How the Booking.com review score works, and how to reply to reviews

Booking.com scores a property out of 10, and that number is not a star rating a guest chose. It is derived from a set of separate category questions answered after checkout, averaged across the reviews currently counted for the property. That single design decision explains almost everything owners find confusing about it: why the score barely moves after a great month, why it can drop without a bad review appearing, and why two properties with identical written feedback can sit half a point apart.

How the score is built

A guest who completes a review answers several category questions and writes optional free text. The categories, not the prose, produce the number. Booking.com explains the mechanics on its own partner help centre, and it is worth reading the current version rather than trusting any summary, including this one, because the weighting has been revised before.

Three consequences follow from it being an average of categories:

1. The written review and the score are only loosely related. A guest can write two warm paragraphs and still score a 7 because they marked value for money low.
2. Your worst category sets your ceiling. Five strong categories and one weak one average out to mediocre.
3. Volume dampens everything. On a property with hundreds of counted reviews, one perfect stay is almost invisible. This is protective far more often than it is frustrating.

The sub-scores, and which one to watch

Two of those rows deserve attention because they behave unlike the rest.

Location cannot be changed and should be read as a constant that sits inside your average. If it is low, every other category has to carry more, and it is worth knowing that permanently rather than rediscovering it each quarter.

Cleanliness is the category most responsive to effort, and the bathroom is most of it. It is also the category most likely to produce a written complaint, which is why it shows up disproportionately in aspect level analysis of hospitality reviews. There is more on reading those patterns in guest sentiment analysis.

Value for money, the one that drags

Value for money is the category most often responsible for a score that will not move, and it is the one operations cannot fix, because it is set by pricing rather than by anything the team does on the day.

A property that raises rates into a strong season without changing anything about the experience will see this category fall, and the headline score with it, while every operational category holds steady. The reviews will read perfectly well. The number goes down anyway.

Why the number moves when nothing changed

• Old reviews stop counting. The score reflects a moving window, so a strong review ageing out lowers the average without anything new going wrong.
• Mix changes. A season that brings a different kind of guest brings different category expectations, particularly on value and comfort.
• Small denominators. A property with forty counted reviews will swing on single reviews in a way one with four hundred never does.

None of these are things to react to. The useful signal is a sustained move in one category, not a decimal point on the headline.

Replying to Booking.com reviews

Property responses are written in the Extranet, under the reviews section, and they are moderated before publication. That moderation delay is the practical difference from Google, where a reply appears immediately. Write it, save it, and do not assume it is live.

Two things behave differently here compared with Google:

• Reviews come only from real bookings. Booking.com invites the guest after a completed stay, so the invented review problem that dominates Google is largely absent. If a review is unfair it is still almost certainly from somebody who was there.
• The audience is further down the funnel. Somebody reading your Booking.com reviews is on a booking page with a price in front of them, which makes a calm, specific reply to a complaint worth more here than almost anywhere else.

The reply shape is the same one that works everywhere: neutral greeting, specific acknowledgement, what changed, close. Never the guest’s name. The reply template library has the wording for each rating band.

The thresholds that matter

The score is displayed with a word attached, and the word is what a guest actually registers while scanning. Crossing a threshold changes the label even when the movement is a rounding artefact, which makes the boundaries worth knowing.

In practice: 8.0 is the point below which the number starts working against you, 8.5 is where a property reads as genuinely good, and 9.0 and above is where the score becomes a reason to book rather than merely not a reason to leave. Below 7.0, the score is the first thing about your listing that a guest notices, and nothing else on the page is being read properly until it improves.

For how Booking.com sits against the other platforms, and where the effort is best spent for your property type, see which review platform actually matters.