How to Test Your Actual Internet Speed (And What the Numbers Mean)

A speed test takes twenty seconds and most people run it in the way least likely to produce a useful answer: over Wi-Fi, from the far end of the house, while a console is updating. Then they compare the result to a plan speed that was never promised at that location anyway.

The short answer

Test on a wired connection, with everything else idle, at least three times across different hours. That result is your line speed. Anything measured over Wi-Fi tests your Wi-Fi, not your internet.

How to run a test that means something

1. 1Plug in with Ethernet if at all possible. If you cannot, stand next to the router on the 5 GHz band and note that you are measuring Wi-Fi, not the line.
2. 2Stop everything else. Pause downloads, close streaming, and check no console is updating in another room.
3. 3Reboot the router if it has been up for months. It costs two minutes and fixes a surprising share of problems.
4. 4Run three tests on different servers or services, and take the best result — you are looking for the ceiling, not the average of a congested moment.
5. 5Repeat at peak time, around 8-10pm. The gap between quiet-hour and peak-hour results is the most revealing measurement you will take.

What the four numbers mean

Download

How fast data reaches you, in Mbps. Divide by 8 for the MB/s your downloads will show. This is the headline number and usually the least interesting one.

Upload

How fast data leaves you. On cable and 5G this is often ten to twenty times smaller than download, and it is what fails first on video calls and cloud backups. See upload vs download speed.

Ping / latency

How long a single packet takes to make a round trip, in milliseconds. Nothing to do with speed. Under 20 ms is excellent, under 50 ms fine, over 100 ms noticeable in games and calls. A connection can be fast and laggy at the same time.

Jitter

How much the latency varies. Low jitter matters more than low latency for voice and video — a steady 40 ms is far better than one bouncing between 10 and 90 ms.

Reading your result

If a wired test consistently returns under 70% of your plan at multiple times of day, that is a legitimate complaint. If a Wi-Fi test does, that is a home networking problem.

Why results vary so much

• The test server. A distant or busy server caps your result regardless of your line.
• Your own hardware. An old laptop with a 100 Mbps Ethernet port cannot report more than 100 Mbps, however fast the line is. Some Wi-Fi adapters are similarly limited.
• Peak-time contention. Shared cable and 5G segments slow when the neighbourhood is online.
• The router itself. Budget routers can struggle to route gigabit traffic, especially with security features enabled.
• Browser overhead. A native app usually measures higher than a browser test on very fast connections.

Turning the result into something useful

A speed test in Mbps only becomes actionable once you convert it:

1. 1Divide by 8 to get MB/s — what downloads will show. Use the converter.
2. 2Compare to a real download. Start a large one and check the peak. It should land within about 20% of the converted figure.
3. 3Estimate a real task. Feed the figure into the download time calculator with the size of something you actually need.

Before you call your provider

Have these ready and the conversation will be much shorter:

• Three wired test results, at different times of day, with timestamps
• The advertised speed and, if published, the peak-time average for your plan
• Confirmation that you tested wired, with other devices idle
• The router's own speed test result, if it has one

Providers deal with a lot of reports that turn out to be Wi-Fi or unit confusion. Arriving with wired evidence moves you past that conversation immediately.