Testing process overview

How We Test VPNs

This page is a practical summary of our VPN testing workflow. It explains the steps we repeat every cycle before a score or ranking is published.

Baseline run

Each cycle starts with no-VPN measurements so VPN performance is compared to a known baseline.

Speed and latency

We measure regional download speed (EU/US/Asia) and latency for consistency across common use cases.

Streaming reliability

We track unlock consistency for major streaming libraries and retest after platform-side detection changes.

Safety controls

Kill switch behavior and leak resilience are verified to reduce accidental exposure when connections drop.

Score model used in production

The current score model is fixed and documented publicly. Commercial terms are not part of the formula. For implementation details, see /methodology.

Metric	Weight
Speed composite	24%
Latency	10%
Logging policy	14%
Ownership clarity	8%
Jurisdiction risk	9%
Audit status	10%
Streaming unlock	11%
Torrent policy	7%
Kill switch reliability	7%

Re-test cadence

Providers are re-tested monthly, with additional runs after major app updates, policy changes, and major streaming-block shifts. Freshness is exposed through visible test dates.

You can see these outputs directly in the VPN Index and the full Transparency Report.

Related research pages

Methodology

Full protocol and governance details.

Transparency Report 2026

Live matrix of tested providers and metrics.

VPN Index

Interactive scorecards and filters by use case.

Editorial Policy

How rankings stay independent from commercial pressure.

Affiliate Disclosure

How the business model works and where boundaries sit.

Contact

Report data issues or suggest test additions.