Verified vs unverified reviews: what a badge actually proves

A "verified" badge feels like a guarantee that the review is true. It is not, on any platform. It establishes something much narrower, and the gap between what it proves and what readers assume is where most of the trouble lives.

What verification can and cannot establish

Verification answers one question: did this person actually deal with this business?

It says nothing about whether their account is accurate, whether they were reasonable, or whether their experience was typical. A verified review from a customer who misremembers is still wrong. An unverified review from someone with a receipt in a drawer is still true.

What verification does is make fraud expensive. Anyone can write an unverified review. Faking a confirmed purchase at scale is much harder, and that difficulty is the entire value of the badge.

The levels, and what each one is worth

Different platforms use different ladders, and most do not publish theirs. Ours has six rungs:

Nothing verified. Someone wrote a review. We know nothing else. It still publishes — excluding these would silently discard people who paid cash or lost the receipt — but it carries the least weight in the rating.

Email confirmed. The reviewer clicked a link in their own inbox. This proves an address exists and someone controls it. It is a low bar, but it stops the cheapest kind of bulk fabrication.

Matched to a business record. The business supplied an order reference and it matches. Worth noting what this does not establish: the business supplied the data, so it proves the business believes this person is a customer. That is meaningful but it is not independent.

Purchase confirmed by a connected system. An order from a linked commerce platform, confirmed without the business as intermediary. This is the first level where the business could not have fabricated the underlying fact.

Identity verified. The reviewer's identity was checked. Rare, and reserved for cases where it matters.

Multiple independent signals agree. More than one of the above, from sources that do not depend on each other.

Why we publish unverified reviews at all

Because the alternative is worse.

A platform that only publishes verified reviews looks rigorous and quietly discards a large part of what actually happened — the cash customer, the person who bought through a third party, the one who no longer has the email. Those experiences were real.

So they publish, they are labelled honestly, and they count for less. A rating built mostly on unverified reviews is worth less than one built on confirmed purchases, and that is not hidden in a methodology page: the mix is shown on every profile, so you can see which you are looking at.

Why a five-star review can produce a low average

This surprises people, so it is worth stating plainly. A single unverified five-star review does not produce a 5.0 here. It produces something lower, because weight follows evidence.

That is deliberate. If one unverified review could set a perfect score, the cheapest possible fraud would also be the most effective.

The weighting is published in full, including the exact multipliers, on the rating methodology page.

What to look for as a reader

The proportion, not the presence. One verified review among forty tells you nothing. Look at what share of the whole is verified, and at what level.

Who supplied the proof. A badge that rests on data the business provided is weaker than one that does not. Platforms that collapse both into a single "verified" label are hiding the more useful distinction.

Whether the platform explains its ladder at all. If "verified" is not defined anywhere, it means whatever the platform finds convenient.

If you wrote a review and it says unverified

You can usually raise it. On your review page there is an option to add proof — a receipt, an order number, a booking confirmation.

We check it, record the decision, and discard the file. The business never sees what you uploaded, and it does not change your words or your rating. It changes only how much your review counts, which is the one thing evidence should change.