Why We Won't Put a Price on a Fix From Review Data

“What will it cost, and when do we get it back?”

Every plan eventually reaches this moment.

The guest data has been sorted. The priorities are clear. There's an action plan with owners and timelines. Then the owner leans forward and asks the question that matters most to anyone signing off on capital.

What will it cost? And when do we get it back?

It's the right question. It's probably the most important question in the room.
And it's the one review data can't answer.

We get asked for this number a lot. Operators and investors who reviewed our model asked for cost to capture, payback periods, and NOI and EBITDA impact.

We understand why. So instead of quietly leaving it out, this article explains why we won't produce it, and what does.

Why do hotels want a cost figure from guest data?

Because capital is scarce, and payback is how owners compare one project against another.

A refurbishment competes with a new spa, a systems upgrade and a dozen other claims on the same budget. A number makes those choices comparable. Without one, the conversation drifts back to opinion.

So when a tool shows a large guest-experience opportunity, the natural next step is to ask what it costs to capture.

The instinct is sound. The problem is the source.

What can review data actually tell you?

Here's the clearest way we can put it:

Guest reviews show where guests feel friction. They don't show why the friction exists, or what it will cost to remove.

That's the core of the hotel review data limits. Reviews are written by guests, and guests describe their experience, not the building's engineering.

A guest writes, “the room was too warm and the air conditioning barely worked.” That's real, useful evidence. But it's evidence of a symptom.

Why does the same complaint lead to very different bills?

Take a real example from our reference property, an anonymized 156-room hotel in the Lake Geneva region.

Its Maintenance & Upkeep theme has 268 guest mentions, and only 49% of them are positive. The friction guests describe includes unstable room cooling, slow elevators and wear on fixtures.

Now focus on cooling alone. A recurring cooling complaint could mean any of these:

A faulty unit. A repair or a replacement part. Small, fast, operating budget.

Aging building fabric. Insulation, glazing or envelope issues that make rooms hard to cool however good the equipment is. Larger, slower and probably capital.

Plant sized for a different use. A system designed for how the rooms were used years ago, not how they're used now. Potentially a major project.

Same complaint. Same reviews. Three very different bills.

Reviews can't tell you which one you're dealing with. A technician can. Until someone has diagnosed the cause, nobody can honestly price the fix.

What goes wrong when someone guesses the number anyway?

A guessed number doesn't stay a guess for long.

It goes into a slide. The slide goes into a board pack. Someone quotes it in a budget meeting. Within a month it's “the number,” and nobody remembers it was built on an assumption about a cause no one had checked.

That's anchoring, and in capital planning it's expensive. A precise-looking cost figure feels more trustworthy than a range of possibilities, even when it's less accurate.

The worst outcome isn't having no number. It's an owner allocating capital on a guess presented as analysis.

So we draw a firm line.

Any cost figure generated from review data alone would be a guess. We'd rather not hand you one.

But you put a number on revenue. Isn't that the same thing?

It's a fair challenge, and it deserves a straight answer.

Yes, we estimate revenue opportunity. For our reference property, it's a range of CHF 1.13M to 1.25M a year. But that figure is a different kind of estimate, built in a different way.

It sits at hotel level, split only by broad pillar, never by individual fix.

It's a ceiling, not a forecast: what might be available if every gap closed at once.

It rests on published hospitality research linking online reputation to rate, occupancy and RevPAR. That relationship has been studied across many hotels.

It describes guest behavior, which reviews do capture.

Cost is different. It depends on the building: the age of the plant, the state of the fabric, the scope of the work. There's no research coefficient that turns “guests say the room is too warm” into a repair bill, because the bill depends on things no guest can see.

So the two numbers aren't the same kind of claim. One estimates how much opportunity exists. The other would claim to know what a specific fix costs. Reviews can support the first. They can't support the second.

What do we use instead of cost?

Instead of pretending to cost the work, we use three things that review data can support honestly.

Effort. Every theme is rated Low, Medium or High. Effort sequences the work. It doesn't price it. A Low-effort fix can start this month. A High-effort one needs planning and approval.

Budget source. Each initiative is routed to the pot that would pay for it: operating budget, pricing and positioning, or capital. That tells you who needs to approve it, not how much it costs.

Horizon. Work is grouped into Quick Wins (0 to 30 days), Programs (30 to 90 days) and a Capital Plan (90 to 180+ days), so slow capital decisions don't hold up fast operational ones.

This is how guest feedback becomes an action plan without inventing numbers. In our reference plan, the phased bathroom and room refurbishment sits in the Capital Plan for exactly this reason. It needs diagnosis, a priced scope and approval before anything moves.

The plan says what to do and in what order. It doesn't pretend to know the invoice.

What does hotel capex payback actually require?

Real payback needs four things.

1. A diagnosed cause. Someone has confirmed what's actually wrong.
2. A priced scope. An engineer, contractor or condition survey has costed the work.
3. A baseline. You know where performance stood before the fix.
4. An observed result. You've measured what changed afterward.

Guest reviews contribute to two of these. They give the first signal that something needs attention. And later, they help show whether guests noticed the difference.

The diagnosis and the price come from the building and the people who maintain it. Any payback figure that skips those steps isn't a payback figure. It's a hope with a decimal point.

So where does costing become real?

Inside an engagement with the hotel.

When a property's own engineering team, or a condition survey, sits alongside the sentiment data, the picture changes. The guest data shows where friction is concentrated. The engineering input explains why, and what it will take. Together, they can support a real cost estimate.

That's a workflow built with a client, not a generic calculator applied from the outside. It takes longer. It's also the only version we'd trust.

The other half of payback, measuring what improved after a fix, is the layer we're building next.

In short: what review data can and can't do

Review data can:

• show where guests experience friction.
• show how often each issue comes up, and how sentiment is trending.
• help rank fixes by controllability and effort.
• support a hotel-level revenue opportunity range
  

Review data can't:

• identify the root cause of a complaint.
• price a repair or renovation.
• calculate payback, ROI, NOI or EBITDA impact on its own.

Knowing the difference is what keeps a useful tool from becoming a misleading one.

Before your next capex review

For any project justified by guest feedback, ask four questions.

1. Has someone diagnosed the cause, or are we assuming it from complaints.
2. Who priced the scope, and what did they inspect?
3. What's our baseline today?
4. How will we know, in three or six months, whether it worked?

If the cost came from the complaints alone, send it back.

The bottom line

Reviews tell you where to look. Engineers tell you what it costs.

Any tool that claims to do both from review text alone is guessing, and it's guessing with your capital.

We'd rather show you exactly where the problem sits, rank what to do first, and leave the price to the people who can actually see the building.