Platform guides

Booking.com reviews: how the score is calculated and what you can reuse

6 October 2026 · JustReview

Booking.com reviews come only from guests who booked a stay through Booking.com, and each one carries a single required overall score from 1 to 10 plus optional text and category ratings. The big number on a property page averages those overall scores across the last three years. Since January 2025 that average has been weighted towards recent stays, and any review older than 36 months is archived and no longer counts. The category scores for cleanliness, staff or location never enter the headline score at all.

That last rule is the one with practical weight for anyone who shows Booking reviews outside Booking. JustReview imports Booking.com reviews into widgets for hotels, guesthouses and apartment owners, and it does not archive anything. On 6 October 2026 we opened the Booking pages of four of those properties in a headless browser and set what Booking displays beside the reviews we hold. The review counts lined up with our last 36 months almost exactly, and more than half of what we had stored for those four was already gone from Booking.

How Booking.com reviews work

Everything in this table comes from Booking’s own guide to Guest Review Scores for partners and its review guidelines for guests, both read on 6 October 2026.

Two rows explain most of the questions owners ask. The score uses one number per review and nothing else, so a guest who marks cleanliness down but still gives a 9 overall moves the score as a 9. And the window moves every day, so the score changes even in a month with no stays.

How the Booking.com review score is calculated

Until January 2025 the rule was simple: the score was the average of all overall scores from the past three years. Booking’s partner guide now says that “every new guest review score you receive will be more important than the existing ones” and that “as a review gets older, its importance decreases”. It adds that for properties with few reviews it limits large swings “by giving the same weight to the most recent reviews”. It does not say how fast the weight falls, so nobody outside Booking can recompute the number exactly.

What you can do is test how far a plain average gets you. We took every property in our data whose Booking import is current and whose page we could open, four in total, and compared three numbers: the score Booking shows, the plain average of all the reviews we hold, and the plain average of only those from the last 36 months.

The count column is the cleanest confirmation of the archiving rule we have seen. Booking’s number sat within six reviews of our 36-month count on all four, while our totals were up to three times as large. We did not reconcile the small surplus on Booking’s side review by review; a few days of new arrivals or a different dating rule would account for it.

The score column tells a split story. On A, B and C the plain three-year average, rounded to one decimal, is exactly what Booking displays, so at that precision the recency weighting left no trace. The all-time average misses on all four, by up to a quarter of a point. Property D is the exception, and it is the small one. Its 14 scores from the window average 9.21, yet Booking shows 9.7. Its four most recent reviews are all 10s and its only weak score, a 5, dates from July 2024. A plain average cannot produce 9.7 from those 14 scores. Only a weighting that leans heavily on the recent 10s gets close. For a property with a handful of reviews a year, recency weighting is worth half a point. For one with a few hundred it barely registers.

Over half the reviews we held were already archived

Here is the part that matters once reviews leave Booking. Of the reviews we held for those four properties, 857 of 1,639, or 52.3%, were older than 36 months. Booking had archived them and no longer counted them. Per property the share ran from 40.7% for the apartments to 68.2% for the villa. Across all 3,277 Booking.com reviews in our database, 1,432 (43.7%) are past the 36-month line.

So a widget fed from Booking is not a mirror of the Booking page. It is the Booking page plus an archive that Booking itself has put away. Whether that is a problem depends on what you want the widget to do:

• For the score, Booking’s window is the honest one. A guest who clicks through to Booking will see the three-year figure, so the number on your site should not claim more.
• For the text, older reviews can still be useful. A three-year-old review that praises your breakfast is still true if the breakfast has not changed. One that complains about a renovation you finished in 2023 is not, and you should hide it.
• Look at the dates either way. On the villa, two out of three reviews we held were older than anything on its Booking page. Ordering by newest first keeps the cards on screen within Booking’s window. Reviews from 2021 sitting next to a “this week” price do not help.

If the number on your site and the number on Booking disagree for other reasons, such as moderation or syncing delays, why review counts do not match covers the usual causes.

Why a 9 on Booking is not a 5 on Google

The 1 to 10 scale changes how guests score, and it shows up the moment you put Booking next to a five-star platform. This is how the 3,277 Booking scores we hold spread, and what each becomes on a five-star display:

Four reviews in ten are a 9 or an 8. On the pages we checked, Booking labelled 9.1 and 9.3 as Superb, 9.7 and 9.8 as Exceptional, and 8.9 as Fabulous. On a five-star display an 8 becomes four stars, which is how a merely average stay often reads on Google.

The clearest view comes from properties that sit on both platforms. Two in our data have a large enough sample on each, and over the same 27 months, from October 2023 to early January 2026, they compare like this:

Same building, same guests, same period. The averages differ by less than a tenth of a star, yet the share of top marks drops by 19 to 26 points on Booking. On hotel E, Google also has more than twice the share of low scores. Booking compresses opinions into the 8 to 10 range, while Google pushes them to the ends. Neither is wrong. They are different instruments, and a visitor who sees both on your page should be able to tell which is which. Our broader look at what a star rating distribution tells buyers explains why a perfect row of fives persuades less than it seems to.

What you can reuse on your own website

Booking reviews are public, and a widget can show them with a link back to your property page. What to carry over, and how:

1. The score, with its date. Show the three-year figure as Booking shows it, out of 10 or converted with half stars, never rounded up to a full five.
2. The review text, with the source visible. Keep the Booking mark on each card so the guest knows which platform’s scale they are reading.
3. Recent reviews first. Lead with the reviews still inside Booking’s window and treat the archive as optional.
4. The two halves of each review, chosen on purpose. Booking asks what guests liked and what they did not, and both answers arrive in one review, which we measured in our guide to adding Booking.com reviews to a website.
5. Not the category scores. They do not drive the headline number, and a widget built on the overall score will not show them.

For a hotel the direct booking page is where this pays off, and our numbers on hotel reviews show why Booking, not Google, is usually the source with the most to say about a property.

Showing Booking.com reviews with JustReview

The reason to use us for this is that you keep everything and decide what to show. Our Booking.com reviews widget connects from the public property URL, refreshes every 24 hours, draws half stars instead of rounding a 9 up to five, and keeps each review it imports, including the ones Booking has since archived. You can put Booking next to Google or any other source in one widget and hide individual reviews that no longer describe your property. Every supported platform is listed in supported review sources.

One caveat, measured on the same four properties. The per-source average our widget shows is a plain average of every Booking review we hold, archived ones included, not Booking’s weighted three-year score. On the four properties it differed from Booking’s figure by between 0.03 and 0.12 stars, the largest gap on the 14-review guesthouse, where recency weighting does the most. If you want the exact Booking number on your page, quote it next to the widget with its date.

Booking.com is part of the Ultimate plan at EUR 39 a month, which also brings review curation and several profiles per platform, and paid plans start with a 7-day free trial; the full list is on the pricing page. Create your account, paste your property URL, and compare our count of reviews from the last 36 months with the count on your Booking page.

How this was measured

All checks ran on 6 October 2026. Booking’s rules are quoted from its partner guide to Guest Review Scores and its review guidelines for guests, read the same day. Scores, counts and labels on Booking come from each property page opened in a headless Chromium browser with an English locale, read from the visible score block and the page’s structured data. Our figures come from the 3,277 Booking.com reviews in the JustReview production database, read with SELECT queries only; the 36-month line was set at 6 October 2023 by review date.

The four properties in the score check are every Booking-connected property in our data whose import is current and whose page we could open. A fifth large property was left out of that check because our copy of its Booking reviews ends in January 2026, and for the same reason the comparison with Google stops on 4 January 2026. Property D was not used in the Google comparison because it had only ten Booking and five Google reviews in that period. Properties are not named because they are our customers. Four properties are a small sample, so read the results as what we saw on them, not as a rule for every listing on Booking.