Methodology

How the Lambah Score works

A ranking you can trust — because it rewards places that are both loved and busy, not one at the expense of the other.

The score

Every place earns a Lambah Score from two honest signals: how highly people rate it, and how many people bothered to. We combine them so neither can be faked alone.

Lambah Score = rating² × volume
rating² · number of ratings → normalised to a 0–100 scale

Squaring the rating is deliberate: it makes the gap between a 4.5 and a 3.5 matter, so a mediocre place can't out-muscle a great one just by being crowded — while volume still counts fully, so a genuinely popular favourite rises to the top.

Why not just… the rating?

Because a naïve "sort by stars" is worthless. It hands #1 to an airport kiosk with 4.9 from 600 ratings over a beloved roastery with 4.5 from 12,000. High average, tiny crowd — statistically meaningless, and easy to game with a handful of friends.

And sorting by volume alone is just as wrong — it rewards busy-but-mediocre. Watch what happens with two real-looking places:

PlaceRatingRatingsSort by volumeLambah Score
Place A3.3 ★2,500wins ✗loses
Place B4.1 ★2,000loseswins ✓
Pure volume ranks the 3.3-star place above the 4.1-star one — just for 500 extra ratings. The Lambah Score (rating²) flips it back: 4.1² × 2000 = 33,620 beats 3.3² × 2500 = 27,225. The better place wins, as it should.

The guardrails

What we never do

We never show the reviews. Not the text, not the authors — ever. The Lambah Score is computed from aggregate signals only. We turn millions of opinions into a single trustworthy number, and the opinions themselves stay private.

💡 Insight, never the words. That's the line we don't cross.
💡 Lambah · لمبة · Home · Brand kit · lambah.app