Methodology

How we test search, SEO and AI data APIs

Documentation can prove that a vendor describes a capability. It cannot prove how that capability performed in our test.

Evidence grades

VERIFIED HANDS-ON means we executed or directly observed the stated behavior and retained evidence.

OFFICIAL DOCUMENTED means a current primary provider source supports the statement, but we did not execute that behavior in the stated test.

THIRD-PARTY identifies outside evidence. NOT TESTED means our stated test did not execute it. UNAVAILABLE / LIMIT records an attempted frozen task blocked by an observed constraint. UNRESOLVED means the available evidence is insufficient to make the claim.

Frozen inputs

We design comparisons before seeing results. Query sets, market, language, device and evaluation rules are frozen. We do not replace inconvenient keywords or missing results after execution.

No fake metric equivalence

Difficulty, authority and trust metrics can use different datasets and formulas. We retain provider labels and compare observable disagreement instead of averaging them into a universal score.

Missing data is evidence

A missing result is not zero. A plan limit is not a product-wide absence. A delayed result is not automatically a failed feature.

Commercial independence

Affiliate links may fund this site. They do not change evidence grades, fill missing values or determine a benchmark result. See our affiliate disclosure.