How We Test

Research and hands-on testing are different evidence types, and our articles say which one they use.

Research-led evaluation

Our initial guidance is based on manufacturer specifications, compatibility documentation, support and security policies, and clearly identified independent evidence. We compare those sources against a defined use case and the five-part evaluation framework.

When we can say “hands-on tested”

We use that phrase only when the article records all four of the following:

• the exact unit and variant tested;
• the date or date range of testing;
• the protocol, setup, and comparison conditions; and
• the observations produced by that work.

If any of those records is missing, the article must describe the work as research, not hands-on testing. We never infer ownership experience or measurements from retailer listings or third-party reviews.

Limits belong in the article

Direct testing cannot prove every use case. Articles identify firmware versions, systems, accessories, network conditions, and other limits that could change the result.