§ M — METHODOLOGY
How we score every business
Every rating on this site is computed by software from public data sources — no human edits, no paid placements, reproducible by anyone.
The Overall Score · 0–10
A weighted blend of four signals: 45% confidence-weighted rating, 25% review volume, 20% online presence, 10% website performance. A 1-review 5-star business cannot outrank a 200-review 4.6-star institution — both quality and track record matter.
The Confidence-Weighted Rating · Bayesian
A trust-weighted version of the customer rating. Pulls businesses with very few reviews toward the city-wide average using a strong prior, so a brand-new restaurant with two 5-star reviews doesn’t outrank a 4.7-star institution with 800 reviews.
The Review Volume · log-scaled
How many customers have actually weighed in. We use a log scale capped at 200 reviews — diminishing returns above that. A business with 100+ reviews has earned a real track record; a business with 3 hasn’t yet.
The Customer Rating · Google stars
The raw Google review average — exactly what you’d see on Google Maps. We don’t modify, smooth, or filter it. If a business has 4.8 stars on Google, we display 4.8.
The Online Presence Rating · 0–10
A broader composite of website performance, customer signal, and visibility on Google. At a glance, how strongly does this business show up online?
The Website Rating · 1–10
Assesses how quickly a site renders on mobile and desktop, and how search engines perceive it — which directly impacts how many customers Google sends to the business. A lower rating means lower Google ranking; a higher rating means more visibility.
Award Eligibility · badges
To earn an award badge — Best of the Best, Category Winner, or Subcategory Winner — a business must be currently operational, have at least 10 Google reviews, and compete in a subcategory with at least 5 businesses. Below those thresholds we still rank and score, but no trophy.
Spot something wrong?
If a number looks off on your listing, email info@bestofcolumbusga.com and we'll re-enrich it on the next pipeline run. We don't adjust scores by hand, but we do fix bugs.