How to Test Your Internet Speed Accurately
Why test your internet speed?
Testing shows whether you are getting what you pay for, helps you find the cause of slow streaming or choppy calls, and gives you evidence if you need to contact your provider. One number on its own does not tell you much, so the goal is a clear, repeatable picture.
A method that gives trustworthy results
- Pause everything else. Close streaming apps, downloads, cloud sync and game updates on every device.
- Use a cable for the first test. Plug a laptop straight into the router with an Ethernet cable. This shows the speed of the line itself. If you have no cable, stand next to the router.
- Restart if you have not in a while. Restarting the modem and router clears minor faults.
- Turn off any VPN. A VPN adds a detour. Test with it on only if you want to measure the VPN.
- Run the test three times. Wait a minute between runs, and note the best and the typical result.
- Repeat at a different time of day. Evenings are busier than mornings on many networks.
- Then test on Wi-Fi. Compare the Wi-Fi result with the cable result in the same spot. The gap shows what Wi-Fi costs you.
How our test measures
The speed test measures ping first, then downloads data for about nine seconds over several parallel connections, then uploads for about seven. It ignores the first two seconds of each phase while the connection ramps up, and averages the rest. It also samples your ping while data is moving, which shows how your line copes under load.
Understanding the results
- Download speed: how fast you receive data, which affects streaming and downloads.
- Upload speed: how fast you send data, which affects calls, backups and uploads.
- Ping: the delay of a round trip. Lower is better.
- Jitter: how much ping varies. Lower is steadier.
- Latency under load: your ping while the line is busy. A big rise over idle ping points to bufferbloat.
Why results vary from run to run
Tests are snapshots of a moving target. Your result changes with the time of day, how busy the local network is, how strong your Wi-Fi signal is and what else is using the line. Differences of 10 to 20 percent between runs are normal. A big, consistent shortfall is the thing to investigate.
Why different tests disagree
Each test uses different servers, numbers of connections and timing. A test that uses your provider's own server often reads higher, because the data stays inside their network. A test that uses a distant server reads lower. That is why you should compare runs from the same tool, rather than comparing numbers from different sites.
Limits of any browser test
- Your device matters. An old phone or laptop may not be able to handle more than a few hundred Mbps.
- Wi-Fi caps speed. Even a good Wi-Fi connection often delivers a fraction of the plan speed.
- Very fast plans need a cable. To test a gigabit line, use a wired computer with a gigabit network port and a good cable.
- Data use. The test uses real data, up to about 400 MB per phase, so be careful on mobile plans.
What to do with your results
If your cable result is close to your plan, the line is fine, and any trouble is in the Wi-Fi or on a device. See why your Wi-Fi may be slow. If the cable result is far below your plan, record the date, time and numbers from several tests and read are you getting the speed you pay for?.