Publication Policies, Scoring Rubric & UK Legal Hub
Built specifically for United Kingdom operation—accounting for UK GDPR, the Data Protection Act 2018, PECR cookie consent, ASA/CAP affiliate advertising codes, and the UK Online Safety Act 2023.
Editorial Policy & Evidence Standards
Our editorial process adheres to strict evidentiary standards, preventing subjective bias or commercial interests from inflating companion platform evaluations. Every score published on Matelytics is grounded in verifiable evidence.
The 5-Tier Evidence Hierarchy
All quantitative metrics and qualitative statements across our seven evaluation pillars are classified according to our formal Evidence Hierarchy (Class 1 to Class 5):
| Evidence Class | Classification Name | Evidentiary Standard | Scoring Treatment |
|---|---|---|---|
| Class 1 | Officially Documented | Primary developer documentation, published whitepapers, official API specifications, verified terms of service, and public architectural blueprints. | Scored (0.0 – 10.0) |
| Class 2 | Independently Verified | Empirical testing conducted directly by Matelytics researchers (e.g., memory recall benchmarks, multimodal latency timing, automated stress tests). | Scored (0.0 – 10.0) |
| Class 3 | Editorial Observation | Qualitative appraisal by experienced editorial reviewers (e.g., interface responsiveness, narrative roleplay depth, persona coherence). | Scored (0.0 – 10.0) |
| Class 4 | Awaiting Verification | Unverified vendor claims or features undergoing technical audit. Score must remain null until complete technical evidence is gathered. | Score Null / Retained Denominator |
| Class 5 | Not Applicable | Feature does not exist by architectural design (e.g., voice synthesis on a purely text-only model). Confirmed absent by editorial team. | Excluded / Weight Redistributed |
The Three Mandatory Publication Gates
To prevent incomplete or hazardous platforms from achieving high aggregate scores through superficial features, Matelytics enforces three mandatory scoring gates:
Pricing & Token Transparency
Audits subscription tiers, recurring billing disclosures, token burn rates, and refund policies.
Privacy & Data Handling
Audits user chat retention, training opt-outs, cloud vs local persistence, and data deletion tools.
Safety & Content Governance
Audits age assurance, content guardrails, reporting workflows, and CSAM/self-harm prevention.
Mandatory Gate Enforcement: If any mandatory gate has insufficient evidence (evidenceClass === 4) or fails minimum standards, the numerical composite score is strictly withheld and marked “Unscored — Awaiting Full Technical Audit”. Even when a composite score is withheld, our full 7-pillar evidence matrix remains published for complete consumer transparency.
Zero-Fabrication & Editorial Independence Guarantee
Commercial affiliate partnerships, advertising deals, or vendor sponsorship requests have zero influence over platform scores, reviews, or comparative ranks. Matelytics editors operate under strict editorial freedom; no platform vendor may purchase a higher score or bypass our mandatory publication gates.