Secure-VPN Benchmark Methodology

Every successful VPN run is paired with a clean no-VPN baseline measured immediately before connection on the same worker and using the same throughput endpoint. Raw Mbps is shown alongside relative performance so a fast or slow baseline cannot silently distort the interpretation.

Relative metrics

• Download retention = VPN download ÷ baseline download × 100.
• Upload change = (VPN upload ÷ baseline upload − 1) × 100.
• Latency overhead = VPN latency − baseline ping.

v1.6.3 quality control

Headline heatmap cells use only measurements whose QC row is rankable. The current QC layer flags missing/zero baselines, baselines far below the recent worker median, packet loss at or above 5%, and download retention below 15% for bounded immediate retesting. Persistent degraded observations can be confirmed as degraded while invalid baseline rows remain non-rankable.

Hourly reporting

Time-of-day views group measurements by provider, protocol, destination and UTC hour. The dashboard displays the median value and sample count. Sparse cells remain blank instead of being interpolated.

Source and destination

Source worker is the geographic machine performing the test. Destination is the VPN exit target. They are separate dimensions, so a future Netherlands or Singapore worker can independently measure the same VPN destination.

Commercial independence

Affiliate payout, CPA, rev-share and conversion rate do not enter technical benchmark metrics or ranking.