Download, upload and ping
Download Mbps shows inbound throughput—streaming, browsing, game patches. Upload Mbps reflects sending—video calls, cloud sync, livestreaming. Ping (latency) is round-trip time to the server; jitter is variation. Gamers and traders obsess over ping; streamers need upload stability.
Wi-Fi versus Ethernet
ISPs’ minimum speed guarantees on regulated products refer to the router—typically tested over wired Ethernet to eliminate wireless variability. A phone on Wi-Fi 2.4 GHz across the house may halve apparent speeds versus what the line delivers. Always test twice: beside the router on 5 GHz, then wired if possible.
Server location and peering
UK-hosted tests better reflect domestic ISP routing. Far-flung servers add geographic latency and may cross congested peering—fine for consistency checks, unfair for blaming local faults. Try multiple test sites when disputing performance.
TCP bottlenecks and browser limits
Browser-based tests may not saturate gigabit lines—use dedicated desktop apps or iperf3 on LAN to separate ISP bottlenecks from test harness ceilings. Small buffers (bufferbloat) inflate ping during loads; cake/FQ_CODEL on advanced routers helps.
Use SwitcherMate’s speed test alongside wired checks to cross-verify marketing claims against your real home setup.
Interpreting volatility
Single tests lie cheerfully. Run clusters: same server, different times; different servers, same time. If variance exceeds ~15% on fibre while wired, investigate bufferbloat, background operating-system updates, or suspicious LAN loops.
Publishing screenshots to forums? Redact your public IP unless you intentionally seek diagnostics from strangers.