Which ratings can you actually trust?

Star averages are easy to inflate and hard to read. We rebuild them from the reviews that actually say something.

The problem with star averages

A restaurant's rating on a review platform counts every review equally: a considered paragraph about a meal and a drive-by star tap with no text weigh exactly the same. Ratings have crept upward for years — on the big platforms most restaurants now sit between 4.2 and 4.8, which makes the differences that matter nearly invisible. And because bare star ratings are cheap to produce, they are also cheap to buy.

The result: a 4.5 can be a genuinely excellent restaurant, or a tourist trap with a review-collection habit. The number alone can't tell you which.

TrustyRating — built from detailed reviews

TrustyRating is built only from detailed reviews — reviews that describe the experience: what was ordered, how the service went, what a regular noticed on their third visit. Reviews that are just a star tap, or a few words of generic praise, don't count toward it. Review-site averages count every review; ours counts the ones that explain themselves.

Think of it like critics vs. audience scores for films: not a claim that other reviews are fake, but a different population of reviewers — people who took the time to say why. Where the two numbers agree, you can relax. Where they diverge sharply, there is usually a story, and we show you both numbers so you can see the gap yourself.

When a place doesn't yet have enough detailed reviews for a reliable number, we say so instead of showing one.

TrustyScore — how much signal exists

TrustyScore is a different kind of measure: not how good a place is, but how much trustworthy evidence exists about it — recent review activity, press coverage from sources we track, and social attention. A high TrustyScore means the picture is well-documented; a low one means the place is flying under the radar (which can be a discovery, not a defect).

What we watch for

Because we read reviews rather than just counting them, certain patterns stand out: bursts of uniform praise from accounts with little history, reviews that describe pressure to leave a rating, headline averages that run far above what the substantive reviews describe. Where reviews themselves describe these patterns, we flag it — plainly, based on what the reviews say.

Questions

Why is a restaurant's TrustyRating lower than its Google rating?

Usually because detailed reviewers are a more critical population than star-tappers — that's normal and the gap is typically modest. A very large gap is worth reading about: it means the people who described their experience tell a different story than the headline number.

Do restaurants pay to be listed or rated?

No. Restaurants can't pay for placement, ratings, or removal of information. There are no ads.

Where does the data come from?

Public reviews across major platforms, press coverage from publications we track, and public social activity — refreshed on a rolling basis. Every number links back to what it was computed from.

Why only Paris?

Depth over breadth. Reading every review of every restaurant is expensive to do properly, so we'd rather cover one city at full depth than many cities thinly. More to come.

I own a restaurant and my rating looks wrong.

TrustyRating is computed the same way for every restaurant, from the detailed reviews your guests wrote. We don't hand-adjust numbers — for anyone. If you believe data about your restaurant is factually incorrect (wrong address, closed, duplicate), contact us and we'll fix it.

TrustyTable is independently built in Paris. Start browsing →