So we built a system that tests continuously and publishes everything — averages, peaks, failures — as it happens.
If you've ever searched for “best VPN” or “fastest VPN”, you've seen the problem: dozens of review sites showing slightly different speed results, usually from a single test run months ago, presented as if those numbers reflect how the VPN performs today.
We wanted VPN speed data that actually means something. Not a screenshot from six months ago. Not a single best-case result. Real, continuously collected performance data that gives you an honest picture of what a VPN delivers.
So far we've run 222,615 individual speed tests, and 142 VPN server locations have results from the last 30 days. Every test records download speed, upload speed, ping, jitter, whether the connection succeeded or failed, and a UTC timestamp.
Rankings and averages use the last 30 days of results, and a location needs at least 5 successful tests in that time to be included. We don't sample or weight results, and every successful test counts equally. Failed tests lower a server's pass rate rather than its speed average.
vpnspeeddata is funded through affiliate partnerships. When you click a “Get deal” link for one of the VPNs we test and buy a subscription, we earn a commission at no extra cost to you. It pays for servers, bandwidth and development.
The test code has no awareness of affiliate agreements. Our testing is fully automated, and we report what we measure. We have affiliate links for both providers we currently test, so we earn whichever one you choose.
We only earn if you trust the data enough to act on it — publishing biased results would destroy the one thing that makes this useful. Our editorial policy sets out how we keep the two apart.