Transparency Dashboard
Real-time platform statistics and authenticity metrics
0 verified
0 today
Legacy average rating (pre-2026 instrument), from 0 ratings
actions taken
The 2026 instrument is not active in this environment, so there are no questions to show. The formulas below describe how it scores answers wherever it is running.
Each 1–5 answer is scaled: score = (value − 1) ÷ 4 × 100.
score = (value − 1) ÷ 4 × 100Brand recall rate = (correct + 0.5 × partial) ÷ answers × 100.
recall rate = (correct + 0.5 × partial) ÷ answers × 100The published score is the unweighted mean of Attention, Clarity, Feeling and Brand recall, shown to one decimal. The “What would you do next?” answers are published as a distribution and are never averaged into the score.
No weighting by reviewer reputation, recency, payment status, or anything else is applied to the published score.
rank value = (25 × category average + n × ad average) ÷ (25 + n)This ordering prevents an ad with 3 ratings outranking one with 400. It is used only to order lists; it is never displayed as a score.
12 complete ratings to publish a score — below 12 we show “early signal” and the raw response count instead.
25 for leaderboard eligibility.
50 for any commercial or aggregate product.
Ratings collected before the 2026 instrument are preserved unchanged and labeled “Legacy score”. They are never converted into the new dimensions and never blended into new scores.
Responses submitted implausibly fast, straight-lined, or showing coordinated patterns are flagged and excluded from published aggregates. Flagged responses are retained (never silently deleted) and audited.
The audit runs nightly, re-deriving those checks from the stored answers rather than from what the browser reported.
- 2026-09-01Introduced the five-question instrument (v1); prior single-star ratings preserved as v0.
How We Ensure Authenticity
Email Verification
All users must verify their email before submitting reviews
One Review Per User
Each user can only review an advertisement once
Transparent Moderation
All moderation actions are logged and can be appealed
How Ratings Are Calculated
Pre-2026 star ratings are a simple average — no weighting, no manipulation, just honest math.
Ads rated with the 2026 five-question instrument are scored differently, and the two are never blended. Read how ads are scored.
See something suspicious?
Our Growing Community
We’re a startup building our community one authentic review at a time. Every number you see is real, even if it’s small. We believe in quality over quantity.
Commercial firewall
What money can and cannot do on this platform. These are commitments, and the tests that enforce them run on every build.
- Scores are never for sale, for edit, or for removal, at any price, under any agreement.
- Commercial relationships are disclosed on the page where they appear, not only in a policy document.
- Sponsored content is ranked in a separate namespace and labeled at every impression.
- The scoring pipeline has no access to commercial records; that separation is enforced in the database, not by convention.
- Any exception ever made is logged publicly in the moderation log.
Loading the request counter…
Our Commitment
We’re building a platform based on trust and authenticity. Every number you see here is real. We’re a startup, and we’re growing organically with real users and real reviews. No bots, no fake engagement, just honest community building.