How We Review
Every article should tell you what was checked, what remains uncertain and why the downside matters.
Last reviewed August 30, 2026 · Maintained by FlawCheck Editorial Desk
1. We label the evidence level
Every review and comparison should identify the strongest evidence behind its conclusions. A launch article is not silently upgraded to a hands-on review.
| Evidence label | What it means |
|---|---|
| Hands-on Review | A contributor personally used or measured the product, and the article states the source, approximate test period, conditions and limitations. |
| Research-Based Review | Analysis based on attributable sources such as official documentation, credible reporting, independent tests and owner evidence. |
| Provisional Analysis | A product is new, unreleased or incompletely documented. Conclusions and scores may change as stronger evidence becomes available. |
| Rumor-Based Comparison | Some information has not been confirmed by the manufacturer. Confirmed facts and unverified claims must be identified separately. |
2. We verify before judging
We prioritize manufacturer documentation and primary sources for specifications, compatibility, warranty terms, recalls and support commitments. Marketing claims are identified as claims. Conflicting regional specifications are called out instead of blended together.
3. We score the compromise
- 0–2: minor trade-offs.
- 3–4: low-impact inconveniences.
- 5–6: meaningful compromises.
- 7–8: significant limitations or likely dealbreakers.
- 9–10: severe defects or safety risks supported by strong evidence.
The score considers severity, frequency, affected users, workarounds, ownership cost, evidence confidence and similarly priced alternatives. It is editorial judgment, not a laboratory certification.
4. Hands-on testing, when performed
The article states the product source, approximate test period, conditions, tools and limitations. We publish measurements only when the procedure can be described. Review samples, loans or products supplied by a manufacturer are disclosed.
5. Research-based launch analysis
For new products, scores are provisional. We compare confirmed specifications and policies with the market, identify unanswered questions and update the article when reliable evidence becomes available.
6. Sources, AI assistance and accountability
AI-assisted tools may help organize research and drafts. Claims still require source verification, AI output is not evidence, and the publisher remains accountable for publication and corrections.
7. What we do not do
8. Commercial independence
Affiliate links may earn revenue without increasing the buyer's price. Compensation does not guarantee coverage, change a score, prevent criticism or provide a brand with pre-publication approval. See our affiliate disclosure.
9. Updates, corrections and responsibility
We update material changes such as firmware, support policy, recalls and corrected specifications. Substantive corrections are recorded rather than hidden. See our corrections policy and corrections log.
This methodology is maintained by FlawCheck Editorial Desk. Questions about evidence, testing labels or corrections can be sent to [email protected].