How TrustScore is calculated
The methodology behind the TrustScore shown on every Ardent Reviews listing.
Every reviewed entity on Ardent Africa gets a TrustScore from 0–100. We publish exactly how it's calculated, because a score you can't understand isn't one you can trust.
The formula
TrustScore is a weighted average of all published reviews, converted to a 0–100 scale. Two factors adjust how much each review counts:
- Recency: recent reviews count more than old ones. A review from the last 90 days counts fully; older reviews count progressively less, down to a quarter weight for reviews more than two years old. This keeps the score reflecting current experience, not history.
- Verification: a review tied to a real, verified interaction (an invitation sent after an actual visit, purchase, or appointment) counts more than an open review with no such link.
What we show alongside the score
- How many reviews the score is based on. A 4.8 from 200 reviews and a 4.8 from 4 reviews aren't the same confidence. We label listings with few reviews as "still gathering data" rather than pretending the score is settled.
- Response rate: how often the entity replies to reviews. This is shown separately from the score itself, since responsiveness and service quality are different things.
Why we don't hide the formula
Some review platforms treat their scoring as a trade secret. We think that's backwards for a tool meant to build trust. If you can't see how a score is built, you can't fully trust it. Publishing the method also means an entity always knows exactly what improves its score: more recent reviews, and more verified ones.
Keeping scores honest
TrustScore only reflects reviews that pass our moderation checks, and entities can't hand-pick who gets invited to review them. Invitations must cover a genuine batch of real interactions, not a chosen few happy customers. See Write a review for how verified reviews work.
If an account is later banned for abuse, its reviews are removed from both the public list and the score at the same time, so a banned reviewer can no longer influence any listing.
"What people say"
Listings with at least 5 published reviews also show a short AI-generated summary of common themes, a quick read of what reviewers keep mentioning, alongside a few recurring positive and negative tags. It's generated only from that entity's own published, already-moderated reviews (never invented), and it's refreshed periodically rather than on every visit, so it reflects the overall pattern rather than the single most recent review.
Last updated: 3 August 2026