How to Test Your Internet Speed Accurately
A bad speed test is worse than useless. It lies to you. You might think your internet is slow when you just tested it wrong — at the wrong time, on the wrong device, with your whole family streaming in the background. Here's how to get a number you can actually trust.
The 5-Step Checklist Before You Test
- 1
-
2
Close everything.
Every browser tab with video, every background app, every cloud sync, every software update — they all steal bandwidth from your test. On Windows: Task Manager → Network column. On Mac: Activity Monitor → Network tab. Find anything using the network and stop it before you test.
-
3
Make sure no other device is active.
A streaming phone, gaming console, or smart TV in another room splits your connection. Either disconnect other devices for a minute or test early in the morning when no one else is online.
-
4
Run the test 3–5 times and average the results.
One test can be a fluke — a quick spike in congestion or a busy server. Three tests give a much better picture. If the numbers jump all over the place (like 95, 42, 88 Mbps), that's actually useful. It tells you something about jitter or congestion on your line.
-
5
Test at different times of day.
Evening results are almost always lower than morning results. Everyone in your neighborhood is online at once. Test at 6am for your best-case speed. Test at 7–9pm to see what you actually get at night. The gap between those two numbers tells you how much congestion is hitting you.
What to Do With the Results
Once you have a solid reading, the pattern tells you where the problem is:
| What the results show | Likely cause | Next step |
|---|---|---|
| Wired fast, Wi-Fi slow | Wi-Fi issue — not the ISP | Fix your Wi-Fi: router position, 5 GHz band, fewer walls |
| Fast in mornings, slow evenings | ISP peak-hour congestion | Document with timestamps and contact your ISP, or adjust when you use bandwidth |
| Always slow no matter what time or connection type | Plan limit or damaged ISP equipment | Check your plan speed, then escalate to your ISP with proof |
| Wild swings between tests | Network instability — jitter or packet loss | Run a jitter/packet loss test. Check cables and your router |
| Consistent and close to your plan speed | No issue | Nothing to fix — slowness you notice is probably the device or the website's server |
Test From Different Devices
Run the same test from your laptop and your phone — both on Wi-Fi. If your phone is way faster, the problem is your laptop. It could be a driver issue, a background process, or something in the settings. If both are slow on Wi-Fi but fast on a cable, your Wi-Fi environment is the problem, not your ISP.
This two-device trick finds the problem fast. It takes two minutes and can save you a 45-minute call with your ISP.
About Test Servers
Speed tests pick the nearest server by default. A local server will always give you faster results and lower ping than a server far away. Both numbers are valid — they just measure different things.
For checking your ISP's performance, the local server is the right pick. It cuts out everything outside your ISP's control. It shows you what your connection can actually do.
Want to know your ping to a game server in Frankfurt when you're in Australia? Test against a Frankfurt server. That number — not the local result — reflects your real gaming experience.
Some tools let you pick the server yourself. Try one auto-selected test and one against a server in a different city. The difference shows how well your ISP routes traffic to different places.
Which Speed Test Tool to Use
The tool you use matters. Our partner speedtest.now is built for exactly this kind of careful testing. It measures all five key numbers at once — download, upload, ping, jitter, and packet loss — using multi-stream testing to fully push your connection. It also has a dedicated loaded latency test (to catch bufferbloat), a gaming performance test, and ISP comparison data so you can see how your results stack up against 230+ providers in 77 countries. All free, no ads.