Editorial & affiliate policy:
evidence and limits.

What a recommendation means here

A recommendation is tied to a stated task and the evidence shown in that article. A short trial can support trying a tool for that task; it cannot establish product-wide accuracy, long-term reliability or suitability for every team.

We distinguish our observed results, official product documentation and editorial suggestions. “Untested” is an evidence boundary, not a failing grade. Different account tiers are named rather than treated as equal-budget comparisons.

How AI is involved

AI assistance is used for research, test-interface operation, documentation, drafting and site development under Marvin Smit’s direction. Articles identify fictional inputs and assistant-operated tests. We do not label an AI-generated illustration as a product screenshot or an editorially corrected answer as an original model output.

Read how we test for the source material, answer keys and checks behind the first articles. Published screenshots are selected from original captures; explanations sit beside them.

Provider links, affiliates and sponsorship

The current articles use ordinary provider links without affiliate tracking. The site has no active affiliate placements or sponsored rankings. We have not added commission rates, partner badges or membership claims without a verified relationship.

If affiliate links are introduced, the relevant recommendation will disclose that the lab may earn a commission. Any paid placement or material free access will be identified where it matters to the reader. Payment does not buy a test result or remove a material limitation.

Links to a provider do not mean it endorses the lab. Purchasing, plan terms, account support and refunds are handled by the provider.

Prices and dates

Pricing examples use USD and identify billing assumptions, seat counts, required onboarding and excluded extras where relevant. A monthly equivalent of annual billing is not described as a month-to-month offer. Account-specific promotions are distinguished from public catalogue rates.

Test dates, source-check dates and publication dates have different meanings. We do not refresh a date merely to make an unchanged article appear new, and we do not imply a new trial when reusing existing evidence.

Correcting an article

Report a suspected error through Contact & corrections. We check the source and the affected claim. Material factual corrections should include a dated note explaining what changed; routine spelling or layout edits do not need a correction notice.

A new vendor claim does not automatically replace an observed result. When a feature changes, the article should distinguish the old test from the newer documentation or retest.

Reusable templates

The lab’s note and email templates are starting points to adapt. A copied prompt does not guarantee correct output or a saved CRM action. Check names, commitments, recipients and dates before using the result in your own work.