How We Test & Review VPN Services

All 56 VPNs on our website have been put through an extensive and fair testing program, ensuring that every VPN is treated exactly the same. We’ve designed this system to make sure we only recommend the very best VPN services.

Full Editorial Independence

At Top10VPN.com we pride ourselves on genuine, 100% editorial independence. What does that mean?

Quality Standards

We are constantly looking to improve our reviews and advice while adhering to rigorous editorial standards of accuracy and honesty.

Every page on Top10VPN.com is compiled by a dedicated editorial team and checked by two or more editors before publishing. Every claim we make can be backed up with cataloged independent research.

We do not rehash or share information from other sites or outlets unless they are a primary source which can be fact checked. We perform all our own research and communicate directly with the VPN services themselves to obtain first-hand information.

All of our research and writing is produced exclusively by our in-house team of VPN researchers, whose work is edited and fact-checked by our VPN experts. The reviews and advice you see on the website is the result of their combined efforts.

Over 30,000 Hours Spent Testing VPN Services

The process of testing a VPN never really ends. Once we select a VPN to review, the work has only just begun.

Every VPN review we publish is continually updated on a weekly basis for the rest of its lifespan. An initial review may involve upwards of 40 hours of testing, but every month will see many more hours spent updating it.

Our VPN experts are intimately familiar with every VPN reviewed on the website. They are constantly retesting VPN speeds, checking if they can still unblock streaming services, updating key information and pricing plans, and dozens more data points.

It’s not just test results that get updated, either. We also regularly assess and update our testing methodologies to ensure they are in line with the fast-changing VPN industry.

Our Tools & Testing Environment

When testing, our main priority is to emulate the results that you would likely get at home. That means we use standard hardware that is readily available to consumers.

Here’s a complete list of all the test devices we use, plus the operating system versions we run:

• Lenovo T480 & Lenovo T480s (Windows 10 & Ubuntu Linux 20.04)
• Apple iMac 2017 Core i5 (macOS Monterey)
• Apple MacBook Pro 2017 Core i5 (macOS Monterey)
• Samsung Galaxy S9 & Samsung Galaxy A8 (Android 12)
• Apple iPhone 8 (iOS 15)
• Apple iPhone 11 (iOS 16)
• Amazon Fire TV Stick 3rd Generation
• Apple TV 4K 1st Generation
• Google Chromecast with Google TV
• Microsoft Xbox One
• Sony PlayStation 4
• Nintendo Switch
• Sony KD65XE

We ensure that every device is kept up to date with the latest available stable OS release. We also routinely refresh our device range when they begin to grow out of date.

These devices are all completely standard, as-sold, with no modifications, hacks, or upgrades. Having our team of experts share this collection of hardware eliminates some of the variables that could otherwise arise between testers. It also means that all VPNs are running on identical hardware, to ensure maximum comparability.

Virtual Machines

On our Mac, Windows, and Linux test devices, every VPN is installed inside its own independent virtual machine instance.

Installing them all straight onto the base machine would mean clogging the hard drive up with dozens of different VPNs, potentially risking clashes and performance issues.

By using virtual machines we ensure that every VPN is given a neutral, unbiased playing field of a completely vanilla OS install. This improves consistency and comparability between test results.

With regards to virtual machine software, for Windows we use Microsoft Hyper-V, for macOS and Linux we use VMWare Fusion, and for our server in China we use VirtualBox.

Our Testing Software

We also use a suite of bespoke testing software. They are designed to test the performance and specific features of each VPN we review. Here’s an overview of the software we use and how they work:

Across all of our reviews and some of our best VPN lists you’ll see ratings assigned for VPNs in various categories. Here is a comprehensive overview of how we test each of those categories.

Privacy & Logging Policy

A VPN’s logging policy is one of its absolute most important attributes. We read every word of every one to fully understand exactly what each VPN logs, where it’s stored, and how long for.

If too much of your data is logged then you could be exposed if a hacker or court gains access to your VPN’s servers.

We want to see as little data logged as possible, especially nothing which could be used to personally identify a user. Aggregated server metadata, for example, is largely fine as it’s far too generic to ever be connected to a user.

Your originating IP address, though, would be unacceptable as it is a direct link between your browsing activity and your real identity.

RAM-based servers are preferable, as they do not store data of any sort beyond the duration of the VPN session. If they are not in place, we like to see logged data deleted as soon as possible, and not being retained for any more than a month.

We also dig into the companies behind each VPN to discern where in the world they’re based, as national laws and jurisdictions can have huge implications for data privacy.

Speed & Reliability

Our speed testing methodology is simple enough for you to replicate it at home.

These tests are always performed in uniquely instanced virtual machines to ensure as much consistency and comparability as possible.

First, we test our dedicated 100Mbps office connection three times using speedtest.net. We use this to determine a control download speed (Mbps), upload speed (Mbps), and ping time (ms).

We then repeat this process with the VPN connected in four locations spread around the world: UK, Germany, East Coast US (ideally New York or Massachusetts), and Australia. We perform three tests per country to create a set of averages.

In the event that servers are not available in those regions we will look for suitable alternatives. If there are no alternatives the VPN has its speed rating capped.

The logic behind this selection is that it gives a full range of distances around the world. You can extrapolate our results to understand how a VPN’s speeds may vary depending on where you’re located and the server you may want to connect to.

We calculate the total loss in download and upload speeds (plus increases to ping time) and convert it into a percentage. We then subtract marks from the speed rating depending on how drastic the speed loss is.

Shorter-distance connections are weighted higher than long distance ones. Reliability and consistency of the connection is also factored in as a variable.

Features, Encryption & Security

A broad and complex category, lots of individual criteria make up this score – most of them relating to the technical side of the VPN.

Streaming

Streaming tests are very simple and reflect the exact experience you would get at home as a typical user.

We have paid subscriptions to multiple popular streaming services from around the world. To test a VPN’s access, we simply connect to a server in the necessary region and try to log in and watch content.

We’ll try a multitude of server locations, protocols, and platforms before coming to a conclusion. When creating a streaming score we weight specific streaming services differently based on popularity. You can see our full streaming rating methodology above.

Here is a list of all the streaming services we have active test subscriptions for:

1. All 4
2. Amazon Prime Video US
3. BBC iPlayer
4. Disney+
5. HBO Max
6. Hotstar
7. Hulu
8. ITV Hub
9. Line TV
10. Netflix (multiple international libraries tested)

Torrenting

We judge VPNs for torrenting based on five key criteria. The most important is speed – the less our download bitrate is affected by the VPN, the better.

1Speed

We measure our download bitrate on our 100Mbps test connection using our dedicated 1GB test file. We then see how the speed drops off when downloading the same file with a VPN connected.It’s the same logic as with our speed testing, only we test solely from a local server connection.

We use qBittorrent as our torrenting client, and we download our own custom-made 1GB test file from a fixed-speed host to keep our results consistent.

2Encryption & Security

We also need to be confident that a VPN can protect and anonymize your torrenting traffic before recommending it. We use our torrenting leaks tool here to make sure that there are no IP and DNS leaks throughout the test.

3Logging Policy & Past Behavior

The third factor is a VPN’s logging policy and past behavior. A VPN should record as few logs as possible and be based in a private jurisdiction in order for us to entrust it with our torrenting data. This rating is taken straight from the Privacy & Logging Policy section of the VPN review.

4Server Network & P2P-Friendly Attitude

We then look at the VPN’s server network, and how P2P-friendly it is. Torrenting policies are categorized as either unlimited, permitted, restricted, or not allowed.

Here’s a breakdown of what these categories mean:

• Unlimited requires a VPN to allow torrenting with no data caps on every single server in its network.
• Permitted is as above, but with perhaps a handful of servers or countries which do not permit torrenting. We will also consider the locations in which torrenting is permitted – there should ideally be a global spread leaving no nation without a realistic option.
• Restricted means that torrenting is allowed, but only on a small minority of servers in the VPN network.
• Not permitted means that you cannot use the VPN for torrenting at all.

5Relevant Extra Features

Finally we consider the VPN’s relevant extra features. A kill switch and port forwarding feature are the two most important, but anything else that could be considered useful for torrenters (such as SOCKS5 proxy) are also taken into account.

Ease of Use

We always strive to review VPNs with scientific rigor and objective metrics. However, some things can’t be expressed in sums and formulas.

This part of our testing covers off all of the intangibles: how does a VPN feel to use? Is it easy to navigate? Are labels clear and straightforward? Are there quality of life features like pre-categorized servers, sorting options, or favorite lists?

We take this into account across every single platform we test on – that’s desktop, mobile, streaming devices, and router. Desktop and mobile are weighted the heaviest due to their popularity, but every device is taken into consideration.

We address all of the differences between platforms, too, so that you know whether a VPN is right for you and your individual setup.

We try to maintain a consistent level of expectation and a mutual understanding of which features are good and which features are bad across the entire testing team.

Customer Support

Online resources (like an FAQ or knowledge center) are the bare minimum for a VPN. They can vary wildly in quality, but simply having something available to help users with troubleshooting is the least we expect.

Alternatively, email or telephone support without online resources to back it up can also be considered the entry level for VPN customer support.

A combination of both is what we want to see most VPNs achieve. We consider the extensiveness of online resources, and the promptness and helpfulness of support agents when testing the VPN here.

We consider live chat to be the pinnacle of customer support. Being able to get a tailored response from a real human (ideally available 24/7) is what separates the best VPNs from the competition.

We put forward real requests to support agents, and we test and fact check all of the solutions they offer.

Across Top10VPN.com you’ll see two different types of ratings for VPNs: overall ratings and use case ratings.

In this section, we’ll explain how our testing data is translated into these numerical ratings, how important we consider each category, and what separates a decent VPN from a great one.

Use Case Ratings

When you want to see the best VPNs for a specific purpose, device, or region, our experts create a specific methodology just for that page.

We consider the most important decision-making factors specifically for that purpose, and assign them a weighting that reflects how much they matter. These subcategory ratings combine to create a unique overall rating for that use case.

Don’t worry if that all sounds a little confusing – you can find a full testing methodology on every best VPN list.

Overall Ratings

An overall rating, which you see on all 56 VPN reviews, is what this page is dedicated to explaining.

Our experts calculate this rating based on a fixed set of testing categories that are important for all VPNs, no matter how you plan on using them. These ratings are formed from years of impartial, first-hand testing data.

We also rate every VPN in three other areas: ‘Price & Value for Money’, ‘Bypassing Censorship’, and ‘Server Locations’.

• Price & Value for Money does not factor into the final overall rating. We believe that price is entirely subjective – what is expensive for one user is affordable for another. The subcategory rating is simply there so that you can compare one VPN to another.
• Bypassing Censorship and Server Locations are also not given their own individual weighting. That is because they are a contributing factor to the 20% allocated to ‘Features, Encryption & Security’.

We highlight it and give it its own subcategory rating, though, as to some readers it is considered an extremely important feature.

Read on to find out how our ratings are calculated and how you should interpret each one.

1Privacy & Logging Policy: 30%

Every logging policy starts off at a 9.9 rating. As we comb through the relevant documents on the VPN’s website, we then subtract points for each negative point we can find.

Some things, such as logging your ISP or country, will only cost 1 point or so. Other more serious logs, like connection timestamps or total bandwidth consumed, will detract more. The absolute worst logs, such as DNS requests or your true IP, will instantly drop the rating to 1.0.

Other factors will apply a modifier to the rating. This includes the VPN jurisdiction, how long logs are retained for, and whether the logged data can be personally tied to your account or not.

Finally, if the VPN has received a third party audit or proven its policy in a court of law, it will receive a flat bonus to its overall rating.

2Speed & Reliability: 25%

Every VPN starts with a speed rating of zero. We add points to it depending on how it performs in each region. Download speeds equal 75% of the total rating (40% on the local connection, 20% to the US, 10% to Germany, 5% to Australia), upload speeds 20% (10% local, 5% US, 4% Germany, 1% Australia), and ping times 5% (local 3%, US 2%).

We then calculate the percentage speed loss between the connection with and without a VPN, plus the added ping time in milliseconds. The greater the loss in speed or increase in ping, the less the VPN’s speed rating increases by.

You can see what our formula sheet looks like above. Every VPN is subjected to this exact same process, with discretionary points awarded or subtracted by the tester depending on consistency and reliability.

3Features, Encryption & Security: 20%

We have a checklist of must-haves in this category which heavily influence its category rating. This includes the presence of OpenVPN or WireGuard, a functioning kill switch, functioning leak prevention, and AES-128 or -256 encryption.

We also check for a range of additional features and bonuses, including (but not limited to): ad blocking, tracker blocking, dedicated IP, Double VPN, Smart DNS, SOCKS, split tunneling, and Tor over VPN.

The server network rating also factors into this section. You can see how we rate it above. The number of countries, the spread of countries, and the number of servers are all factored in.

Finally, we consider its ability to unblock censored internet. We create a rating by assessing the VPN’s consistency on our Shanghai server over the past year. If it does not work at all in China, we consider its additional features (like obfuscation) to assess whether it could feasibly work in a less strictly-censored country (like Russia or the UAE).

4Streaming: 10%

Streaming services are not weighted equally. We consider US Netflix and BBC iPlayer to be the most valuable based on their popularity, therefore they are weighted higher.

After that, we consider a variety of Netflix libraries or a selection of other, less popular services to be equally valuable. The Netflix libraries we test for include the UK, France, Germany, Italy, Japan, and Australia. The other streaming services we test are listed here.

The absolute highest-scoring VPNs will be able to consistently unlock a good number of streaming services in a wide variety of geographical regions.

5Torrenting: 5%

Torrenting ratings are arrived at through a purely mathematical equation.

We measure our download bitrate on our 100Mbps test connection using our dedicated 1GB test file. We then see how the speed drops off when downloading the same file with a VPN connected. This is worth a maximum of 3 points, depending on download speed loss.

The encryption and security rating of the VPN is scaled down to equal two points of the overall torrenting rating. The logging policy and past behavior of the service is also worth up to 2 points.

The array of torrenting-friendly servers is worth a maximum of two points. The fewer available across the VPN’s network, the lower the rating. The global spread of these servers is also considered.

The additional features available that aid with torrenting make up the final two points. The kill switch and port forwarding are the highest-weighted.

6Ease of Use: 5%

Most VPNs perform well in this category as, fundamentally, VPNs are simple apps that tend to largely resemble one another.

To lose substantial points in this category, our VPN testers will have to experience difficulty installing or signing up to the service, struggle to work the app in day-to-day use, or encounter regular bugs.

To gain substantial points the VPN will have to be extremely intuitive to use across a broad range of platforms. Having an app for Windows, macOS, iOS, and Android is the bare minimum – there should also be apps for Linux, Fire TV, or router.

We will always provide plenty of screenshots in this section so that you can better understand any points we make about the VPN’s appearance and usability.

7Customer Support: 5%

We break customer support down into five different tiers as described below.

Within those tiers it is up to our VPN testers to decide which precise rating the service should receive depending on the quality of the support team’s responses. Slow and inaccurate responses will receive a lower rating than responses which are received promptly and with lots of helpful information.

Just like everything else we test, we routinely quiz VPN support agents to make sure that standards aren’t slipping. We ask them deliberately technical questions that we already know the answer to so that we can accurately assess them.