RankPlot uses more than one research method. Every article should state what was actually checked and must not turn documentation research into a hands-on claim.
Our Evidence Labels
Controlled Test
RankPlot ran a defined test in an environment it controlled. The article should publish the date, software and browser versions when available, hosting and cache conditions, page workload, device profile, run count, aggregation method, raw or row-level results, and known limitations. Results apply to that setup; they are not universal field performance.
Hands-On Check
RankPlot accessed and used the named product or workflow for the specific actions described. The page should identify the scope and date. A signup, installation, checkout, or interface check does not prove long-term reliability, customer support quality, or business results.
Documentation-Based Review
RankPlot compared current official product pages, documentation, pricing pages, support policies, public repositories, or other cited sources. This method can document features and published terms, but it is not a verified purchase or performance test.
Editorial Judgment
Recommendations such as “best for beginners” or “better for a specific workflow” are reasoned conclusions. The page should show the criteria and tradeoffs behind the judgment. Editorial scores are not converted into customer ratings or laboratory measurements.
Controlled Performance Test Standard
- Define one question and the variables being compared.
- Keep the host, content, plugins, workload, cache state, and test profile consistent unless the variable itself requires a change.
- Run repeated samples and retain failed or invalid-run notes instead of selecting only favorable results.
- Publish the aggregation method and row-level or raw result access when it is safe to do so.
- Separate lab data from Core Web Vitals field data and from business outcomes such as revenue or conversion.
- State what the test did not measure.
The current public example is the 120-run Astra and Hello Elementor dataset. It covers one host, one test date, three build stacks, two page workloads, desktop and mobile profiles, and cold and warm cache states. It does not establish universal theme speed, revenue, bounce rate, or conversion effects.
Commercial and Product Research
For hosting, themes, plugins, and SEO tools, RankPlot checks the claims relevant to the article rather than applying one unsupported site-wide scorecard. Volatile facts such as price, renewal terms, plan limits, refund rules, support access, and version numbers should use dated official sources. If a product was not purchased, installed, or tested, the page must say so or avoid implying otherwise.
Updates, Corrections, and Conflicts
- Material factual corrections are made when reliable evidence shows a claim is wrong.
- Prices and product policies are rechecked when a page is substantially updated; old review dates are not presented as current verification.
- Affiliate eligibility does not decide rankings, scores, or whether a product is included.
- Vendors may provide corrections or source links but do not receive approval over the final conclusion.
For the wider rules, read the Editorial Policy. For commercial links, read the Affiliate Disclosure. Questions about a test or correction can be sent through the contact page.
Last reviewed: August 31, 2026.