VPNGuide.co.uk

How we test and score VPNs

Every number on VPNGuide.co.uk comes from a documented procedure. This page explains exactly how our tests are run and how the scores are produced.

Speed testing

  • Test connection: 500 Mbps symmetric UK full-fibre line
  • Test location: London, United Kingdom
  • Device: desktop PC, wired Ethernet, no other active traffic
  • Operating system: Windows 11 and macOS, latest public VPN app versions
  • Protocol: fastest protocol offered by each provider (WireGuard, NordLynx or Lightway)
  • Server selection: provider-recommended server plus one manual server per region
  • Number of tests: 10 runs per server, per session, morning and evening
  • Calculation: mean of all runs after discarding the highest and lowest result
  • Baseline: unprotected speed measured immediately before each VPN run

Speed loss is the percentage difference between the baseline and the VPN result from the same session. We never compare a VPN result against a baseline taken on a different day.

Score weighting

The overall VPNGuide Score is a weighted average of seven category scores. The weights are fixed for all providers.

  • Speed — 20%. Average UK download speed, speed loss and latency across repeated tests.
  • Privacy — 20%. Jurisdiction, logging policy, independent audits and ownership transparency.
  • Security — 15%. Encryption, protocol implementation, kill switch and leak protection.
  • Streaming — 15%. Reliability with BBC iPlayer, ITVX, Netflix UK and other UK services.
  • Features — 12%. Split tunnelling, multi-hop, ad blocking, dedicated IP and platform coverage.
  • Ease of use — 9%. Setup time, app clarity and support quality.
  • Value — 9%. Effective monthly cost, renewal pricing and device allowance.

Commercial independence

Scores are never adjusted because a provider pays a higher commission, and we include providers with no affiliate programme at all — Mullvad is one example. Where a figure such as server count is supplied by the provider rather than independently verified, we label it as provider-supplied.

Re-testing schedule

  • Prices checked at least every 14 days
  • Server counts checked at least every 30 days
  • Speed tests repeated at least every 60 days
  • Full reviews revisited at least every 180 days

Corrections

If you believe any figure on this site is out of date or wrong, contact us and we will check it. Corrections are applied to the underlying data, so every page updates at once.