A ranking only matters if you can trust the process behind it. This page sets out the rules the editorial team follows on every review, every report, and every score change.
1. Independence
The lab decides the scores. Commercial relationships do not influence rank, sub-score, inclusion, or the contents of a pros / cons list. Whoever is running a test does not know which provider pays a commission and which doesn’t.
We do not accept paid placement, sponsored slots, or “featured” upgrades. We have turned down offers larger than our annual revenue. We will again.
2. Methodology before opinion
Every score on this site is generated by the published 100-point rubric — speed (35%), privacy (30%), streaming (20%), value (15%). The rubric is public; the lab rig is public; the raw data is downloadable. Read the full breakdown at /methodology.
3. One rubric, every provider
The same battery of tests runs against every VPN we cover, on the same hardware, within the same testing window. Providers without an affiliate programme are tested against the same rubric as providers with one.
4. Sourcing and verification
- Primary measurements come from our lab. Speed, kill-switch failure rate, DNS leak rate, and streaming unblock rate are measured in-house.
- Privacy posture draws on independent audits, app source-code releases, jurisdiction analysis, and warrant canaries — all linked from the review.
- Provider claims are treated as claims, not facts, until verified.
- Two-reader rule: every published review and report is read by a second editor before publication.
5. Bylines and accountability
Every review and feature is signed. The author named on the page is the person responsible for the work — including factual accuracy. Where a piece is co-written or fact-checked by another team member, that is acknowledged inline.
6. Conflicts of interest
Editorial contributors disclose any current or prior employment, contract work, or personal subscription that could be perceived as a conflict before working on a review. Where a real or apparent conflict exists, that contributor does not score or edit the affected provider.
7. AI and automation
AI tools may be used internally for summarisation, copy-editing, or data extraction. No published review, score, or ranking is written by an AI without editor sign-off, and no provider claim is taken on AI’s say-so. The lab measurements are run and validated by humans.
8. Corrections policy
- Material errors (wrong score, wrong fact, broken claim) are fixed as soon as we verify the mistake. A correction note is appended to the page, dated, and explaining what changed.
- Score revisions from re-testing or methodology updates are logged on the affected review with the previous and current values, and the test date.
- Methodology changes are applied retroactively. We do not grandfather old scores when the rubric shifts — every provider is rescored under the new rules.
9. Right of reply
Providers may dispute a finding. If they can reproduce a test and produce conflicting measurements, we re-run the test and publish both results with dates. We do not quietly remove criticism; if we change a rating in response to provider feedback, the change is logged.
10. Funding and disclosures
We are funded by affiliate commissions. The full disclosure — including what does and does not change because of an affiliate relationship — is published at /affiliate-disclosure.
11. Report a problem
Errors, suspected bias, or anything else that seems off: editorial@vpntrackr.com. Include the URL and the specific claim or score you’re questioning.
These standards were last reviewed on 2026-05-18. They evolve as the lab and the threat landscape evolve; the current version is always at this URL.