HIZTESTI.COM
Dizüstü bilgisayar başında çalışan, gözlüklü bir kişi

Measuring correctly

Your device may be slowing down the speed test

Last updated: 5 min read

It is your device's processor that runs the speed test. If your computer or phone is bogged down with other work during the test, the measuring code cannot run fast enough and the result comes out below your real speed. That is a limit of your device, not your internet. Closing open apps and measuring again usually reveals the difference.

Internet is fast, so why is the result low?

A speed test is not an abstract measurement; it is a program run by your device's processor. It has to process incoming data, do the counting and draw it to the screen. If your device is already struggling at that moment (many apps open, an update running, the system overheating), the measuring code has to wait and cannot process the data at the speed it arrives. The result drops because of your device, not your internet.

This is the diagnosis most likely to cause an injustice: people see a low result and blame the ISP, when the bottleneck is at this end of the table.

How do you know it is the device?

  • Your device is slow in general: apps open late, the fan runs constantly, everything stutters.
  • The result changes markedly from test to test, more erratically than Wi-Fi fluctuation could explain.
  • Heavy work is running during the test: video editing, gaming, many browser tabs, a backup.
  • When you test on another device on the same line, the result is clearly higher.

That last point is the most reliable test: measure on a second device connected to the same Wi-Fi or router. If the result is high there, the problem is the first device, not your line.

What to do

  1. Close open apps, especially video, gaming and editing software, and anything updating in the background.
  2. Close browser tabs you are not using; let the speed test stand alone.
  3. Restart the device. It is usually the fastest way to clear accumulated background work.
  4. If the device is hot, wait a while. Overheating processors slow down to protect themselves (thermal throttling); performance returns once they cool.
  5. Then run the test again. If the result rises noticeably, the earlier low number was your device's limit, not your internet's.

What do manufacturers say?

Device performance being variable is acknowledged by manufacturers themselves. Apple publishes a help page for a slow iPhone or iPad, recommending as first steps restarting the device, keeping it updated, and freeing up storage.

If your iPhone or iPad is running slow, learn what you can do to help.[1]

The same logic applies to Windows PCs. Microsoft lists closing unused apps as one of the first steps to improve performance:

Having many apps or browser tabs open can slow performance.[2]

On Windows you can press Ctrl + Shift + Esc to open Task Manager, close the apps you are not using, and re-run the test. In short: a device without enough resources, or busy with other work, can produce a slow result on even the fastest internet.

If you want to know how we tell a device-limited measurement from a line-limited one, our methodology paper is open. Measurement methodology

Frequently asked questions

Can an old computer not make full use of fast internet?
Usually it can, but there are limits. A very old device's processor or network card may not be able to handle very high speeds (e.g. gigabit); the result then caps at the device's ceiling, not your line's. For everyday speeds, older devices are generally fine.
Should I test on my phone or my computer?
Both are valid, but they measure different things. A phone usually measures over Wi-Fi and with a weaker processor. If you want your line's real capacity, a computer that can connect by cable gives the most reliable result.
Do antivirus or security tools affect the result?
They can. Some security software scans passing traffic, which both occupies the processor and can slow the measurement. If in doubt, disable it temporarily and re-test.
Does restarting the device really make a difference?
Often, yes. A restart clears accumulated background work and freed-up memory, leaving more resources for the measuring code. If the low result was device-related, the difference is usually visible after a restart.

Sources

Did this guide help?

Run the test again