How to Run a Fair VPN Speed Test

Almost everyone has done it. You connect the VPN, open a speed test website, see a smaller number than usual and conclude that the VPN is slow. Or you see a larger number and conclude that it is fast. Both conclusions are usually wrong, because a single speed test taken in a hurry measures a dozen things at once and the VPN is only one of them.

A fair VPN speed test is a small experiment. You decide what you want to learn, change one thing at a time, repeat the measurement, and write the results down. This article gives you a method you can follow in about fifteen minutes, with a checklist, a log template and a guide to reading the numbers. It does not name any provider and it does not promise a particular speed, because your result depends on your line, your device and your distance from the server. If you want to understand why a tunnel changes speed at all, why internet speed changes with a VPN explains the mechanics. This article is about measuring.

What a VPN speed test can and cannot tell you

A speed test is a short, artificial download and upload between your device and a test server. It reports three things: download speed, upload speed and latency (often labelled ping). Cloudflare's explanation of latency is a good plain reference if the term is new to you.

What a speed test can tell you:

• How much your throughput changes when the tunnel is on compared with off, on the same line, at about the same time.
• Whether one VPN location performs noticeably better than another from where you are.
• Whether a change you made, such as switching from Wi-Fi to a cable, has helped.

What it cannot tell you:

• How fast any specific website or app will feel. A website loads from its own servers, which may be far away or busy.
• Whether the VPN is secure or private. Speed and protection are separate questions, and how to test a VPN covers leaks and the kill switch.
• A universal rating. Your numbers belong to your line, your device and your hour of the day.

Keep this limit in mind. The goal is a fair comparison on your own setup, not a score to post online.

Decide what you want to learn first

Before you open any test, write one sentence describing the question. Different questions need different tests.

Your question	What to measure	Main number to watch
Will video streaming be smooth?	Download, repeated	Download in Mbps
Will video calls hold up?	Upload and latency	Upload and ping
Is the VPN OK for online gaming?	Latency and stability	Ping and its variation
Is a large download too slow?	Sustained download	Download over 30 or more seconds
Which location should I pick?	The same test at each location	Download plus ping
Did my change help?	Before and after, same conditions	Whichever you changed for

If you skip this step you tend to chase the biggest download number, which is rarely what matters for calls or games. Write the question at the top of your log so you remember what "better" means.

Build a baseline before you connect

The baseline is the most important part and the one most people skip. A baseline is your speed with the VPN off, measured the same way you will measure it with the VPN on. Without it, a VPN number means nothing, because you do not know what your line delivers.

Prepare the environment:

1. Pick one device. Use the same phone or laptop for every run.
2. Pick one connection. Cable if possible. On Wi-Fi, stand in the same place near the router and do not move.
3. Close competing traffic. Pause cloud backups, system updates, torrent clients, game launchers and streaming on other devices in the house.
4. Restart nothing in the middle. Do not restart the router between runs unless you are testing the router itself.
5. Use one test site and one test server. Many sites let you pick the test server. Note its name and location and keep it for the whole session.
6. Note the time. Home networks slow down in the evening when neighbours are online.

Then run the VPN-off test three times, waiting about thirty seconds between runs. Record download, upload and ping each time. Take the median of each column. These three numbers are your baseline.

The step-by-step test with the VPN on

Now repeat the same procedure with the tunnel up. Keep everything else identical.

1. Connect the VPN app and wait until it clearly shows connected. If you use XPASS, the free VPNclient app signs in with the username and password on your My Pass page; see the download page for the official apps.
2. Check that your normal pages load before you test, so you know the tunnel is really passing traffic.
3. Run the same test site with the same test server three times, thirty seconds apart.
4. Record download, upload and ping for each run.
5. Take the median of each column.
6. Switch the VPN off and run the baseline again once. If the new baseline is far from the first one, your line changed during the session and the comparison is unreliable. Start again later.

That last step, a closing baseline, is what separates a fair test from a lucky one. Home connections vary minute by minute. Bracketing your VPN runs with two baselines tells you how much of any difference could be ordinary drift.

Be careful with the test server. Some test sites default to the server nearest your IP address. When the VPN is on, your apparent location changes, so the site may pick a different test server than it did during the baseline. That alone can change the result. Lock the same test server for both measurements, or at least note that it changed.

A log you can copy

Keep a plain table in a notes app or spreadsheet. Here is a layout that works.

Run	Time	Connection	VPN	Location	Down (Mbps)	Up (Mbps)	Ping (ms)
1	20:05	Cable	Off	n/a
2	20:06	Cable	Off	n/a
3	20:07	Cable	Off	n/a
4	20:10	Cable	On	Location A
5	20:11	Cable	On	Location A
6	20:12	Cable	On	Location A
7	20:15	Cable	Off	n/a

A calculation to keep handy is the percentage change: take the VPN median, divide by the baseline median, and subtract from one. If the baseline download was 200 Mbps and the VPN median was 150 Mbps, the VPN result is 75 percent of baseline, a drop of 25 percent. Make a note of both the percentage and the raw number, because a 25 percent drop from 500 Mbps still leaves plenty for most tasks, while a 25 percent drop from 20 Mbps may be noticeable.

Those figures are illustrations of the arithmetic, not typical or promised results.

Reading the numbers sensibly

Once the log is filled in, look at it with these questions.

Is the spread small? If the three VPN runs are within roughly ten to fifteen percent of each other, the median is meaningful. If one run is half of another, something else is moving, so run more tests or choose another time of day.

Is the drop in line with the cause? Encryption and the extra path add overhead. A modest drop is normal. A very large drop usually points at distance, a busy server, a weak Wi-Fi spot or a protocol setting rather than encryption itself. Why is my VPN so slow lists those causes in order.

Is the bottleneck your own line? If your baseline is already close to the speed of your plan, the VPN may be limited by the plan itself. Nothing in a VPN can exceed what your line delivers.

Does ping change in a way that matters? For browsing, an extra few dozen milliseconds hardly matters. For gaming or remote desktop, ping and its variation matter more than raw download. See does a VPN reduce ping for what to expect and why.

Does upload look odd? Upload is often lower than download on home lines, and the tunnel adds overhead in both directions. Judge upload against your own baseline, not against the download number.

Testing different locations fairly

If you are choosing between locations, keep every variable except the location constant. A workable routine:

1. Run the baseline with the VPN off.
2. Connect to Location A, run three tests, record the median.
3. Disconnect, wait a minute, then connect to Location B and repeat.
4. Return to Location A for a final check. If its result has moved a lot, the line changed during the session and the comparison is weak.

Do not pick the winner by one test. A location that looks best at 14:00 may be quite different at 21:00, so test at the time of day when you actually use the VPN. For the reasoning behind distance and location, read how to choose a VPN server location.

If your app lets you change settings such as the connection type, change one setting at a time and test after each change. If you change three things at once you will not know which one helped. When in doubt, leave app defaults alone unless support suggests otherwise.

Common testing mistakes

• Testing once. A single result is an anecdote. Use at least three runs per condition.
• Comparing different test servers. The VPN changes which test server a site auto-selects, which can mislead you more than the VPN does.
• Testing during a download or update. Another process silently eats your bandwidth and blames the VPN.
• Switching between Wi-Fi and cable mid-test. The difference between them can be bigger than the VPN effect.
• Testing straight after connecting. Give the connection a few seconds to settle, but do not wait minutes either; if it needs minutes, that is a finding in itself.
• Believing a big download number means a fast experience. Latency, DNS and the distant website decide how pages feel.
• Using a very short test and ignoring sustained speed. Some links start fast and settle lower. For large downloads, watch a longer transfer.
• Comparing results with another person. Their line, device and city are different, so their numbers do not transfer to you.
• Testing a tunnel that is not really working. If your pages do not load normally, fix the connection first with how to check if your VPN is working.

Scenarios: three people, three tests

The evening streamer. Pim wants to watch video in the evening on a laptop. She tests with a cable at 21:00, three runs with the VPN off and three on, using one fixed test server. Her download median drops moderately, which she notes is still far above what the video needs. She stops worrying and enjoys the show. Her lesson: compare against the requirement, not against the baseline alone.

The remote worker on video calls. Arun spends the day in meetings. Download barely matters to him; he tests upload and ping. His VPN-on upload is lower than baseline but comfortably above what a call uses, while ping goes up by a visible amount and varies. He tries a nearer location and repeats the same three runs. The ping falls and steadies, so he keeps that one. His lesson: test the number your task depends on.

The student on shared dorm Wi-Fi. Mei's results swing wildly, even with the VPN off. Her baseline runs vary by more than half. She concludes that the network itself is unstable and that no VPN comparison will be fair there. She tests later at night when fewer people are online and gets consistent baselines. Her lesson: if the baseline is noisy, the test is not ready.

When the result is bad: next steps

If the fair test shows a large drop that really hurts your use:

1. Try a nearer location and repeat the test. Distance adds latency.
2. Switch from Wi-Fi to a cable or move closer to the router.
3. Test at another time of day to see whether the server or your local network is busy.
4. Restart the VPN app and, if needed, the device. Then retest.
5. Check for something else using bandwidth.
6. Read how to improve VPN connection stability for reconnect and network tips.
7. If it is still poor, note your log, the times and locations you tested, and send these to the provider's support. A tidy log helps a support agent far more than "it is slow".

Summary

A fair VPN speed test needs a clear question, one device, one connection, one test server, a baseline with the VPN off, at least three runs per condition and a closing baseline to check for drift. Read the median, not the best result, compare it with what your activity needs, and remember that a speed test measures a short burst, not how every website will feel.

When you are ready to run the test on your own devices, the pricing page shows the prepaid passes, and a short pass is enough to try the method for yourself.