Why is my WiFi slow? Find the real fault

/01 — DIAGNOSE FIRST

Seven tests, in this order, before you change anything

Changing settings before you have measured anything is how people end up three hours in with a worse network than they started with. Do these first and write down three numbers each time — download, upload and ping.

1. Wired versus wireless, in the same spot, in the same minute. This single test splits every fault into “the line” or “the WiFi”.
2. At the router versus in the problem room, recording band, channel and signal strength each time.
3. 11am versus 8–10pm. Congestion faults only appear in the evening.
4. Latency and packet loss, not just megabits. Most “slow” complaints are actually latency complaints.
5. Latency under load — start a big download and watch the ping. This is the test almost nobody runs and it explains more faults than any other.
6. The provider’s status page. Ten seconds, and it occasionally ends the whole investigation.
7. On copper lines only, the master test socket — and only where the faceplate has a horizontal split across it. If there is no horizontal bar, do not unscrew it.

/03 — INTERFERENCE

Interference: what actually matters in a UK home

• Microwave ovens leak around 2.45 GHz, and the peak sits over 2.4 GHz channels 9–11 — so channel 1 is the least affected. This is a genuine, measurable effect for the two minutes the oven runs.
• Baby monitors: DECT models work around 1.9 GHz and do not affect WiFi at all. The 2.4 GHz frequency-hopping ones hop across the whole band and genuinely do. It is worth knowing which you have before blaming it.
• Cheap analogue 2.4 GHz cameras and video senders transmit continuously and are among the worst offenders in any house.
• USB 3.0 ports, external drives and unshielded cables radiate broadband noise centred right in the 2.4 GHz band. Move any 2.4 GHz hub or dongle away from them — this one surprises people.
• Neighbouring networks are usually the dominant interferer in flats and terraces, and there is no setting that removes them.

An honest limitation: a WiFi analyser app only shows you networks. If the noise floor is raised with no networks visible, the source is something that isn’t WiFi — and proving what needs either a spectrum analyser or the patient business of switching suspects off one at a time.

/05 — WHEN IT IS THE LINE

When it is the line, not the WiFi

This is the section that costs us work and we are writing it anyway.

If a wired test at the router shows 8 Mbps down and 0.7 Mbps up, no mesh, extender or powerline adapter changes that number. Not a better one. Not a more expensive one. The honest advice is to look at the package, the provider, or whether full fibre has reached your street — not at hardware.

Copper lines and stability management: if a line keeps resyncing, the network automatically makes it more stable and slower — interleaving adds a few milliseconds of latency, and banding typically drops the sync speed by 10–20% to buy reliability. Fix the noise source first, then ask the provider to reset it. Rebooting the router every day because it feels better is the fastest way to a permanently slower line.

Your right to leave, stated accurately: Ofcom’s speed codes of practice let you exit — but only with providers signed up to them, only where speed falls below the minimum guaranteed figure (not the advertised average), and only where the fault is on the provider’s network. In-home WiFi problems do not qualify, and anyone telling you otherwise is setting you up for a disappointing phone call.