Is This Complaint New or Chronic? Reading What Guests Actually Said
One theme score can describe a long-running problem or a sudden drop, and the number alone won't tell you which. This article gives three questions to ask about any guest complaint: its direction, the evidence behind it, and whether it follows the calendar. A worked example from one anonymized hotel shows how the reading changes what you ask next.

SENTEEZ
Guest Experience Insights

Key findings at a glance
- A score is a snapshot. The same 40% positive can mean a year-long plateau or a fall from 70% three months ago.
- Three questions. Ask about direction, evidence and the calendar before you act on a complaint theme.
- Thin months need care. Months with fewer than 10 mentions are flagged as low-confidence, and themes under 50 mentions are excluded from revenue scoring.
- Worked example. Maintenance & Upkeep sits at 2% positive from 52 mentions, against an 85% healthy benchmark, a gap of 83 points. Flat and low is the chronic shape, and the largest group (22 mentions) is about climate controls guests can't adjust.
- Direction isn't cause. Reviews tell you where to look, and the property tells you what changed.
Two hotels, one score, two different problems
A theme score of 40% positive can describe two very different situations. In one hotel, the score has sat at 40% for a year. In the other, it stood at 70% three months ago and has just fallen. (Both hotels are an illustration, not real properties.)

The number is identical, but the right response isn't. A long plateau points to something built into the property or its routines, and it usually calls for a considered plan. A recent drop points to something that changed, so the first move is a question: what happened in the last few months?
A single percentage can't tell these two apart. That's why reading a theme should come before acting on it.
This article walks through three questions to ask about any guest complaint theme:
- Is it getting worse, getting better or staying flat?
- Is there enough evidence behind that reading?
- Does it follow the calendar?

From there, we look at what guests actually describe, using a worked example from one anonymized hotel analyzed by SENTEEZ. Everything here comes from publicly available reviews.
Why a score alone cannot tell you
A theme score is a snapshot. It's the share of Review Mentions about one theme, such as Maintenance & Upkeep, that guests expressed positively. A Review Mention isn't a unique review. One guest review can touch several themes, so the count reflects mentions, not guests.
Like every average, a snapshot can't show direction. We often say that a strong average is a hiding place, not a finding, because it can cover a weak hotel. The reverse is just as true for a weak average. A low score can hide a slow recovery, a steady decline or a flat line, and each of those asks for a different next step.
Two other things sit behind every score, and they're easy to miss. The first is volume: a score built on 400 mentions and one built on 40 aren't equally solid. The second is content: two hotels can share a score while their guests describe entirely different problems.
The Explain It layer exists to open up that single number. For any theme, it shows three things: how sentiment has moved month by month, how much evidence sits behind that movement, and what guests are actually describing. The next sections take these in turn.
Question one: is it getting worse, better or flat?
Start by reading direction. The monthly sentiment trend plots the positive and negative share of mentions for one theme, month by month, so you can see how the picture has moved instead of only where it stands today.
Three shapes come up often, and each one prompts a question rather than a conclusion.
A sudden drop. Sentiment falls sharply in one or two months after a stable stretch. The natural question is what changed: a renovation, a staffing change, a system update, a new supplier, or a shift in the kind of guest staying. Reviews show that something changed. They don't say what.
A slow slide. Sentiment erodes gradually over many months. This pattern can point to wear, aging or deferred attention building up, although reviews alone can't confirm the cause. The question is how long the slide has been under way and whether anything was done along the way.
A long plateau. Sentiment stays at roughly the same level throughout the window. This is the chronic case. The complaint isn't an incident but part of how the property is experienced. The questions turn to what's built into the asset or the routine, and what it would take to change it.
There's a fourth reading, and it's improvement. A steady rise after a low point can be the first sign that something is working, so it's worth recording before it's forgotten.
One caution applies to all four. A trend is a record of what guests said in past months. It isn't a forecast of what they'll say next quarter, and it should never be presented as one.
Question two: is there enough evidence?
A trend line looks authoritative whatever sits beneath it. A month with 60 mentions and a month with four are drawn with the same confidence unless something tells the reader otherwise.
Small months are noisy. When four guests mention a theme in a month, one more positive or negative comment moves that month's score by 25 points. That swing says more about sample size than about the hotel.
For this reason, the Explain It view marks low-confidence months, which it defines as months with fewer than 10 mentions. It's the monthly version of a rule we apply across the platform: aspects with too few mentions, or too few hotels reporting them, are flagged and kept out of conclusions instead of being presented with false confidence.
A few practical reading rules follow from this:
- Read the shape across the whole window before reading any single month.
- Give less weight to flagged months, and say so when presenting them.
- Treat the total number of mentions behind the theme as the real measure of how much can be concluded.
- When almost every month is flagged, say that the monthly movement is too thin to interpret, and rely on the overall level and the consistency of direction instead.
That last point matters. Honest analysis sometimes ends with “the monthly detail is too thin to read,” and a trend chart that admits it is more useful than one that hides it.
Question three: does it follow the calendar?
Some complaints rise and fall with the seasons. Cooling is the obvious example: complaints about rooms that are too warm are more likely in hot months, and heating complaints in cold ones. A dip that looks new may simply be the same dip coming back.
A year of monthly data lets you ask the question. Look at whether the weak months cluster in one season and, if the data allows, whether the same months looked weak a year earlier. If they do, the better question is how the property handles that season, not what went wrong this quarter.
Use cautious wording. Twelve months can suggest a seasonal pattern but can't prove one, because each season appears only once. The phrase to use is “check whether it tracks the calendar,” not “this is seasonal.”
It also helps to know what the chart doesn't show. The Explain It trend plots the share of positive and negative mentions, not how many mentions fall in each month. A seasonal pattern might show up in the volume of complaints without any movement in the percentage.
What guests actually describe: the sub-theme breakdown
Direction and evidence tell you how a theme is moving. The sub-theme breakdown tells you what it's made of. Two hotels at the same score can have guests describing quite different things, and the action that suits one won't suit the other.
The hotel data on our public platform shows how much a single score can contain. In the markets analyzed so far, Maintenance & Upkeep sits at similar levels in several places: Interlaken scores 24.1% positive across the 19 hotels that report it, Lucerne 24.7% across 22, Lausanne 25.2% across 8, and St Ives in the UK 25.6% across 7. The scores are close. What guests describe isn't identical.

In Lucerne, the recurring patterns center on cooling, with aging fixtures, plumbing and in-room electrical items alongside. In Interlaken, climate control also appears, joined by bathroom and plumbing faults, building works and slow repairs. These are recurring patterns drawn from samples across hotels, not counts, and they're signals rather than proof of cause.
These markets differ in type, hotel mix and traveler base, so the point isn't to rank them. The point is that a similar score can sit on top of different guest experiences, and only the breakdown shows which. The regional figures are also separate from the property example below and aren't directly comparable with it.
A worked example: one theme, three questions
The example below comes from one anonymized hotel analyzed by SENTEEZ using public reviews. Its name, location and size are withheld. The theme is Maintenance & Upkeep.
The snapshot. Maintenance & Upkeep stands at 2% positive from 52 mentions, against SENTEEZ's 85% healthy benchmark, a gap of 83 points. On score alone, this looks like an emergency. The trend and the breakdown say more.
Question one: direction. The negative share sits at or near 100% in most months of the window. It eases to the high 80s in three months and to roughly two-thirds in May 2026, then returns to the top by June. The positive share is at zero throughout, apart from one small bump in February 2026. Nothing is rising or falling in a sustained way. The reading is flat and low, which is the chronic shape.
Question two: evidence. Every plotted month is flagged as low-confidence, with fewer than 10 mentions. Across 11 plotted months, 52 mentions average under five a month, and the total sits only just above the 50-mention floor used for revenue scoring. That changes how the chart should be read. The May dip and the February bump each rest on a handful of comments, so neither should be presented as recovery. What can be said, with appropriate modesty, is that across the whole window the guests who mentioned this theme were almost uniformly negative.
Question three: calendar. The window runs from July 2025 to June 2026, with no point plotted for April 2026. The negative share stays high in warmer and cooler months alike, so nothing in the chart visibly follows the seasons. With so few mentions in each month, though, a seasonal pattern can't be ruled out, and the chart shows shares, not volumes.

What guests describe. The breakdown splits the theme into groups. In paraphrase:
| What guests describes | Mentions | Positive |
|---|---|---|
| Room climate controls that are centrally restricted and unresponsive, leaving rooms too warm | 22 | 0% |
| Visible wear in rooms, such as tired bedding, worn finishes and damaged fixtures | 12 | 0% |
| Televisions that do not tune or work properly, and poor signal | 9 | 0% |
| Leaking air conditioning units, lift breakdowns and water supply interruptions | 5 | 0% |
| Prompt, courteous problem solving by individual maintenance staff | 2 | 100% |
These groups account for 50 of the 52 mentions. The largest, at 22 mentions, is about climate controls that guests can't adjust. That's a different kind of problem from worn furnishings or an unreliable lift, even though all of them sit under the same 2%. The strength is small, only two mentions, but its direction is clear: guests praise the people who fix things while criticizing what they're asked to fix. With so few mentions, treat it as a signal worth watching, not a finding. For each group, the product also suggests an action, an owner role and a priority.
What the reading changes. For this theme, the useful question isn't “what went wrong last quarter?” The trend shows nothing recent that would explain it, because the complaint is persistent. The question becomes which groups can be addressed through settings and routines, and which need a planned program. That sorting is the job of the layers described below. The reading doesn't say how long any fix will take, what it will cost or what it will do to guest scores.
What the trend cannot tell you
A trend shows direction. It doesn't show cause. If sentiment on a theme fell in a given quarter, the reviews can't say whether a renovation, a staffing change, a pricing shift or a change in guest mix was responsible. Only the people who run the property know what changed.
We keep coming back to one line, and it's worth saying plainly:
Reviews tell you where to look, and the property tells you what changed.
Two further limits are worth stating. First, a trend describes what guests said, not what happens inside the hotel. It's a signal, not proof. Second, opportunity isn't attribution. If a theme improves after an action, the data can't prove that the action was responsible, and a trend should never be used to promise an outcome.
Where this fits in the workflow
In the Revenue Opportunity workflow, Size It shows where the largest gaps against a healthy benchmark sit. Every theme row there opens Explain It, which is where the reading in this article happens: the monthly trend, the evidence behind it and the sub-theme breakdown.
Explain It informs judgment. It doesn't feed the Priority Score, which Prioritize It calculates separately from each theme's score lift, its controllability and the effort involved. Plan It then turns that ranking into a sequenced plan with owners and 90-day checkpoints. What you read in Explain It lets a team ask better questions of the ranking, such as whether a high-priority theme is a recent problem or a long-standing one.
Track It, which looks at whether improvements hold over time, is in development.
Frequently Asked Questions
Key answers regarding hospitality sentiment diligence and data methodology.
There's no fixed number. Twelve months lets you look for a seasonal pattern, but the more useful measure is how many mentions sit behind each month. The Explain It view flags months with fewer than 10 mentions as low-confidence.
Not at all. Chronic means persistent, not unfixable. It tells you the complaint is part of how the property is experienced, which shifts the question from “what changed?” to “what's built in?”
No. A trend describes what guests said in past months. It isn't a forecast.
Themes with fewer than 50 mentions are shown but excluded from revenue scoring, and any reading of them should be described as preliminary.
See what your guests are really saying about your hotel.
A free, analyst-built snapshot of your guest sentiment and how you compare with one competing hotel, delivered within 48 hours.
- Your hotel's rating, with a clear summary
- One-competitor benchmark: where you win and lose
- Your top recurring guest themes
- Delivered within 48 hours
- Free, with no commitment

Written by SENTEEZ
Guest Experience Insights
Related Insights

Same Report, Two Readers: What a General Manager and a Hotel Owner Each Need
A general manager and a hotel owner can open the same guest-experience report and ask very different questions of it. This article shows how each reader should approach the report, where their views meet, and what it can't tell either of them. The thread running through it: the biggest opportunity and the first move aren't the same thing.


What a Stronger Review Score Could Be Worth: A Modeled Example
This article walks through one anonymized hotel, from guest feedback gaps to a hotel-level revenue range. It shows which inputs are real and which are illustrative, how the range splits across four broad pillars, and why the result is a ceiling, not a forecast. It also states where the model falls short, and closes the series by laying out what a modeled example can and can’t support.


What Guest Reviews Reveal About a Hotel’s Deferred CapEx Before You Buy
Guest reviews can’t show a hotel’s capital plan, but recurring Maintenance & Upkeep complaints are an early signal of postponed spending. Three reads sort that signal: the town baseline, the town ceiling, and whether the weakness travels alone. In three Swiss towns upkeep is positive in only about a quarter of Review Mentions, so the market comes first. The article ends with three hypothetical targets, the disclosures worth requesting, and the limits of the method.
