Every year, hotels renew software contracts they quietly hate. The reason is almost always the same: switching costs feel larger than the ongoing pain. So the pain continues, and nobody adds it up.
I’ve been trying to add it up.
The visible costs vs. the invisible ones
The visible cost of a hotel tech stack is easy to find it’s on the invoice. PMS licence, channel manager subscription, booking engine fee, maybe a revenue management tool. Most mid-market hotels spend between $15,000 and $60,000 per year on software licences, depending on property size and configuration.
The invisible costs are larger. They show up in:
Staff time absorbed by workarounds. When software doesn’t do what staff need, staff find workarounds. A front desk agent who manually copies reservation data from one system to another is spending time that costs money. If that agent spends 45 minutes per shift on workarounds across a 200-room property with 6 front desk agents per day, that’s 4.5 hours of paid labour lost to bad software every single day.
Error recovery. Bad integrations produce bad data. A rate plan mismatch that overbills a guest requires a supervisor to intervene, a refund to be processed, and a follow-up email to be sent. The guest still leaves unhappy. The cost of that incident staff time, payment processing, and the probability of a negative review is not zero.
Guest friction. A digital check-in that times out because the keylock API didn’t respond is not just a technology failure. It’s a guest standing at their door at 11pm with luggage, frustrated, calling the front desk. The front desk agent drops whatever they were doing. The guest’s first impression of the property is a broken door. That experience affects the review score, which affects booking volume, which affects revenue.
Missed revenue. A room upsell system that doesn’t trigger because the PMS didn’t pass the right room type code is missed revenue. Every night. Quietly. Never visible on any report.
The denominator problem
The reason these costs stay invisible is that hotels don’t have a denominator to measure them against. They know the licence fee. They don’t know the cost of the workarounds, because those costs are absorbed into salaries that get paid regardless.
This is how you end up renewing a contract for software you hate: the licence fee is concrete, the alternatives require capital, and the ongoing cost of the current system is diffuse enough to be invisible on any single line item.
What good looks like in practice
The best-run hotel technology stacks I’ve studied share one characteristic: they minimise the surface area of manual intervention. Reservations flow from the booking engine to the PMS without a human touch. Room status updates from housekeeping reach the front desk in real time. Guest requests go directly to the relevant department without a phone call.
This sounds obvious. It is not common.
The delta between a hotel where the technology works and one where it doesn’t shows up most clearly in staff behaviour. In a property with good technology, the front desk agent is present with the guest. In a property with bad technology, the front desk agent is present with the screen.
The switching cost fallacy
The argument against changing systems is always switching costs: the time to migrate data, the time to retrain staff, the risk of downtime during transition.
These are real costs. But they are one-time costs. The cost of staying on a bad system is perpetual.
The calculation that hotels rarely do: if the current system costs $X in hidden operational expense per year, and switching costs $Y in one-time disruption, the payback period for switching is Y divided by X. For many hotels, that number is under eighteen months.
The reason the switch doesn’t happen is not that the economics are wrong. It’s that the hidden costs are never measured, so the denominator in the calculation is always zero.
What I’d build if I had the data
The tool I want to build and can’t yet, because I don’t have access to real operational data is a hidden cost calculator for hotel technology. A simple model that takes: number of staff, average shift length, estimated workaround time per shift, error rate, average cost of error recovery, upsell miss rate. The output would be an annual number that sits next to the licence fee on the same invoice.
I suspect that for most mid-market properties, it would make switching look very cheap by comparison.