A clear score. No hidden formula.
The first version keeps the maths simple and the rules visible.
An arithmetic average.
We add the 1–5 star ratings from published reviews that have not been deleted, then divide by their count. We round the result to one decimal place. For example, ratings of 5, 4, and 3 produce a score of 4.0 from 3 reviews. Rating distributions show how many of those reviews awarded each star level; rounded percentages may not sum to exactly 100%.
What counts — and when.
Pending, rejected, withdrawn, and deleted reviews do not count. When a review is published or withdrawn, the score and distribution update directly from database records. Editing a review moves it back to pending; it starts counting again after approval. A business with no published reviews displays “No reviews yet” rather than a rating of zero.
The same weight for every published review.
Claiming a profile, paying for a service, receiving an invitation, or providing purchase evidence does not change the weight of a published rating. Reviews are not weighted by recency in this first version. We display the review count so you can judge the amount of evidence behind a score.
Demo data is separate from real activity.
Fictional demo businesses and their sample reviews carry a Demo label. Demo pages are excluded from search engine indexing and the production sitemap. Production directory entries, badges, counts, and scores come from saved records. We do not invent public platform statistics.