How ratings work
Every score on AffSensor is derived from moderated user submissions and verified data. Nothing is paid for and nothing is estimated by us. Here is exactly how the numbers are built.
The Trust Index (0–100)
The Trust Index is a single 0–100 number that summarises how much a network has earned publishers' and advertisers' trust based on public, verifiable signals. It is a weighted sum of six inputs, capped at 100.
Paid placement does not affect the Trust Index. Featured and sponsored slots buy position in labelled ad units only. The number, the sort order, and every input below are computed purely from moderated user data and verification status.
| Signal | Weight | How it works |
|---|---|---|
| Publisher rating | +32 pts | Publisher-side weighted average (0–5 stars) scaled to the full weight. Reviews without vertical/GEO/duration context count at half weight; older reviews decay to a floor of 0.2 over ~2 years. |
| Advertiser rating | +12 pts | Advertiser-side weighted average (0–5 stars) scaled to this weight. Advertisers rate traffic quality, fraud transparency, volume, reporting and account management. |
| Payment reliability | +32 pts | Share of verified "paid" votes among all verified payment votes. Unverified votes are excluded. The most recent confirmed payout is shown alongside the number so recency is visible. |
| Verification level | +12 pts | +6 pts when the network is identity-verified, +12 pts when it is data-verified. Self-reported networks get 0. |
| Code of Conduct | +6 pts | +6 pts when the network has signed the public Code of Conduct and the application is verified. Applied or unverified networks get 0. |
| Contract risk | −15 pts | Deducts up to 15 pts. A warning-tier T&C flag removes 6 pts; a predatory T&C flag removes 15. Clean or unrated contracts do not deduct. |
| Maximum positive contribution | 94 pts | Total is clamped to the 100-point ceiling; the contract-risk penalty is then subtracted. |
Formula
trust_index = clamp(0, 100, round(
(publisher_avg_rating / 5) × 32 // publisher rating
+ (advertiser_avg_rating / 5) × 12 // advertiser rating
+ (payment_reliability_% / 100) × 32 // payment reliability
+ verification_bonus // 0 / 6 / 12
+ code_of_conduct_bonus // 0 or 6
− contract_penalty // 0 / 6 / 15
))How missing data lowers the score
Nothing is invented to fill a gap. If a signal has no data, it simply contributes zero — which lowers the ceiling of the score in real terms.
- No publisher reviews: the publisher-rating component is 0, so up to 32 points are unavailable.
- No advertiser reviews: the advertiser-rating component is 0, so up to 12 points are unavailable.
- No verified payment votes: payment reliability is 0, removing up to 32 points and hiding the "last confirmed payout" recency marker.
- Self-reported only: no verification bonus (0 pts vs +12 when data-verified).
- No Code of Conduct: no CoC bonus (0 pts vs +6).
- Contract not yet audited: no penalty either way. A network with audited-clean T&Cs shows the same 0 contract-risk penalty as one that hasn't been audited, so we surface the audit status separately on the profile.
The practical effect: a brand-new listing with no reviews, no verified payouts, no verification, no CoC, and no contract audit sits at 0. A fully-vetted network with strong two-sided reputation and a clean audit approaches 100. Most established networks land between these extremes.
The weights above are read directly from src/lib/trust-index.ts, the same module the ranking code calls. This page and the computed score cannot drift apart.
Two-sided reputation
Networks are rated separately by publishers and by advertisers. Each side has its own sub-ratings and its own weighted average. The two scores appear side-by-side on every network profile and directory row.
- Publisher score — payments, offers, support, tracking.
- Advertiser score — traffic quality, fraud transparency, volume, reporting, account management.
Weighting inside a rating
- Recency: newer reviews weigh more; a review's weight decays linearly to a floor of 0.2 over roughly two years.
- Context: reviews that disclose vertical, GEO, and duration count at full weight. Reviews without that context are halved.
- Verification: reviewers marked as verified affiliates are surfaced first on the profile.
Payment reliability
The payment reliability percentage is the share of verified "paid" votes among all verified payment votes in the last window. Unverified votes are excluded from public numbers. Each network profile also displays the most recent confirmed payout so recency is visible, not just the ratio.
Moderation
User submissions are held as pending until reviewed. Only approved submissions count toward public ratings and the Trust Index. Suspicious content is auto-flagged by an authenticity classifier and reviewed manually.
Featured placements
Featured and sponsored slots are labelled and never affect the Trust Index or ranking within rating-sorted views. Ranking sort orders (Trust Index, publisher rating, advertiser rating, payment reliability, review volume) are computed purely from moderated data.
Corrections
If you believe a specific review contains a factually false statement, use Report a legal concern. We will investigate and act per our Review & Content Policy.