From claim to useful conclusion.

How a review moves forward

1. Research: identify the product, audience, pricing and claims that matter.
2. Evidence: record source quality, dates, observed behavior and meaningful limitations.
3. Review: connect the record to practical fit, score dimensions and a clear verdict.
4. Publication Gate: check required evidence, metadata, trust, commercial and content conditions.
5. Human approval: an authorized editorial decision confirms that the review can be published.
6. Published: show the review publicly with review and recheck dates so it can be corrected or refreshed.

How to read our evidence labels

Official or vendor claim

A statement found in a first-party product, pricing, help or policy source.

Observed or hands-on evidence

A behavior seen during a declared test session, with the access and plan recorded.

Editorial assessment

Our reasoned interpretation of fit, trade-offs or usefulness, kept separate from a vendor fact.

Not confirmed

The available record does not support the claim yet.

Limitation

A plan, access, scope or evidence boundary that changes how a finding should be read.

Scoring, dates and corrections

Scores use product-specific evidence across ease of use, output quality, AI usefulness, value, flexibility and trust and support. A score is withheld or described as provisional when critical testing is incomplete. Pricing and product claims carry the date they were checked; published reviews can receive later rechecks as products change.

If a factual error is supported by new evidence, we correct it and may update the review or recheck date. Editorial assessments can remain unchanged when the factual basis still supports them. See the Corrections Policy and How We Test for practical detail.