HOW WE RANK.
Every gym in our directory is scored using a data-driven system designed to surface the best facilities, not just the most popular ones.
The Problem with Simple Averages
A gym with a single 5-star review shouldn't outrank a gym with 200 reviews averaging 4.7 stars. Simple averages reward obscurity and punish popularity. We use a Bayesian average to solve this: a statistical method that balances rating quality with review volume to produce fairer rankings.
Three Ranking Factors
Average Rating
The gym's average Google review score on a 1–5 star scale. This is the primary quality signal.
Review Volume
The total number of reviews. More reviews mean more confidence in the rating. A volume bonus rewards well-established gyms.
Sentiment Analysis
AI-powered analysis of actual review text to detect patterns that star ratings alone can miss: coaching quality, facility cleanliness, and community culture.
The Bayesian Average
At the core of our ranking system is a Bayesian average, a technique used by platforms like IMDb and Amazon to produce fair rankings when data volume varies widely between entries.
Formula
Score = (v / (v + m)) × R + (m / (v + m)) × C + bonusRThe gym's average rating from Google reviews (1–5 stars).
vThe number of reviews the gym has received.
mThe minimum reviews threshold (set to 5). This is the "prior". Gyms with fewer than 5 reviews are pulled toward the baseline, preventing untested gyms from dominating rankings.
CThe assumed mean baseline (set to 4.0). Before any reviews come in, every gym is assumed to be "average" at 4.0 stars. This prevents new gyms with one perfect review from jumping to #1.
bonusA logarithmic volume bonus (capped at 0.5) that gives a small edge to gyms with a high number of reviews. This rewards consistently well-reviewed facilities without letting volume alone determine rank.
How It Works in Practice
Consider two gyms. Gym A has 3 reviews with a perfect 5.0 average. Gym B has 150 reviews with a 4.7 average. Under a simple average, Gym A would rank higher. But with only 3 reviews, we can't be confident that score reflects reality.
Our Bayesian system pulls Gym A's score toward the 4.0 baseline because it has fewer than our minimum threshold of 5 reviews. Meanwhile, Gym B's large review count means its score barely shifts from its true 4.7 average, and the volume bonus pushes it even higher.
The result: proven quality rises to the top, while unverified ratings are treated with healthy skepticism.
Beyond the Stars
Star ratings tell part of the story. Our AI-powered sentiment analysis reads the actual text of reviews to detect patterns that numbers miss:
- check_circleCoaching quality and instructor expertise
- check_circleFacility cleanliness, equipment condition, and safety
- check_circleCommunity atmosphere and welcoming culture
- check_circleValue for money and membership flexibility
- check_circleClass variety and scheduling convenience
Each gym receives a sentiment score from −1.0 to 1.0 and a label (Excellent, Good, Mixed, or Poor) based on this analysis. This helps surface gyms where reviewers consistently praise what matters most to athletes.
What We Don't Do
- closeWe don't accept payment for higher rankings
- closeWe don't manually adjust scores to favor specific gyms
- closeWe don't penalize gyms for not having a Revgear partnership
- closeWe don't use data older than 12 months in our calculations
Ready to find a top-ranked gym near you?
Explore the Directoryarrow_forward