VPN Comparisons: How to Compare Two VPNs Head to Head

Every "A vs B" page on VPNRadarLab lines up two similar providers from one segment and compares their overall score, six sub-scores, prices, refund window, audit status, protocols, jurisdiction and platform count, marking the stronger side of each row. Speed tests are not part of the picture: the site reads what providers publish, nothing else.

What a head-to-head page shows

Pairs always come from the same segment, so a consumer VPN faces a consumer VPN and a browser-built tool its own kind; a business ZTNA product never meets a free mobile app. The hub holds 300 such matchups: 148 between consumer VPNs, 73 in security suites, 50 among free or mobile apps, 11 for business VPN and ZTNA, 10 for tools built into a browser or OS, and 8 between decentralised VPNs. Those counts describe the matchup list alone, not the wider rankings elsewhere on the site. Every row is filled from published sources: policies, audit reports, app pages, the network map and price pages. Where a provider stays silent, the row says so.

Where the overall score comes from

The total runs from 0 to 10 and is the weighted sum of six sub-scores on the same scale. On the base weights used across the main ranking, privacy and jurisdiction take 30%, security and protocols 20%, price and fair terms and censorship bypass 15% each, network and streaming 10%, platforms and support 10%. The full criteria sit on the methodology page.

Two rules shape the result more than any other factor. A red flag, meaning the provider states it keeps connection IPs, has a public security incident, or falls under Chinese jurisdiction, caps the overall score at 6.0 no matter how strong the rest looks. An undisclosed feature scores zero for that criterion, which is a blank rather than a deduction; deductions apply only to stated facts such as connection or session logging, an incident on record, PPTP or L2TP as the only protocols, or a large renewal jump. A no-logs policy earns 2 points, an independent audit adds 3 more, court confirmation adds 1, and renewal penalties start once the second term costs 1.3 times the cheapest long plan per month.

Why the winner changes from page to page

Matchup pages always use the base weights, so their totals match the main ranking. Topic lists do not: the leading sub-score takes a bigger share (45% on privacy-led pages, 40% on security, price, bypass and network pages, 35% on platform-focused ones) and the other five split the remainder in their usual proportions. A preset can therefore reorder two services that sit close on base weights.

NordVPN against Surfshark shows the effect. On the main ranking the scores are 5.58 to 5.30: NordVPN leads on privacy and security, Surfshark on price, network and platforms. On the security-led list the gap widens to 5.97 against 5.49. On the price-led list Surfshark rises to 5.42 while NordVPN drops off the list entirely. See the NordVPN and Surfshark page for the full row-by-row split.

Red flags and fine print that settle a close call

One public incident can outweigh excellent encryption claims on the scoreboard. The audit row is the sharper tie-breaker: both sides of a matchup can claim "no logs", yet only one may have let an outside firm verify it. When a provider does publish its kill switch, it has already told you more than many rivals that stay silent.

What a comparison here will not tell you

Speed figures would measure one route at one moment, not the service, so this site runs none; streaming and "works in country X" are the provider's own claims, not verified facts. App store and Trustpilot reviews appear on service pages but never move the score. During the refund window, run your own test: connect from your home line, from mobile data, and from the server locations you actually need, then decide. The NordVPN and ExpressVPN page puts both refund windows side by side.

FAQ

Is a higher overall score always the better pick?

Only if the sub-score matching your use points the same way. A slim lead built on platform count is worthless when you need censorship bypass, where the main ranking's median sub-score is just 0.5 and 292 of 377 services score under 2.

Why does one VPN show different scores on different pages?

Topic pages apply their own weight presets; matchup pages and the main ranking use one shared set. Same data, different priorities.

Why do some matchups pit two weak services against each other?

Many readers arrive with a ready list of inexpensive or free apps. For them, the page shows which provider discloses more and which hides a red flag.

Short answer

Read the six sub-scores before the headline total, then check three things that scores alone can hide: a red flag capping the result at 6.0, an independent no-logs audit, and the renewal multiplier behind the intro price. A matchup's totals follow the base weights while topic lists reshuffle the same providers by preset, so a winner on one page can lose on another without anyone changing the facts. If the call is still close, the refund window is your laboratory.