Time of Day, Server Choice and VPNs: Why Speeds Change

The same line, different results

If you run a speed test at lunchtime and again at nine in the evening, you may get two quite different answers. Nothing is wrong with your equipment. Several outside factors change from one moment to the next, and three of the biggest are the time of day, the server you test against and whether a VPN is active. Knowing how each one works helps you read your results sensibly and avoid chasing problems that are not there.

Time of day and shared capacity

Home internet is usually a shared resource somewhere along the way. Your line goes to a local network point, and that point serves many homes. Beyond it, the provider's wider network carries traffic for a large number of customers. When many people are online at once, for example in the evening when families stream video and play games, the shared parts can become busy. The effect is often called congestion.

Congestion does not affect everyone equally or every day. It depends on how the provider has built its network, how many customers share your section, and what they are doing. Some people never notice it. Others see a clear dip in the evening.

• Busy hours: speeds may be lower, and ping and jitter may be higher, when many people are online.
• Quiet hours: late night and early morning tests often show the connection at its best.
• Your own household: evening is also when your own family is most likely to be using several devices at once.

To see whether time of day affects you, test at the same spot with the same device in the morning, afternoon and evening for a few days. Write the results down. A steady pattern of lower evening speeds suggests shared congestion or heavy household use. No pattern suggests the cause lies elsewhere.

Server choice and distance

A speed test measures the path between your device and a test server. It does not measure your connection to the whole internet at once. This means the server you pick affects what you see.

A server close to you has fewer network hops and less distance to cover, so ping is lower and speeds are often higher. A server far away adds delay, and the path may pass through busy links outside your provider's control. The server itself can also be a limit. If it is busy serving many tests, or if it has less capacity, it may not be able to fill your connection, and the result will look lower than your line can really do.

• Use the automatically selected nearby server as your normal baseline.
• If a result looks odd, try another nearby server and compare.
• When you compare tests over time, use the same server each time.
• Treat a far-off server as a separate experiment, not as your main number.

The same idea applies in daily life. A website hosted far away may feel slower than a local one even on an excellent connection, because the distance and the route add delay that has nothing to do with your plan.

How a VPN changes the picture

A VPN, or virtual private network, sends your traffic through an encrypted tunnel to a server run by the VPN service, and from there on to the internet. This is useful for privacy and for connecting securely to a workplace, but it changes how the connection behaves.

• Extra distance: your data takes a detour through the VPN server, which adds delay and raises ping. If that server is in another country, the effect is bigger.
• Encryption work: scrambling and unscrambling data takes processing effort on your device and on the server, which can reduce speed, especially on older hardware.
• Shared VPN capacity: the VPN server serves many users, and its own load influences your result.
• Different test server: with a VPN on, many tests pick a server near the VPN exit rather than near you.

None of this means a VPN is bad. It means a speed test with a VPN on measures the whole chain, not only your line. If you want to know what your provider delivers, switch the VPN off before testing. If you want to know what you get while using the VPN, test with it on, and record that result separately and clearly labelled.

Other things that vary quietly

• Background activity: automatic updates, photo backups and streaming on other devices can appear and disappear without notice.
• Weather and physical lines: some types of connection are affected by conditions along the line or by wireless signal quality.
• Device state: a device that has been running for a long time, or has many apps open, can test lower than after a restart.
• Router load: a router that has not been restarted for a long time may behave less smoothly.

A simple way to separate the causes

Change one variable at a time. First, test with a cable and no VPN, using a nearby server, at a quiet time. That is your best case. Then change one thing: test in the evening, or switch to WiFi, or turn the VPN on, or pick a distant server. Compare each result with your best case. The difference tells you how much that single factor costs you.

What to do with the findings

If evening speeds are consistently and substantially lower than your best case on a wired test, keep a dated log and share it with your provider when you ask about it. If the difference only appears with a VPN, that is the cost of the tunnel, and you can try a nearer VPN location or a different protocol setting. If one test server looks slow while others are fine, ignore that server. The aim is to understand the pattern, so that you can tell a real fault from normal variation.

Frequently asked questions