Rating methodology
Every scored provider receives a transparent total from 0 to 10: up to four points for result evidence, two for identity and accountability, and four for public transparency. The score measures what CSR can check. It is not a recommendation, safety guarantee, or profit forecast.
Completeness and source verifiability.
Operator and accountability strength.
Four visible one-point checks.
A public-research score, a CSR source review, and a disclosed human override have different provenance. Each rating card names its basis and assessment date. A 0/10 means no points were established in a completed assessment; it does not mean scam, unsafe, or unprofitable. Unscored means no deterministic 4 + 2 + 4 assessment is published. A disclosed human override may still set the official final score without inventing component points.
Calculation
For a computed assessment, total = result evidence + identity/accountability + public transparency. Software recomputes that total from the component decisions, and every point needs a dated reason and source reference. A direct human override is identified separately and may publish an exact final value while the deterministic assessment remains Unscored.
Commercial firewall
Buying a result-sheet service does not buy rating points. The resulting evidence is judged under the same rubric as unpaid evidence. Profitability is not an input: a complete loss-making record can earn the same evidence points as a complete profitable record.
Component one
How complete, loss-inclusive, and independently traceable is the result record?
| Points | Published rule |
|---|---|
| 0 | No usable, loss-inclusive result record is available. |
| 1 | A dated public record visibly includes losses, but remains provider-published and unreconciled. |
| 2 | CSR or another identified reviewer reconciled a limited record, while the original archive or full period remains incomplete. |
| 3 | A substantial source-linked, loss-inclusive record was reconciled with explicit scope, exclusions, and calculation limits. |
| 4 | CSR independently accessed or captured original sources and calculated a complete named scope with dates, exclusions, and limitations. |
Component two
This measures the strength of accountability, not popularity.
Anonymous, ambiguously mapped, or unsupported operator identity.
An attributable person, claimed entity, meaningful longevity, or jurisdiction exists, but legal or accountability details remain incomplete.
A named accountable operator or entity is confirmed with jurisdiction and an accountable contact or correction route. Dated internal CSR confirmation is labelled as internal; it is never disguised as a public filing.
Component three
One point for each check established in the dated public-research record.
Silence inside a completed research packet scores zero for that check. Absence of evidence is not evidence that the opposite is true.
Automatic scoring
A provider receives a score when CSR has a dated, correctly mapped research packet with an explicit decision for every component. Source files and URLs, capture dates, reasons, the component sum, and an assessment digest stay attached. An unresolved mapping, no packet, or a materially stale packet is Unscored.
Human decisions
A clear current decision from the human CSR team is sufficient authority for this rating lane. An exact 0-10 value is labelled human override with authority, date, scope, and limitations. The builder creates the dated record during implementation; no separate email, legal or formal review, ownership or disclosure investigation, conflict attestation, public filing, signature, second approver, or pre-existing file is required. An existing public-research calculation stays visible, or remains Unscored when none exists; component points are never reverse-engineered from the override.
Separate customer lane
Customer experience is never blended into the CSR Audit Team score.
The score appears only after at least five qualifying, approved customer-experience reviews. Research-only notes, rejected or incentivized reviews, and duplicate active reviews from the same reviewer do not count.
Newer reviews receive more weight, only the most recent eligible review from a reviewer contributes, and the score includes a stabilizer equivalent to seven reviews at 7.0/10. CSR publishes its one-year recency half-life. This follows the inputs described by Trustpilot's Trust Centre and its review guidelines; it does not claim to reproduce Trustpilot's private fraud systems or undisclosed constants.
Only the Verified Customers Rating may emit Review or AggregateRating structured data.
Citation rule
Quote the provider, exact total, 0-10 scale, 4+2+4 component scores, assessment date, basis, limitations, and canonical provider URL together. Never present this rating as customer sentiment, a profitability forecast, a provider-wide safety verdict, or proof that CSR and a provider are affiliated.
Corrections
Providers and readers can submit corrected identity, terms, result records, or claim evidence through the evidence and correction route. A material change triggers reassessment; paid status cannot block a downgrade or buy an upgrade.