1. Lock the item identity
We record the exact model, variant, manufacturer, and stable identifiers before gathering evidence. One variant’s evidence is never silently applied to another.
2. Match sources to claims
Manuals and primary documentation establish specifications. Independent testing may establish performance. User reports are labeled as reports, not converted into prevalence claims.
- Primary manuals, regulators, standards, and specifications
- Transparent independent tests and expert reviews
- Retail listings only for identity and availability
- User discussions only as attributed sentiment
3. Compare consistently
Comparison and best-for pages declare criteria before drawing conclusions. Best-for means a fit for a stated need, not a fabricated universal winner.
4. Show limits and maintenance
Every article displays review and update dates, evidence limitations, source health, and a next-review date. Material model changes, recalls, contradictions, or broken links trigger re-review.