What the test measures
When you press the button, your browser asks this site's server for a very small file ten times, with a short pause between requests, and times how long each answer takes to arrive. The first request is left out of the result, because it also includes setting up the connection. The other nine give:
- Typical response: the middle value (median). Half the requests were faster, half slower. It is the fairest single figure.
- Fastest and slowest: the spread.
- Jitter: the average change from one request to the next. Low jitter means a steady connection.
This is a measure of delay, not of speed in Mbps. A connection can download quickly and still respond slowly, and the reverse. The two are explained in ping, latency and jitter.
How to read your result
| Typical response | Word shown | What to expect |
|---|---|---|
| Under 60 ms | Very quick | Pages open at once; calls and games feel immediate |
| 60 to 150 ms | Good | Normal for home broadband and good mobile data |
| 150 to 300 ms | Fair | Pages pause slightly; fast games feel behind |
| Over 300 ms | Slow | Calls have gaps; pages feel sticky |
Jitter of a few milliseconds is excellent. If jitter is as large as the response time itself, the link is unsteady, which usually means weak Wi-Fi or a busy line.
How this differs from a real ping
The classic ping tool sends a bare network packet, which browsers are not allowed to do. This test sends a full web request instead, so each round trip includes a little extra work at both ends. Expect figures somewhat higher than a command-line ping to the same place. The server is also part of a worldwide content network that answers from a location near you, so the result reflects the path from you to that nearby location, not to a far country.
It is best used for comparison: the same test in two rooms, on Wi-Fi and on mobile data, at noon and at night, or before and after moving the router.
Getting a trustworthy reading
- Run it three times and look at the pattern, not one result.
- Pause downloads, video and cloud backups on all devices first.
- Test once beside the router and once where you normally sit. A big difference points to Wi-Fi, not the line.
- Switch off a VPN to compare; a VPN adds its own delay.
- On a phone, test on Wi-Fi and again on mobile data.
If the result is poor
- High response everywhere, even beside the router: the line or the provider's network is slow to respond. Restart the router; if it persists at all hours, report it.
- Good beside the router, poor elsewhere: a Wi-Fi coverage problem. Work through fixes for slow Wi-Fi.
- Fine most of the time, poor in the evening: the network is crowded at peak hours.
- Wild jitter: someone is uploading or downloading heavily, or the Wi-Fi signal is weak or crowded. Try the 5 GHz band or a cable.
Privacy
The test fetches a small text file from this site, exactly as opening any page does. Nothing about the result is sent anywhere or saved; it is worked out and shown in your browser.
Does the test use much data?
No. Ten requests of a few hundred bytes each, far less than opening one web page.
Why is the first request slower?
It includes finding the server and setting up an encrypted connection. That is why the test leaves it out.
Can I test the response of a different server?
Not from a browser page; browsers limit what a page may contact. Use the ping command on a computer for that.