How we review
Every review site says it is independent. We would rather explain exactly how our assessments come about, which sources we use, what we cannot establish and where the money comes from. Then you can decide for yourself what our verdict is worth.
What we are and are not
VPNwijzer is a small, independent Dutch initiative. We have no test lab, no server room and no team of measurement engineers, and we do not pretend otherwise. You will therefore never read a sentence on this site like "we measured 847 Mbit/s on the Amsterdam server", because we did not measure that.
What we do is what many comparison sites skip. We collect prices ourselves from a Dutch connection, with a reference date, because VPN providers show different amounts per country. We read audit reports to see who did the audit, when, and what exactly was examined, instead of copying "externally audited" from a press release. And we trace claims back to a source you can open yourself; those sources are listed at the bottom of every review.
Our five criteria
Every service is assessed on the same five points, in this order of weight. For each criterion we say what we use and where our knowledge ends.
1. Privacy
What we look at: the logging policy as the provider words it, and above all whether an outside party has examined it. We note who did the audit (Deloitte, KPMG, Securitum, Cure53 and so on), on what date, and what the scope was. The last point matters: a no-logs assurance engagement examines something different from a pentest of the apps, and both are sold as "audited" in marketing. We also look at the jurisdiction, whether the servers run on RAM, and who owns the brand.
Where our knowledge ends: we cannot look inside the server room. An audit is a snapshot by a third party, not a continuing guarantee. If something is known about a provider that tested its no-logs policy in practice, such as a court case or a seizure, we weigh that more heavily than a report, because it shows what was actually stored.
2. Speed
What we look at: independent lab measurements from parties set up for it, with the measurement date. We also look at whether there are servers in the Netherlands and which protocol the service uses. We deliberately distinguish throughput from latency: the first decides whether streaming runs smoothly, the second whether gaming is playable. Services that score well on throughput can score badly on latency, and that is exactly the kind of difference a single number wipes out.
Where our knowledge ends: speed depends on your connection, your distance to the server and the time of day. Every number in a review is an indication of the ranking between providers, not a prediction of what you will get.
3. Price over the full term
What we look at: the entry price of the longest subscription, the full amount you pay upfront, and the price at which the subscription renews afterwards. Those three together determine what a service costs; the amount on the button does not. We calculate the real cost over four years, assuming automatic renewal, and that really does change the ranking. The working-out is in Wat kost een VPN echt? (Dutch).
If a provider does not publish its renewal price anywhere on the pricing page, we say so explicitly and it counts as a drawback. Not because that price would be high, we do not know, but because when you sign up you cannot know what you are getting into.
Where our knowledge ends: promotions change weekly and sometimes per visitor. Some providers show amounts excluding VAT. Our prices are therefore always indicative and carry a reference date; check the current price with the provider before you pay.
4. Ease of use
What we look at: which platforms have apps, whether they are available in Dutch, how many steps it takes to connect, and whether basic features such as the kill switch can be found without a manual. We pay extra attention to features that exist on one operating system and not on another, because that is rarely in the providers' own comparison tables.
Where our knowledge ends: ease of use is partly taste. What an advanced user calls "plenty of settings", a beginner calls "confusing". We therefore try to describe for whom something works well instead of attaching a number to it.
5. Streaming reliability
What we look at: whether there are dedicated streaming servers and what independent sources recently reported about which services work.
Where our knowledge ends: of the five criteria this one spoils by far the fastest. Streaming services actively work against VPNs, so what works today may stutter next month, at every provider. We therefore word this deliberately cautiously and never promise that something works. What you can realistically expect is in our streaming guide (Dutch).
How a review comes about
The order is the same every time: collect the facts first, form a verdict afterwards, not the other way round.
One. Collect prices, plans and terms from the provider's Dutch pricing page, from a Dutch connection, with the date. Two. Look up the renewal price, even if it is in the small print, and note when it is missing. Three. Read audit and transparency reports for who did them, when and the scope. Four. Go through the provider's history for incidents, acquisitions and ownership changes. Five. Find independent measurements for speed, latency and streaming. Six. Only then write, with the drawbacks in the same text as the strengths and with a source list in which every non-obvious claim can be traced.
When we update a page we always repeat steps one and two, even if nothing has changed in substance, because prices go out of date fastest.
Our sources
Every review ends with a source list. For each source it says what we took from it and when we looked at it, so you can check our reasoning instead of having to believe it. We use four kinds of sources: the provider itself for prices, plans and terms; audit reports and their announcements; independent trade media and test sites for measurements we cannot do; and reporting on incidents, preferably from the period in which they happened.
We mark links to providers so that they pass on no search engine value; links to independent parties we link normally. A source reference is expressly not a recommendation to buy anything there.
If something is factually wrong, we would like to hear it at u00w4j9@proton.me. We correct quickly and state on the page that something has changed. Quietly making a correction does not count as correcting for us.
How we keep prices up to date
The comparison data lives in one place, in a machine-readable file with a reference date: data/vpns.json. Everything you see in the comparison tool is generated from it, so the table and the text cannot drift apart. That file is public, so you can see for yourself which figures we use and from what date. Prices are checked weekly; articles show the month of their last substantive update.
What we do not do
We do not write invented test results. We do not use scare tactics; a VPN is a useful tool, not protection against everything, and what a VPN does not do (Dutch) is stated just as clearly. We do not write "best VPN ever" language or artificial urgency about discounts that will still be there next week.
We also do not publish pages that exist only to rank for a search term. Concretely: we once made discount, Black Friday and free-trial pages per provider and removed them again. They added nothing in substance to what was already in the reviews, they fell under what Google calls scaled content abuse, and for at least one provider they conflicted with the partner terms. Those URLs now redirect to the page that really answers the question. We say this here because it is more honest than pretending it never happened.
How we make money, and what that does and does not influence
VPNwijzer earns affiliate commissions: if you take out a subscription through a link on our site, we receive a fee for it. It costs you nothing extra. Four things matter here.
Paid placement does not exist here. Providers cannot buy a position in our comparison, a review verdict or a place in the comparison tool. The order in the comparison tool is sorted by price by default and changes only through your filters; the recommendations on the comparison page are editorial and argued on the merits. The size of the commission plays no role in either.
Not every provider pays us, and you do not see that in the verdict. At the moment we have a partner programme with NordVPN, Surfshark, PureVPN and Proton VPN. We do not have one with CyberGhost, Private Internet Access, IPVanish and ExpressVPN; those links earn us nothing and are there because the services are relevant. That is why Private Internet Access is at the top of our price comparison, while we earn nothing from it.
Some services deliberately have no programme. Mullvad, for example, a provider with an excellent privacy reputation, does not do affiliate. That is why Mullvad is missing from many comparison sites, and we think you should know that.
The incentive exists, and we name it. Commissions differ per provider and per subscription type. Our defence is structural: fixed criteria for every review, drawbacks that are always named, even for the services we earn the most from, and this page, so you can hold us to it.
Conflicts of interest we have to report
NordVPN and Surfshark have both been part of Nord Security since the merger. We have a partner programme with both and so earn from two brands of the same owner. That makes our comparison of those two (Dutch) a comparison in which we win whichever way you choose. That is why that fact is at the top of that page instead of below.
Also good to know: of the eight services we cover, five are owned by two companies. Kape Technologies owns ExpressVPN, CyberGhost and Private Internet Access; Nord Security owns NordVPN and Surfshark. The market is narrower than it looks, and we say so for every service where it applies.
Recognising affiliate links
Every page with affiliate links says so at the top, not tucked away at the bottom. The links themselves are marked with rel="sponsored", the technical standard that also tells search engines they are commercial links.
All outgoing buttons also run through an address of our own on our domain, starting with /go/, which forwards you to the provider. That lets us count how often a button is clicked. We do that without a cookie on our domain and without personal data; we see numbers per page, not who clicked. Our full policy is in the affiliate disclosure and privacy statement, and you can read who we are on About.