A single number at the top of a listing is the quickest way to judge a hotel, and also one of the easiest to misread. A score of 4.4 sounds precise. In practice it comes from a particular scale, a particular group of reviewers and a particular formula, none of which you can see at a glance. This article explains how numeric rating scores generally work, why they tend to bunch toward the top, and how to use them next to the written comments. It does not rank or recommend any hotel.
What a rating score usually is
In most systems, guests are asked to give a score, often as stars from one to five, or as a number out of ten or some other scale, and the platform averages them. Some platforms also ask for sub-scores such as cleanliness, location, staff and value, then combine those.
Details vary and are often not fully published. Platforms may:
- Weight recent reviews more heavily than old ones.
- Exclude reviews they judge to be unreliable.
- Round scores in particular ways.
- Show only a rating, or show a rating plus a breakdown.
- Count only verified stays, or accept anyone.
So the same hotel can show slightly different scores on different sites, and you should not treat scores from two platforms as directly comparable.
Scales are not the same
A score out of five and a score out of ten are different scales. As a simple conversion, 8.4 out of 10 corresponds to 4.2 out of 5, because 8.4 divided by 2 is 4.2. But that is only arithmetic. Sites with different scales may also have different habits among their reviewers, and different ideas of what counts as “average,” so even converted numbers do not always line up.
Why scores cluster high
If you browse widely, you may notice that most hotels show scores in the upper part of the scale. There are several general reasons this happens:
- Most stays are fine. Many guests have an acceptable experience and rate it positively.
- Self-selection. People who are happy, or who are prompted by a friendly request, are often more willing to rate.
- Social norms. Some reviewers treat anything below the top as a harsh judgment, so “good” becomes five stars and “bad” becomes three.
- Rating requests. Hotels or platforms that invite reviews may reach satisfied guests more often.
- Filtering. Properties with very poor scores may be less visible or may close.
The effect is that the useful range of a scale is narrower than it looks. The difference between 4.2 and 4.5 may reflect a few disappointing stays, while a score much lower than that is often a stronger signal. How much weight small gaps deserve is uncertain, so it is wise not to treat a tenth of a point as proof of anything.
Sample size changes everything
A score from a handful of reviews is fragile. A score from hundreds is steadier. Here is a worked example. Suppose a new listing has five reviews that total 24 points, an average of 4.8. If a sixth guest gives 2 stars, the total becomes 26 across six reviews, which is about 4.33. One review moved the score by nearly half a point.
Now suppose a different hotel has 200 reviews averaging 4.5, so they total 900 points. The same 2-star review brings the total to 902 across 201 reviews, which is about 4.49, essentially unchanged after rounding. Same review, very different effect.
The lesson: always check how many reviews sit behind a score. A very high score with very few reviews says little. A moderate score with many reviews may be more dependable.
What an average hides
Averages hide the shape of the data. Two hotels can share a score but offer very different experiences: one with a few excellent and a few poor stays, another that is consistently middling. If a platform shows a breakdown by star level, look at it. A visible group of low scores, even if small, tells you what the bad outcomes look like, and reading those comments tells you whether they would bother you. Our guide to negative reviews offers a calm method for doing that.
Sub-scores can also help. A hotel might score well on location and poorly on cleanliness. If cleanliness matters most to you, the overall number may be less useful than the specific one.
Read the number next to the text
The text answers the questions the number cannot:
- What exactly did people like or dislike?
- Are the complaints recent or old?
- Do the reviewers resemble you?
A rule of thumb is to use the score to filter and the text to decide. Scan the numbers to remove clear mismatches, then read recent reviews of the remaining candidates for repeated themes. Our article on review patterns shows how to do that tally.
Where the score and the text disagree, such as a high score with detailed recent complaints, give more weight to the text. If you suspect the numbers are being pushed up artificially, our piece on fake reviews explains general warning signs.
Compare like with like
When you compare hotels, try to compare within one platform, at a similar review count and with similar types of travelers. A score for a downtown business hotel and a score for a beach resort come from different kinds of guests with different expectations, and the numbers may not mean the same thing. This is particularly true when you compare across a travel style that is different from your own, since a family-focused stay may be rated on different priorities than a quiet couples’ stay.
A short checklist for reading scores
- How many reviews are behind the number?
- What is the scale, and how are recent reviews treated?
- What does the spread look like?
- Which sub-scores match your priorities?
- What do the most recent written comments say?
No score can promise how a particular stay will go. Used as a filter rather than a verdict, though, it is a helpful first step.