How We Evaluate Tools

A published, repeatable method — so you can judge our conclusions for yourself.

Last updated: 3 October 2026

When we show a score

We only show a score after hands-on testing by a named person, using the criteria below. Tools that haven't been tested show an evaluation checklist instead of a score. We never show star ratings based on anything else, and we don't display aggregated "user ratings".

What we look at

Core job performance (30%) — how well it does the main thing it promises, measured on a fixed, repeatable test task for its category.
Ease of use (20%) — time to first useful result for a new user; clarity of the interface.
Value and pricing transparency (20%) — what the free plan genuinely allows; clarity of paid plans, renewal pricing and limits.
Privacy, security and data control (15%) — published policies, data-use settings, export options.
Support and documentation (15%) — quality of help documentation and responsiveness to a test question.

Category-specific tests

Each category has its own fixed test so results are comparable. For example:

Grammar tools: a standard passage with a known set of errors, plus a clean passage to measure false positives.
SEO tools: keyword estimates compared with Google Search Console data for a test site; a site audit on a site with deliberately introduced issues.
Video tools: producing a 60-second video from the same 150-word script.
Automation tools: building the same three-step workflow and testing error handling.
Hosting: load-time checks from several regions, plus renewal-price verification.

Recording results

Each test records who tested it, the date, the plan used and the findings. Tool pages show this information alongside any score. Tests are repeated when a product changes significantly.