1. Establish a baseline you can repeat
Choose one device and one test service. Pause optional downloads, cloud backups and streams on your own devices. Keep the browser tab visible. Record whether a VPN is active; only disconnect it if it is optional and you are permitted to do so. Do not disable security software or workplace protections for this test.
Run three completed tests in your usual location, waiting for each to finish. Write down download, upload, ping, jitter and request loss separately. Keep failed attempts in your notes as failures, not zero-speed results. Use the median of the completed measurements and record the range.
Tests transfer real data. Check your allowance before repeating them on a metered connection. Stop if testing would interrupt an important call or use data you cannot spare.
2. Change location, not everything else
Move the same device near the router and repeat. Keep the same Wi-Fi network and test service. A large, repeatable improvement makes the wireless path a useful next place to investigate: walls, placement, interference or the link to an extender. It does not identify which of those is responsible.
If your router exposes separate bands, compare them while close to the router and label the band for every run. A higher-frequency band can offer more capacity nearby but is not automatically better through walls. Do not compare a busy evening run on one band with a quiet morning run on another and call that a band comparison.
3. Compare a wired connection, when available
Connect the same computer by Ethernet to your own router, and confirm it is using Ethernet rather than Wi-Fi. Repeat the baseline. If a different device is required, write that down: the comparison now includes a device change, so the result is less conclusive.
Check the negotiated link speed in the device's network settings. A 100 Mbps link can constrain a faster plan. A worn cable, a port limit or an adapter can be relevant, but a result near 100 Mbps alone does not prove any particular component is broken. Do not connect directly to an optical terminal or change provisioning settings without your provider's instructions.
Read the pattern
| Repeatable observation | Next check | What it cannot prove |
|---|---|---|
| Wired is stable; far-room Wi-Fi is poor | Router placement, band, extender path | That buying a new plan will help |
| Only one device is slow near the router | That device's radio, adapter and background traffic | That the provider is throttling it |
| Several wired devices are slow at busy times | Repeat at a quiet time; log household traffic | Whether congestion is at home or upstream |
| Only one website or game is slow | Its service status and route, using its own diagnostics | That the whole internet connection is slow |
Worked example: avoid the wrong upgrade
Illustrative values, not measured customer data: three bedroom runs give 22, 26 and 24 Mbps; near-router runs give 91, 95 and 93 Mbps; wired runs give 94, 96 and 95 Mbps. The medians are 24, 93 and 95 Mbps. This pattern makes the bedroom wireless path worth investigating before paying for a faster plan.
It still does not establish the line's maximum capacity. The wired adapter or service under test might impose another limit. After moving the router, repeat the bedroom test under comparable conditions rather than relying on the attractive near-router result.
Prepare a useful support request
Include the affected activity, dates and local time zone, device, wired or wireless connection, all completed results and failed attempts, and the checks you tried. Ask the provider to investigate repeated wired degradation or disconnections. Do not send Wi-Fi passwords, router administrator credentials or unredacted account information in a public forum.
A concise report might say: “Video calls break up between 20:00 and 21:00. I paused backups and repeated three tests over Ethernet. Here are the results and times. The same device was stable earlier.” This describes evidence without assigning an unproven cause.