingPMS application failureCard terminal, OTA syncVendor price or contract term
Written paper procedure, printed arrival listEvery layer, for a whileNothing technicalPrinting and one training hour
Support contract with stated response timeSlow vendor responseYour own line and powerAnnual contract fee

Map each option to the layers it protects. A second line does nothing for a PMS application failure; a fallback does nothing for a dead card terminal. Most hotels need two small options, not one large one.

Calculate the break-even in hours: if the option costs a fixed amount a year and your hour costs a known amount, the division gives the number of hours it must prevent. If your log shows fewer hours than that, the option is a preference, not a saving. Accept that resilience has its own cost and does not reach zero; the target is a shorter, calmer outage with a known procedure, not a promise that nothing will fail.

Ask the vendor for recovery time and recovery point in writing before you sign, and keep the answer next to the worksheet. Before signing anything, compare uptime and response time terms in hotel software contracts with what your own log shows you actually need.

How do you review an incident so the next count is easier?

Write one page per incident within a week, while the details are still fresh, and keep the pages in one place. The pattern across a year is worth more than any single event.

Separate cause from cost. The cause goes to the provider or the vendor as a support ticket; the cost goes to the budget and to the next contract negotiation. Write the page within a week; after that, the arrival count is reconstructed from memory and the goodwill line is gone.

Review the pages once a quarter and look for the pattern: the same hour of day, the same provider, the same layer. One repeated failure is a decision; ten different ones are weather. Feed the log into two decisions only: the resilience budget and the renewal of the support or connectivity contract. Keep the pages for the length of the contract and one renewal, so the next negotiation starts from your own record rather than from a feeling.

ChecklistFields for the one-page incident record

  • Date
  • Start and end time, to the minute
  • Systems affected
  • Arrivals and departures inside the window
  • Person-hours by role
  • Postings recovered and written off
  • Goodwill given and its value
  • Cause as understood at the time
  • What changed afterwards
  • Who reviewed it

Keep the record itself honest and unedited, including the incidents that turned out to be nothing. A log with only dramatic entries cannot be used to set a budget. When you evaluate a new system, the same record discipline applies in the early weeks: follow daily and weekly checks in the first 30 days after a PMS go-live, and rely on records you can check afterwards rather than on memory. If you want to see how a modern hotel operating system handles billing corrections and night audit, the same worksheet questions will tell you what to ask.

What to do next

Start a one-page incident log today, before the next outage, and time the next one with a stopwatch. The four lines, arrivals and departures inside the window, manual hours including the recovery tail, postings and goodwill, only become an argument for a second line or a fallback when they come from your own record rather than from a bad afternoon everyone remembers differently.

Frequently asked questions

Does moving to a cloud PMS remove downtime?

It moves the failure points rather than removing them. The internet line, the local network, the card terminal and the identity notification stay yours, and a provider's outage is now also your outage. Ask for recovery time and recovery point in writing and keep a paper procedure either way.

Should I include lost room revenue in the worksheet?

Only where you can show a real lost sale, which in practice means a walk-in turned away. Everything else belongs in a note field. A total built on rooms you might have sold will not survive the first question in a budget meeting.

What hourly cost should I use for staff time?

A loaded hourly cost: wage plus employer contributions and overheads, taken from payroll rather than from the gross wage. Ask your accountant to confirm the figure once, then use the same figure for every incident so the months are comparable. Without a consistent figure, the hotel system downtime cost worksheet cannot be compared across incidents.

How long does it take to build the first worksheet?

The first one is always an estimate, built after the fact from the day's arrivals list and the memory of the desk. The second one is real, because it is timed. That is why the log matters more than the first total.

Is a second internet line always worth the cost?

It depends on your own log. Divide the annual cost of the line by your cost per downtime hour, and compare the result with the hours your log shows. If the log shows fewer hours, the line is a preference rather than a saving, which is a legitimate decision as long as it is made knowingly.

AI Otel EditorialEditorial team

Drafted with AI assistance from AI Otel’s published fact base. Not yet reviewed by an editor.

We can brief you on this topic.

If this topic touches your work, we can walk you through what is built and what is designed.

Request a briefing