Reviews Policy

The short version. We publish no guest’s words. Where a property page shows guest feedback, you are looking at two things and only two: Booking.com’s own numeric scores, and sentences we wrote ourselves by counting how often a subject came up. We host no reviews and accept none, so we cannot moderate or remove one. What we can do — and what this page commits us to — is correct or withdraw anything we have published about them.

Last updated: 3 September 2026

Why we have this policy

No South African law requires a website to publish a page like this one, and we are not going to pretend otherwise. What the law does require — under the Consumer Protection Act 68 of 2008 — is that we do not make false, misleading or deceptive representations to you. Summarising other people’s reviews on pages we earn commission from is an easy way to breach that without ever intending to.

So this page is a standard we set ourselves. It is here partly to keep us honest and partly so that you can check whether we are. If we depart from it, you have something concrete to hold us to.

What we publish, and what we never publish

Two things appear in a guest feedback block, and nothing else:

Booking.com’s category scores — Cleanliness, Location, Staff and so on, out of 10, exactly as Booking.com computes them from every review it holds for that property. We do not calculate them, adjust them or round them in our favour.
Sentences we wrote — drawn from a table of how many reviews mentioned a given subject. They are our words describing a count, not a paraphrase of anybody’s review.

We publish no review text at all. Not a quotation, not a “guests said” snippet, not a lightly reworded version of one. A review is written by the guest, and the licence to publish it sits between that guest and Booking.com. Being a Booking.com affiliate gives us no part of it. South African copyright law does permit quotation in narrow circumstances, so this is a deliberate rule we set ourselves with room to spare rather than the bare legal minimum — and we would rather a future editor knew that than discovered it and assumed the whole constraint was imaginary.

We also publish nothing that identifies a reviewer: no names, no countries, no travel dates, no profile details.

We host no reviews — so here is what we can and cannot do

This matters more than anything else on the page, and it is the reason our policy does not read like a review site’s.

You cannot leave a review here. There is no form, no rating widget, no comments section. We accept no submissions from guests, from property owners, or from anyone else.
We therefore cannot moderate, edit or remove a review. We do not hold the reviews. If a review on Booking.com is fake, incentivised or unfair, only Booking.com is in a position to take it down, and it is to Booking.com that a report about it has to go. We are not going to pretend to a takedown process we do not have.
The first line of defence is Booking.com’s. The reviews behind every score and summary here were collected and screened by Booking.com under its own review rules, and we rely on those controls. We do not independently verify individual reviews, and we do not claim to.
What we are responsible for is the output. Everything we derive from those reviews and put on this site is ours: the summaries, the counts we quote, the scores we choose to display and how we display them. That is what this policy governs, and it is where we can act.

How a summary is made

The method is mechanical on purpose. An impression of a property is easy to write and impossible to check; a count is neither.

We read a capped sample of reviews per property. That is the reason every claim carries a denominator: “Raised in 6 of 20 written reviews” tells you how thin or thick the evidence is. Where a property has hundreds of reviews on Booking.com, ours is a sample of them and the page says so.
Subjects are counted before anything is written. A theme table is built from the reviews in the sample, and the sentence is then written from the table. Nobody writes a summary from an impression and looks for numbers afterwards.
Criticism is never dropped for being unflattering. The “worth knowing before you book” column has the same heading, the same weight and the same position on every property, and it is never hidden. If nothing recurring came up as a drawback, the column says exactly that rather than disappearing. Suppressing minority complaints on pages we earn commission from is the specific failure this design exists to prevent.
We do not turn counts into adjectives. “Raised in 6 of 20 written reviews” is a fact. “Guests love it” is a claim we could not stand behind, so we do not make it.
Below 8 written reviews, no summary is published. A theme drawn from three sentences is noise presented as a finding. Under that floor the page shows the scores and states plainly that there is no written summary and why.

How we show the scores

Labelled, always. Booking.com scores are out of 10, from Booking.com, read on a date we print. A number with no scale and no source beside it is ambiguous, and an ambiguous score is misleading whether or not it is accurate.
Shown whether they flatter or not. We do not hide a score for being low, and there is no threshold below which a property’s scores disappear. Displaying only the good ones would misrepresent the whole set.
Compared honestly. Each score is shown against the average for the same category across the properties we hold that score for, with that number and its denominator printed beside it. The bar keeps a true 0–10 scale; we do not crop the axis to make a small difference look decisive.

Where you will see this, and where you will not

Coverage is uneven, and we would rather set that out than let the gaps imply something. Not every property here carries a guest feedback block, and of those that do, not every one carries a written summary:

Some properties show scores and a written summary.
Some show scores only — either because too few of their reviews carry written comments to draw a theme from, or because they clear that floor but no single subject recurred often enough to report. The page says which of the two applies, because they are not the same thing.
Some show nothing at all. That is an absence of data, not a verdict on the property. A new listing, or one with few reviews, will look like this and it means nothing bad.

A property has never been excluded for scoring badly, and a summary has never been withheld because it was unflattering.

We deliberately do not print a running tally of how many properties fall into each of those three groups. A count like that is correct on the day it is typed and quietly wrong for every day after — and a stale accuracy claim on a page about accuracy is worse than no claim. Each property page tells you its own position.

How current it is

Booking.com’s totals and averages move as new reviews arrive, so a score copied once and left alone drifts away from the truth. This is not hypothetical, and the clearest evidence we have is our own. We recorded review totals for twenty-nine properties on 3 September 2026, and read their reviews the next day. In that single day, two of the twenty-nine had moved: one from 595 reviews to 594, the other from 469 to 470 — one down, one up. Neither number was ever wrong. An undated one would simply have become wrong while sitting on the page. Two things follow from that:

Every block is dated. It states the month we read the reviews and, where a summary is published, the date range the sampled reviews were written in. The summaries currently on the site were read in September 2026. We have read a sample of reviews for eight of the places listed here, and published a written summary for each of those. For twenty-one more we read only enough to collect Booking.com's category scores, so you will see scores and no written summary — and the page says so rather than implying we read them and found nothing. The remaining twenty-four have too few reviews on Booking.com to say anything useful about, so they carry no guest-feedback section at all. That is deliberate: a theme drawn from two or three opinions is noise presented as a finding. If a date looks old to you, it is old — that is what printing it is for.
We intend to re-read them at least every three months and rebuild the summaries from scratch each time rather than patching the old ones. Three months is our judgement of the point at which a small sample stops being a fair picture of a property; if we find scores drifting faster than that, we will shorten the interval rather than leave the stated one unmet.

We are not going to claim a refresh history the site does not have. The date on the page is the honest answer to “when was this last checked”, and it is set by the pipeline that builds the data, not typed by hand.

What we do not do

We do not verify reviewers. We have no way to confirm that any individual review was left by a real guest who really stayed.
We do not inspect or vet properties. Nobody from this site has visited them, checked them or approved them. We are not an inspectorate and no property here carries our endorsement.
We do not solicit, incentivise or write reviews — not for properties we list, not for ourselves, not anywhere.
We do not sell placement. No property has paid to appear on this site or to appear higher up a page, and no property can pay to change, soften or remove what its summary says.

We earn a commission — and what that does not buy

Leisure Isle Accommodation earns a commission from Booking.com when you book a stay after clicking through from this site. It costs you nothing extra and it is the only way this site makes money. You should know that while reading anything here about how good a property is.

It does not change what the summaries say. We earn the same rate whichever Booking.com property you choose, so there is no property we are paid more to flatter — and the counting method leaves no room to flatter one anyway, because every claim is tied to a number that can be checked against the data behind it. The affiliate disclosure sets out the commercial arrangement in full.

Tell us if something here is wrong

This is the part of the policy we can actually act on, so please use it. If you are a property owner, a guest, or simply a reader who thinks a summary or a score on this site is inaccurate, unfair or out of date, tell us and we will look at it.

What happens then:

We check the published sentence against the counts it was built from. If it does not match, it is wrong by definition and it comes down.
If the underlying figures have moved on, we re-read the property’s reviews and rebuild its summary rather than editing the old wording.
If we cannot satisfy ourselves that a summary is a fair account of the sample, we remove it and leave the scores. Removing something we cannot stand behind is always available to us and costs nothing.
If the complaint is really about a review itself — that it is fake, malicious or was written by someone who never stayed — we will tell you so plainly, because it needs to go to Booking.com. We can still remove our summary in the meantime.

Use the contact form. Tell us which property and which line, and it helps if you say what you think the accurate position is. We would rather fix it than defend it.